Professional Documents
Culture Documents
COAL A N D COAL-RELATED C O M P O U N D S
STRUCTURES, REACTIVITY AND CATALYTIC REACTIONS
This Page Intentionally Left Blank
Studies in Surface Science and Catalysis
Advisory Editors: B. Delmon and J. T. Yates
Series Editor: G. Centi
Vol. 150
Preface ................................................................................................................................ ix
1 M e t h o d s of C l a s s i f i c a t i o n a n d C h a r a c t e r i z a t i o n of C o a l ........................................ 1
1.1 Classification of Coal ................................................................................................ 1
1.2 Proximate Analysis and Elemental Analysis ............................................................. 3
1.2.1 Collection and Adjustment of Sample ............................................................... 3
1.2.2 Proximate Analysis ......................................... , ............................................. .... 3
1.2.3 Ultimate Analysis .............................................................................................. 4
1.2.4 Calorific Value (Heating Value) ....................................................................... 4
1.3 Solvent Extraction ..................................................................................................... 4
1.4 Various Analytical Methods ...................................................................................... 6
1.4.1 Nuclear Magnetic Resonance (NMR) ............................................................... 6
1'4.2 FTIR (Fourier Transform Infrared) Spectroscopy ............................................. 12
1.4.3 High-resolution Transmission Electron Microscope (HRTEM) ....................... 18
1.4.4 Characterization of Coal Aggregate Structure by X R D .................................... 20
1.4.5 Pyrolysis Gas Chromatography/Mass Spectrometry ......................................... 24
1.4.6 Ruthenium Ion-catalyzed Oxidation of Coal ..................................................... 31
1.5 Tritium Tracer Methods for Coal Characterization ................................................... 34
1.5.1 Hydrogen Exchange Reaction of Coal with Tritiated Water ............................ 36
1.5.2 Determination of Aromatic Hydrogen around Functional Groups of Coals
in Reaction of Coal with Tritiated Water .......................................................... 45
1.5.3 Catalytic Hydrogen Exchange Reaction of Coal with Tritiated Gaseous
Hydrogen ........................................................................................................... 49
1.5.4 Effects of Particle Size of Coal on Catalytic Hydrogen Exchange Reaction
of Coal with Tritiated Gaseous Hydrogen and Water ....................................... 56
1.5.5 Elucidation of Catalytic Hydrogen Exchange Rate in Coal under Reductive
Atmosphere ....................................................................................................... 65
1.5.6 Hydrogen Transfer between Coal and Tritiated Organic Solvent ..................... 72
vi Contents
In the 20th century, coal played an important role in the development of economics and in-
dustry not only in Japan but throughout the world. As raw materials for industrial produc-
tion such as fuel for the generation of electricity, cokes, tars, gases from coke oven, synthe-
sis gas from coal gasification and liquid fuels from liquefaction, coal has been used exten-
sively in various industries for electric power, iron, tar and gas production, the production
of synthetic chemicals and fuel, etc. Although the necessity for coal will change with
changing prices of other energy sources such as petroleum and natural gas, the importance
of coal as energy and carbon sources will not change in the future.
Coal is more abundant than petroleum and natural gas. Further, coal is not localized
but can be used by many more countries than petroleum. Therefore, if we can establish
coal utilization technology, coal will bring about a great contribution to human life and so-
ciety. On the other hand, shortage of petroleum and natural gas are anticipated in the sec-
ond half of the 21st century. To compensate, the use of coal is expected to gradually in-
crease during the 21st century. In the future, the development of the coal utilization tech-
nology will become more and more important to insure the supply of liquid fuels for trans-
portation and carbon sources for the manufacture of chemicals and plastic materials.
In order to develop such technologies, the elucidation of the structure of coal will be a
fundamental key study. Further, more efficient coal utilization technology must be estab-
lished to meet environmental legislation. One of the key technologies for this purpose is
catalysis. Given the situation, there is urgent need for a text book which covers both scien-
tific and practical sides of coal utilization with and without catalysts. This volume aims to
provide an English description of the basic and practical aspects of the science and technol-
ogy of coal utilization with and without catalysts. The actual structure of coal, the chem-
istry included in the reactivity of coal, the methods to elucidate the structure of coal and re-
action mechanisms of coal conversion, the most important catalyst for converting coal to
liquid and gas, the role of the catalysts in coal conversion, the problems in the process engi-
neering, and how to meet environmental regulations are discussed in detail.
The recent progress in studies on the structure and reactivity of coal made over the last
century is summarized and reviewed with emphasis on both fundamental and applied as-
pects of the science and technology for coal processing in the presence and absence of cata-
lysts. The book consists of six chapters: The first chapter describes the classification and
characterization methods of coal. The second chapter describes the chemical and macro-
molecular structure of coals. The third chapter describes the catalytic and noncatalytic py-
rolysis of coal, coal tar, and coal tar pitch. The fourth chapter shows the catalytic and non-
catalytic liquefaction of coal. While the fifth chapter introduces the catalytic and noncat-
x Preface
alytic gasification of coal. The final chapter details the microbial depolymerization of coal
which involves catalysis using live organisms.
We would like to express our gratitude to Ms. Cecilia M. Hamagami, and Mr. Ippei
Ohta of Kodansha Scientific Ltd. and Dr. Danhong Wang, a postdoctoral fellow at Tokyo
Univerisity of Agriculture and Technology, for their invaluable assistance in the prepara-
tion of the English manuscript.
April 2004
Toshiaki Kabe
Atsushi Ishihara
Eika Weihua Qian
I Putu Sutrisna
Yaeko Kabe
Methods of Classification and Characterization of Coal
higher heating value and lower moisture and volatile matter content. As bituminous coal
exhibits caking behavior, this coal is useful for making coke and is important for iron and
steel production. Anthracite is the highest rank of coal and is very low in volatile matter.
This coal does not form coke when heated. Argonne premium coal samples, which are typ-
ical standard coal samples, are classified using this method of classification. Their proxi-
mate and ultimate analyses are listed in Table 1.2 and 1.3, respectively.
To produce good coke, coals with different caking properties are blended. To choose
the appropriate coals for blending, coal petrography, by which the textural component of
coal is analyzed and classified by microscope, is available. Coal texture includes mi-
crolithotype and maceral. Microlithotype and maceral in Japanese coals are shown in Table
1.4 where Vitrite, Clarite, Durite and Fusite are included in the former, and Vitrinite,
Cutinite, Exinite, Degradinite, Inertinite, Semifusinite, Fusinite, and mineral matter are in-
Table 1.2 Proximate Analyses for the Argonne Premium Coal Samples a
Proximate analysis
aMoist -- moisture content, ash -- ash content, vol. mat. = volatile-matter content, F carbon = fixed carbon con-
tent, cal. val. -- calorific value, rec = analysis on an as-received basis from the mine, dry = analysis on a mois-
ture-free basis. [Reproduced with permission from Karl S. Vorres, Users Handbook for the Argonne Premium
Coal Sample Program, 11 (1989)]
1.2 ProximateAnalysis and Elemental Analysis 3
Cal.
Carbon H y d r o g e n Nitrogen Sulfur Chlorine Oxygen val.
dry maf dry maf dry maf dry maf dry maf dry maf maf
Coal (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) (kJ/kg)
Beulah-Zap(ligA) 65.85 72.94 4.36 4.83 1.04 1.15 0.80 0.70 0.04 0.04 18.19 20.34 28340
Wyodak(subC) 68.43 75.01 4.88 5.35 1.02 1.12 0.63 0.47 0.03 0.03 16.24 18.02 29871
Illinois#6(hvCb) 65.65 77.67 4.23 5.00 1.16 1.37 4.83 2.38 0.05 0.06 8.60 13.51 32887
Blind Canyon(hvBb) 76.89 80.69 5.49 5.76 1.50 1.57 0.62 0.37 0.03 0.03 10.76 11.58 33988
Lewiston-Stockton(hvAb) 66.20 82.58 4.21 5.25 1.25 1.56 0.71 0.65 0.10 0.12 7.69 9.83 34267
Pittsburgh(hvAb) 75.50 83.20 4.83 5.32 1.49 1.64 2.19 0.89 0.11 0.12 6.63 8.83 34930
Upper Freeport(mvb) 74.23 85.50 4.08 4.70 1.35 1.55 2.32 0.74 0.00 0.00 4.84 7.51 36076
Pochontas#3(lvb) 86.71 91.05 4.23 4.44 1.27 1.33 0.66 0.50 0.19 0.20 2.17 2.47 36695
aDry -- analysis on a moisture-free basis, maf = analysis on a moisture to ash-free basis. [Reproducedwith per-
mission from Karl S. Vorres, Users Handbook for the Argonne Premium Coal Sample Program, 11 (1989)]
cluded in the latter. Maceral contains different c o m p o n e n t s and structures even in the s a m e
kind of coal, a fact which often affects not only coking properties but also reactivities of
liquefaction and gasification.
Volatile Matter: The volatile matter content is obtained from the decrease in weight when a
sample in a platinum crucible is heated in an electrical furnace for 7 min at 925 __+20 ~
Fixed Carbon: The fixed carbon content is calculated by subtracting the sum of the mois-
ture, ash and volatile contents from 100.
(JIS M 8812)
Coal
I Pyridine extraction
I I
Residue Extract
Ichloroform
I /..I
Chloroform insoluble Chloroform soluble
I Petroleum ether
I I
Petroleum ether insoluble Petroleum ether soluble
IEthyl ether
I I
Ethyl ether insoluble Ethyl ether soluble
Acetone
I I
Acetone insoluble Acetone soluble
Fig. 1.1 Classification by Wheeler method. [Reproduced with permission from Kimura, H. et al., Chemistry and
Industry of Coal, 102, Sankyo Pub. (1977)]
close to the boiling point of the solvent, however, decomposition or oxidation may occur,
especially in solvents with higher boiling points. There are few studies of extraction at
room temperature.
A mixed solvent such as alcohol-benzene is often used to obtain better extraction yield.
The mechanism of the synergetic effect of the mixed solvents observed was explained by
the increase in the penetration of solvents into coals by coal swelling (Iino and Matsuda,
1984). Further, a CS2-pyridine mixed solvent gave high yields in the extraction of some bi-
tuminous coals at room temperature (Iino and Matsuda, 1983).
Recently Iino et al. found that a mixed solvent of CS2 and 1-methyl-2-pyrrolidinone
(NMP) is extremely effective for dissolving coal at room temperature (Iino et al., 1985,
1988, 1989). This system dissolved 40-70% of bituminous coal although the anthracite,
subbituminous coals and lignites did not give high yields. In this method, a coal sample (10
g) was extracted with 250 ml of CS2-NMP mixed solvent (1:1 by volume) under ultrasonic
irradiation (38 kHz) for 30 min at room temperature. After centrifuging at 14000 rpm for
60 min, the supernatant was separated by decantation. The addition of fresh mixed solvent
to the residue, ultrasonic irradiation and centrifuging were repeated until the supernatant be-
came almost colorless. After filtration and evaporation of CS2 and NMP below 90 ~ the
wet extract obtained was immediately fractionated into AS, PS and MS fractions, using ace-
tone and pyridine, respectively, according to the procedure shown in Fig. 1.2. The extrac-
tion yield was calculated based on Eq.(1.4):
[1-(residue(g)/coal feed (g))] x 100
Extraction yiels (wt % daf)= (1.4)
[1 -(ash (wt % d. b.)/100)]
From the results of the characterization of the raw coals, extracts and residues, it was sug-
gested that reactions between the coals and the solvents do not occur to a significant extent
during the extraction. The synergistic effect with the mixed solvent, CS2 and NMP, has
been explained by increasing solubility and diffusibility of the extracts and increasing
swelling of the coals. The mixed solvents of CS2 with quinoline, pyridine, and THF gave
6 1 Methods of Classification and Characterization of Coal
Coal
CS2-NMP
extraction
I
Extract Residue
Acetone
I
Acetone Acetone
soluble insoluble
(AS) (AI)
Pyridine
I I
Pardon Pyridine
soluble insoluble
(PS) (PI)
Fig. 1.2 Extraction and fractionation procedures. [Reproduced with permission from Iino, M. et al.,
Fuel, 68, 1989, Elsevier (1989)]
(1.5)
Total carbon C
Substitution H,~/x+O
o-= (1.6)
Aromatic site HMx+O+Har
where C represents the number of total carbon, Har the number of aromatic hydrogens, Ha
the number of hydrogens at a position, H e the number of hydrogens at fl and beyond the 13
position except the terminal methyl group, H r the number of hydrogens at the terminal
methyl position, and O the number of hydroxy groups, x is the atomic ratio of hydrogen to
carbon at a, t3 and beyond t3 positions and is usually assumed to be the value of 2.
Special attention has been focused on the recent development of 13C solid-state NMR
spectroscopy, which has enabled us to estimate the structure of coal without destruction. In
the solid-state NMR, broad peaks are obtained because of the chemical shift anisotropy
(CSA). To remove this CSA and to obtain a sharp peak, the magic angle spinning (MAS)
where the probe including a NMR sample is inclined by 54.44 ~ is also used (Schaefer and
Stejskal, 1976). The relaxation time for ~3C is very long compared with that of 1H. To
overcome this point, the cross polarization (CP) method is used (Pine et al., 1973). In this
method, ~H in a sample is initially magnetized and then 13C is magnetized by the transfer of
its magnetization with the CP method, which enables a high signal/noise (S/N) ratio and
short relaxation time. In recent reports, however, it is believed that the CP method does not
give quantitative information (Franz et al., 1992; Snape et al., 1989). Instead of the CP
method, much attention has been focused on the single pulse excitation (SPE) method, in
which 13C nuclear is directly magnetized. In this method, the pulse delay is takes a long
time to magnetize ~3C directly. Therefore, the disadvantages are a longer measurement time
and a poorer S/N ratio than the CP method.
The example of SPE/MAS 13C-NMR spectrum of a bituminous Pittsburgh #8 coal is
shown in Fig. 1.3 (Murata et al., 2001; Kidena et al., 1999). The NMR measurement condi-
tions are as follows: Chemagnetics CMX-300 spectrometer With higher magnetic field
(7.1T), pulse width 1.5 ~ts (45~ pulse delay 100s, MAS frequency 10.5 kHz, proton decou-
pling 83 kHz, and scan number 1000-2000. The spectrum shown in Fig. 1.3 could be di-
vided into twelve Lorenzian curves. Assignments, chemical shift, and half width of each
peak are summarized in Table 1.5. Using these parameters, the curve fitting treatment was
performe d and the results of carbon distribution, which is calculated from the ratio of area
of each peak are summarized in Table 1.6. Aromatic carbons were classified into four
classes, O-beating aromatic carbons (Car-o), alkylated aromatic carbons (Car-R), tertiary aro-
matic carbons (Car_H), and bridgehead carbons (Can). Tertiary aromatic carbons and bridge-
head carbons were difficult to distinguish because their chemical shifts are very similar to
each other. To estimate the parameter concerning average aromatic cluster size in coal, the
bridgehead carbon must be determined. To overcome this problem, Pugmire and co-work-
I I I I I I I
ppm 250 200 150 100 50 0 -- 50
Fig. 1.3 SPE/MAS ~3C-NMR spectrum of Pittsburgh # 8 coal. * -- SSB. [Reproduced with permission
from Kidena, K. et al., J. Jpn. Inst. Energy, 78, 870 (1999)]
Carbon types PC UF PT ST BC IL WY BZ
ers used dipolar dephasing (DD) pulse sequence (Solum et al., 1989). Instead of this
method, Nomura and co-workers (Kidena et al., 1999) determined the fraction of aromatic
hydrogen to total hydrogen (Har). They used the following equations for aromatic hydrocar-
bons.
C a r - Car-o -]-- Car_R .71_Car-H -~- CBH (1.8)
where C and H represent the contents of carbon and hydrogen in coal, Car represents the
fraction of aromatic carbon to total carbon, Har represents the fraction of aromatic hydrogen
to total hydrogen and Zb represents the mole fraction of bridgehead carbon. C and H can
be calculated from the elemental analysis and Car can be calculated from the solid-state 13C-
NMR data. To calculate Har, the solid-state 1H-NMR data of coal were used.
Nomura and coworkers developed CRAMPS (Combined Rotation And Multiple Pulse
Spectroscopy) to avoid the broad spectra of 1H-NMR because of 1H-1H dipolar interactions.
The solid-state 1H-NMR spectra were measured under the condition, 3.5 kHz of offset and
1.5 kHz of MAS. The results are shown in Fig. 1.4. The fight peak is assigned to aliphatic
POC
15' 1'0 5 O' --5Ippm 1'5 1'0 5' 0--5ppm'
Pi
i i i i i i i i i i
B Illinois #6
i i l i i i i i i
Wyodak
,r
l I l l
1'5 10 ; 6--51ppm 15 10 ; 0--;ppm
Fig. 1.4 1H-CRAMPS(BR24) spectra of the sample coals. [Reproducedwith permission from Kidena. K. et al.,
J. Jpn. Inst. Energy, 78, 873 (1999)]
10 1 Methods of Classification and Characterization of Coal
Table 1.7 Hydrogen Aromaticity (H~r), Fraction of Bridgehead Carbon and Protonated Aromatic Carbon
of the Sample Coals
Carbon distribution (%)
Coal Har Bridgehead carbon Protonatedaromatic carbon
Pocahontas #3 0.40 39.5 23.2
Upper Freeport 0.36 32.4 23.7
Pittsburgh #8 0.29 27.6 22.3
Stockton 0.36 25.8 26.9
Blind Canyon 0.24 20.7 20.8
Illinois #6 0.27 20.8 21.1
Wyodak 0.33 13.8 28.4
Beulah-Zap 0.37 13.8 29.6
[Reproduced with permission from Kidena, K. et al., J. Jpn. Inst. Energy, 78, 869 (1999)]
hydrogen and the left peak to aromatic hydrogen. F r o m this data, nar was determined and
Car-H and CBH could be calculated as shown in Table 1.7.
P u g m i r e and c o w o r k e r s h a v e found the relationship b e t w e e n the mole fraction of
bridgehead carbons (Zb) and the n u m b e r of carbon atoms per aromatic cluster (C) (Solum et
al., 1989). As shown in Fig. 1.5, there are two limiting cases of polyaromatic hydrocarbon
condensation: (a) primary catenation and (b) circular catenation. The structures in the pri-
mary catenation and the circular catenation belong to the C4n + 2H2n + 4 and C6n2H6n families
of polycondensed aromatic hydrocarbons (PAH), respectively. If Zp is the mole fraction of
carbon at a peripheral position (e.g., a C-H or a C-O carbon), the mole fraction of bridge-
head carbons Zb can be expressed as Eq. (1.12).
~b -- 1 -- ~p (1.12)
Zp = H/C (1.13)
0.8 -
0.7 -
0.6-
0.5
Zb 0.4
0.3
0.2 -
0.1 -
0 10 20 30 40 50 60
Carbons per cluster
Fig. 1.6 Plot of the mole fraction of bridgehead carbons, Zb, vs. C where C is the number of carbon atoms per aro-
matic cluster. The solid curve is for the combined model, the upper dashed curve is for the circular cate-
nation model, and the lower dashed curve is for the primary catenation model.
[Reproduced with permission from Solum, M.S. et al., Energy Fuels, 3, 191 (1989)]
n=(C--2)/4 (1.15)
where C is the number of carbon atoms in the aromatic cluster. The linear Zb" is derived
from Eqs. (1.12)-(1.15).
The value for Zb" of circular catenation is derived from Eqs. (1.12), (1.13), (1.17), and
(1.18). Zb,,=I_~/~I~/C (1.19)
When Zb" is plotted against C, the upper dashed curve is drawn as in Fig. 1.6.
When C is less than 14 carbons, Zb" governs the relationship for Xb. However, at higher
n u m b e r s o f carbon, Zb" best expresses Zb. Pugmire and coworkers used the hyperbolic tan-
gent function, Eq. (1.20), to transfer the dependence between the two limiting functions Xb"
and Xb".
1-- tanh ( ( C - 19.57) / 4.15) 1 + tanh ( ( C - 19.57) / 4.15) ,,
Zb: 2 Zb t-'~- 2 ~b (1.20)
Using this equation, the plot of Zb vs. C is expressed as the solid curve in Fig. 1.6.
12 1 Methods of Classification and Characterization of Coal
Nomura and coworkers correlated Zb and C with carbon content as shown in Fig. 1.7
where Zb can be obtained from their SPE/MAS 13C-NMR spectrum data, and C was calcu-
lated from their Zb using Eq. (1.20). When this correlation between Zb and carbon content
was compared with Zb in Pugmire's report, the values of Zb for Nomura's group were higher
in higher rank coal. It was concluded that the Zb values estimated by Pugmire's group
might be lower because the CP method used at the same time has poorer accuracy in the de-
termination of quarternary carbon (Kidena et al., 1999).
@
PC
0.5 (a) UF .Q) (b)
PT ~ ' " UF"e
0.4 B2
~t:4-)
~'~-S
L 22 e'"
O'PT
,, 0.3 IL C,]ViK
~ ( ~O HV 18 BC ti LS
9
Fig. 1.7 2'b and C of the sample coals.
PC: Pocahontas # 3, UF: Upper Freeport, PT: Pittsburgh # 8, HV: Hunter Valley,
ST: Stockton. BC: Blind Canyon, MK: Miike, TH: Taiheiyo, IL: Illinois #6,
WY: Wyodak, BZ: Beulah Zap, SB: South Banko, YL: Yalloum.
[Reproduced with permission from ed. Iino, M.; Murata, S. et al., Primary and Higher Order Structures
of Coal and Their Influence on Coal Reactivity-Final Report on "Research for the Future" Coal
Research Project-, 4 (2001)]
1.4.2 F T I R ( F o u r i e r T r a n s f o r m Infrared) S p e c t r o s c o p y
A. Estimation of Coal Structure by FTIR
FTIR spectroscopy is one of most useful methods for directly analyzing the functional
groups in coal and coal-related compounds (Painter et al., 1982, 1985, 1987; Solomon,
1981, Solomon et al., 1982, 1987, 1990; Solomon and Carangelo, 1982, 1988; Gaines,
1988; Martin and Chao, 1988; Smyrl and Fuller, 1982; Fuller and Smyrl, 1984; Snyder et
al., 1983; Mu and Malhotra, 1991; Sobkobiak and Painter, 1995; Chen et al., 1998; Miura
et al., 2001). This method enables the nondestructive analysis of coal, especially the deter-
mination of functional groups.
Studies of Argonne coal samples by FTIR were carried out by Solomon and coworkers
(Solomon and Carangelo, 1982, 1988; Solomon et al., 1982). The results from FTIR func-
tion group analysis on the Argonne Premium Coals are shown in Table 1.8. The IR spectra
were measured using KBr pellets. Absorption peaks at 3400, 3030, 2900, 1700 and 1200
cm -1 are assigned to a hydroxy group, aromatic hydrogen, aliphatic hydrogen, carbonyl
and ether, respectively. The carbonyl content, which includes undeterminable factors, was
estimated by the relative peak area.
Table 1.8 FTIR Functional Group Analysis on the Argonne Premium Coals (wt % dmmf) a
1. Upper Freeport (mvb) 3.43 0.11 2.08 5.62 0.37 0.66 0.71 0.71 22.87 0.63 1.75 0.75
2. Wyodak (subC) 3.03 0.33 1.73 5.09 0.34 0.52 0.78 0.43 20.20 23.86 5.25 5.0
3. Illinois #6 (hvCb) 3.41 0.23 2.07 5.71 0.36 0.69 0.78 0.60 22.73 4.48 3.75 2.25
4. Pittsburgh (hvAb) 3.60 0.16 2.07 5.83 0.36 0.67 0.80 0.60 24.00 0.86 2.5 1.88
5. Pocahontas #3 (lvb) 1.97 0.06 2.19 4.22 0.52 0.60 0.73 0.86 13.93 1.92 1.0 1.25
6. Blind Canyon (hvBb) 4.79 0.16 1.90 6.85 0.28 0.51 0.80 0.58 31.93 8.70 2.5 4.0
7. Lewiston-Stockton (mvb) 3.48 0.23 2.12 5.83 0.36 0.67 0.67 0.79 23.20 3.59 3.75 1.75
8. Beulah-Zap (ligA) 2.02 0.34 1.58 3.94 0.40 0.46 0.74 0.37 13.47 24.67 5.5 5.0
a Except carbonyl : relative peak area.
bHal = wt % hydrogen as aliphatic hydrogen, Hon = wt % hydrogen as hydroxy hydrogen, Har = wt % hydrogen attached to aromatic groups. Cal = wt % carbon in
aliphatic groups, Oon = wt % oxygen in hydroxy groups, and Oether - " w t % oxygen in ether groups.
c 1 Adj = one adjacent hydrogen which is attached to an aromatic carbon: 2 Adj = two adjacent hydrogens attached to an aromatic carbon.
a Peak height at 1700 c m - 1(arbitrary units).
[Reproduced with permission from Solomon, P.R. et al., Energy Fuels, 4, 52 (1990)]
Table 1.9 Data on Coal and Chars Produced at 0.5 ~ Heating Rate, 3-min Hold Time (wt % dmmf)
coal 1.93 0.43 1.50 3.86 0.39 12.89 17 6.83 6.69 2.38
200 ~ char 1.88 0.38 1.74 4.00 0.43 12.55 17 6.02 5.63 0.3 2.38
300 ~ char 1.77 0.41 1.36 3.54 0.38 11.80 18 6.50 7.25 3.0 1.77
400 ~ char 1.67 0.37 1.62 3.66 0.44 11.10 17 6.00 > 9.0 9.8 ND b
500 ~ char 0.78 0.22 2.28 3.28 0.69 5.20 6 3.50 6.75 22.1 1.07
600 ~ char 0.32 0.19 3.36 3.87 0.87 2.10 6 3.00 4.38 30.0 1.00
700 ~ char 0.11 0.11 3.70 3.92 0.94 0.73 1 1.75 5.25 35.4 ND b
800 ~ char 0.00 0.08 3.01 3.09 0.97 0.00 0 1.25 7.00 43.0 ND b
Existence of the bands was determined by a curve resolving method based on the sec-
ond derivative of the spectrum and eye and brain examination. They also clarified that the
frequency shifts from the free OH and the intensities of the bands provide fundamental pa-
rameters that can be used to determine structural characteristics such as bond lengths and
the enthalpy of bond formation. However, no definite methods have been presented to esti-
mate the strength distribution of hydrogen bonds probably because of the difficulty in ana-
lyzing the OH stretch frequency bands associated with water bound to coal in a complex
manner and the presence of more than one band in the region of the spectrum. If the KBr
pellet method was employed, water adsorption by the KBr matrix must also be taken into
account.
Apart from coal hydrogen bonds, a number of works on various hydrogen bonds appear
in various reviews and several books (Hadzi and Thompson, 1959; Vinogradov and Linnell,
1971; Schuster et al., 1976). Hydrogen-bonded adduct formation reactions between the OH
functional groups of phenols and/or alcohols and various bases have been investigated in
detail in liquid phase using FTIR and thermodynamic considerations. The A H (the enthalpy
change) values for various combinations of phenols and bases were estimated by applying
the van't Hoff equation to the equilibrium concentrations. When the hydrogen bond is
formed, the IR wavenumber of the O-H stretching vibration of the phenol shifts to a low
wavenumber and generates heat (AH < 0). The relationship between the enthalpy change,
AH, and the OH wavenumber shift, A VoH, were examined for various phenol-base and alco-
hol-base combinations (Amett et al., 1970; Joesten and Drago, 1962; Purcell and Drago,
1967; Epley and Drago, 1967; Drago and Epley, 1969; Amett et al., 1974), and a linear rela-
tion was found to hold between A H and A VoH by many investigators (Drago and Epley,
1969).
Miura and co-workers have attempted to develop a method to quantify the hydrogen
bonds in coal based on the findings reviewed above (Miura et al., 1997, 1999, 2002a). The
hydrogen bond formation reaction of OH can be defined as follows:
C o a l - OH § B---~ C o a l - OH ... B ; AH (1.21)
where Coal-OH represents a hypothetical state in which the OH exists as a free OH group.
The bases, B, in this case represent basic reagents that have electron donors such as O and
N in OH, COOH, C=O and C - N=C groups.
The amounts of individual hydrogen bonds, i.e., the amounts of OH contributing to dif-
ferent hydrogen bonds, were determined by the curve resolving method. First, the absorp-
tion band ranging from 2400 to 3750 cm-1 was regarded to consist of the following 10 ab-
sorption bands, 7 OH stretch bands and 3 CH stretch bands.
3640 cm-1; free OH groups
3530 cm-1; OH-re hydrogen bonds
3400 cm-1; self-associated n-mers (n ~ 3)
3280 cm-~; OH-ether O hydrogen bonds
3150 cm-~; tightly bound cyclic OH tetramers
16 1 Methodsof Classification and Characterization of Coal
The peak positions are slightly different from those proposed by Painter et al. (1987),
and the hydrogen bonds attributable to COOH dimers were also taken into account judging
from the second derivative of the spectrum and brain and eye examination. Assuming the
Gaussian distribution for each absorption band, the spectrum was curve resolved to obtain
the intensity of each band (Solomon et al., 1982).
Next, the absorptivity (extinction coefficient) of each band must be determined to con-
vert the individual intensities to the corresponding amounts of OH groups. According to
the following relationship between the absorptivity of the hydrogen bonded NH, aNH, and
that of free NH, aNH.0found by Detoni et al. (1970)"
a N H - - aNH,0 ( 1 + 0.0141 A VNH) (1.22)
where AVNHis the NH wavenumber shift. This relationship shows that the absorption inten-
sity of the hydrogen bonded NH increases in proportion to A VNH. Based on the report for
the absorptivity of NH stretching band by Detoni et al., they arbitrarily assumed a similar
relationship for the change in absorption intensity of hydrogen bonded OH, and estimated
the proportional constant from the intensity and the AVon values measured for several mod-
el compounds (benzoic acid and 2-naphtol) as
aOH -- aon,0 ( 1 d- 0.0141 AYon) (1.23)
where aon is the absorptivity of the hydrogen bonded OH and aon.0 is the absorptivity of
the free OH. They were able to calculate the amounts of OH contributing to different hy-
drogen bonds using Eq.(1.23) when they can estimate aoH,0.
Larsen and Basker suggested that the coal hydrogen bonds were treated in a similar
way as the phenol-base hydrogen bonds (Larsen and Basker, 1987). Using Eq.(1.24) ob-
tained by Drago and Epley (1969), they estimated the strength of individual hydrogen
bonds in coal from the A v o n values.
-- AH -- 0.067 (A Yon) + 2.64 (kJ/mol) (1.24)
Figure 1.8 shows the i n s i t u FTIR spectra ranging from 2200 to 3700 c m -1 measured
for eight coals. Very beautiful and reproducible spectra could be obtained by use of the
proposed measurement technique for all the coals. Very flat base lines were obtained at all
temperatures, and the spectra for POC were almost the same above 70 ~ suggesting that
the spectra were not affected by temperature below 300 ~
All the spectra measured for the coals were analyzed by the procedure described above
to estimate the change in the strength distribution of hydrogen bonds (HBD) through heat-
ing. Fig. 1.9 shows the result of peak division and the HBD obtained for ND as an example.
The spectra ranging from 2200 to 3650 c m - ~ were divided into nine peaks by a curve-fit-
ting method. The six types of hydrogen bonds were abbreviated as follows: HB l" OH-z at
3530 cm -1, HB2" OH-OH at 3400 cm -1, HB3" OH-ether at 3280 cm -1, HB4" cyclic OH-
OH at 3150 cm -1 HBS: OH-N at 2940 cm -~ and HB6" COOH-COOH at 2650 cm -~
, 9
These are in the order of increasing strength. No free OH bands were detected for any of
the coals. The strength distributions of hydrogen bonds, the amount of OH corresponding
to the j-th peak, (non)j VS. the strength of the j-th hydrogen bond, (-AH)j relationships, for
1.4 Various Analytical Methods 17
, I
i
30 ~ IL
<
, i i , , i i
UT o ~. PITT
,.Q
<
_
v- -
i , i , i , i i i i i i i ,
,~ UF POC
~= 270 ~ 30-90 ~
ND are shown in Fig. 1.9. The strengths of individual hydrogen bonds represented by - A H
values are HBI: 10kJ/mol-OH, HB2: 17kJ/mol-OH, HB3: 25kJ/mol-OH, HB4: 35kJ/mol-
OH, HB5: 50kJ/mol-OH, HB6: 70kJ/mol-OH. Since it is assumed that all adsorbed water
was desorbed by 150 ~ the distribution at 150 ~ is the distribution of the coal dried and
heated to 150 ~ The dried ND coal is rich in rather weak hydrogen bonds: 2.5 mol-
OHNg of HB1, 1.3 mol-OHNg of HB2, and 1.1 mol-OHNg of HB3. When the raw ND
coal was heated from 30 ~ to 150 ~ the amounts of weaker hydrogen bonds such as
HB 1, HB2, HB3 and HB4 decreased significantly. Since this decrease is due to the desorp-
tion of adsorbed water, it is found that water is interacting with coal hydroxys in rather
weak interactions. When the coal was further heated to 270 ~ all hydrogen bonds de-
creased. This decrease is judged to be due to the decomposition reactions of carboxylic
groups. Thus it was clarified that the in situ DRIFT technique and the analysis method pro-
posed by Miura and co-workers were powerful methods for examining the change in hydro-
gen bonds in coal.
18 1 Methods of Classification and Characterization of Coal
DuB (kJ/mol-OH)
0 50 100 150 200
i | i | i i
0| , , , --
I , I , I t
150 ~
~~ 46 f 150~
lit t
r~
<
0 a , I i
, ! , ! ! I I I I
270 ~ 6 270 ~
~ I
3600 3200 2800 2400
o0 I,20 l l , I40 ,I
60 80
W a v e n u m b e r (cm -~) --AH (kJ/mol)
Fig. 1.9 Six hydrogen bonded OH stretching bands and three CH bands obtained by a curve resolving method and
the strength distributions of hydrogen bonds estimated by the proposed method for ND coal.
[Reproduced with permission from Miura, K. et al., Energy Fuels, 15, 607 (2001)]
HRTEM images of BZ, IL, and POC coals are shown in Fig. 1.10 (a-c), respectively.
Nearly 5 to 6 different spots were observed for each coal sample. The lattice fringes can be
seen very clearly and individual layers are easily discernible in the images. This finding is
especially important since HRTEM technique can now be used to observe the coal structure
directly. The previous direct observation studies failed to observe any lattice fringes. One
explanation given for this failure is that coal is an electron beam-sensitive material and its
structure could be altered during observation in high-resolution lattice fringe mode where
the area exposed to beam is very small. If irradiation damage is the reason for not observ-
ing the lattice fringes, then perhaps by working with a very low intensity beam, irradiation
damage can be avoided or at lest retarded. They used a low intensity beam, and observed
fringes. The irradiation damage was indeed one of the reasons for not being able to observe
the fringes in raw coals. However, simply reducing the intensity beam is not sufficient to
avoid irradiation damage. Reducing the beam intensity can only delay the irradiation dam-
age.
Fig. 1.10 HRTEMimages of (a) Beulah-Zap coal, (b) Illinois #6 coal, and (c) Pocahontas# 3 coal.
[Reproduced with permission from Sharma, A. et al., Energy Fuels, 14, 516 (2000)]
Therefore, in order to have as much time as possible to observe the fringes, they first
observed the samples at high magnification and then at low magnification as the low mag-
nification image is only of morphological interest. This approach gives more time to ob-
serve and image the fringes before irradiation starts damaging the fringe structure. The ap-
proach of first observing the sample at low magnification and then at high magnification
may not be appropriate for raw coal observation. The problem of not being able to focus
the images on the fluorescent screen due to a very low level of illumination was overcome
by using a highly sensitive digital camera combined with a TV system. By this approach
the images can be focused by observing the fringe structure on the TV screen. In some ob-
servations, a Cu grid was used without any polymer support to rule out any possibility of
observing anything else other than coal. Therefore, the fringe structure presented here is
due to coal only. In BZ, IL and POC coals, most of the layers are curved and poorly orien-
tated. Some layers can be seen as forming stacks. It is known that coals contain small con-
densed aromatic units that tend to stack parallel to each other, but the orientation and pla-
narity are imperfect because of the presence of heteroatoms and hydroaromatic portions
(Millward, 1979). The HRTEM images of the raw coals presented in this study also show
that the fringes or layers in coals are poorly orientated and most of these are nonplanar (Fig.
1.10 (a-c)). This observation is in accordance with the above understanding (Millward,
1979). Moreover, the stacking number and layer size increase with coal rank. From XRD
measurements, Hirsch (1954) reported that layer size increases sligtly until 90% carbon
20 1 Methods of Classification and Characterization of Coal
content is reached. From a semiquantitative analysis (Sharma et al., 1999) of the three
coals, it was found that the layer length increases from 0.7 nm for BZ (C = 72%) to 0.9 nm
for POC (C = 91%), and the average of layers per stack increases from 2.8 for BZ to 3.0 for
POC coal, showing some difference among the three coals.
In this application of the HRTEM technique, a unique image of an onion-like structure
was observed in POC coal, as shown in Fig. 1.11. This is direct evidence of the presence of
a fullerene-like structure in a coal. Such a structure resembles the TEM images of spher-
oidal carbon particles published by Iijima (1980). Kroto and Mckay (1988) studied the for-
mation mechanism of these structures and suggested that if such structures are soot-like,
they would be abundant in the earth's early atmosphere. This prediction is in good agree-
ment with the finding of such structures in coal by HRTEM.
Fig. 1.11 Onion-like structure in Pocahontas # 3 coal. [Reproduced with permission from Sharma, A. et al.,
Energy Fuels, 14, 516 (2000)]
"-,~u~,-" \[ . . . . ~ Hongei
ul
m . 1. . _ ^ ontas
C
(wt%, daf) ~ ~ J ~1
% ~~ K-~, Upper Freeport
H..~__, ~ ;:~iUah-zap
I I I I I I I
BZ BZ
520 ~
460 ~
400 ~
370 ~ 400 ~
350 ~
350 ~ 300 ~
300 ~ 250 ~
Original 200 ~
Original
10 20 30 40 50 60 10 20 30 40 50 60
2 0 (deg) 2 0 (deg)
4000 4000
WY IL
3000 100 ~ ~ 3000 1O0 ~
= 2000 2000
.,-~
i
400 ~ 400 ~
350 ~ ~ 350 ~
1000 300 oC 1000 300 ~
250 ~ 250 ~
Original 0 Original
10 20 30 40 50 60 10 20 30 40 50 60
2 0 (deg) 2 0 (deg)
Fig. 1.13 XRD profiles of low-rank coals before and after the heat-treatment.
[Reproduced with permission from Watanabe, I. et al., Energy Fuels, 16, 19 (2002)]
100 ~ 170 ~
3000
' ' UF UF
2000 300 ~
400 ~ 250 ~
1000 350 ~
300 ~
250 ~
Original Original
ll0 210 310 410 53 60 10 10 310 40
2 0 (deg) 2 0 (deg)
3000 , ,
1 in
"~ 2000
1000
400 ~ C
~ 350~
~ 300 ~
250 ~
01 i J , , ~ Original
10 20 30 40 50 60
2 0 (deg)
Fig. 1.14 XRD profiles of high-rank coals before and after the heat-treatment.
[Reproduced with permission from Watanabe, I. et al., Energy Fuels, 16, 20 (2002)]
I 180 ~ ]
I26
1400 . / ~ ~ . . .. . ,400
1200 9 i~o t
1000 ........................ I I ~ ~300
800 ..... 400 ~
\ ~ , . ~ 3 5 0 ~ .~ 200
= 600 \ ~ ~ 3 0 0 ~
400 \ /~------~~~ 250~ fi 100
2O0 ~_~ ~ ~oo oc -
0 . . . . . Kaw 0
10 20 30 40 50 60 15 20 25 30
2 0 (deg) ] 2 0 (deg)
80 ~
I I I26
1400 ~ . ~ 400
I~o 4
1200
1000 ........................p , . ~ 300 I
800 ........ 400
= 600 / \ \ ~ 370 ~~ ~ 200
'~
400 \ x...../~~~ 350 ~ ~ 100
200 \ / ~ 300 ~
0 -- Raw
10 2'0 3'0 4'0 5'0 6'0 0 15 20 25 30
2 0 (deg) 2 0 (deg)
Fig. 1.15 Smoothing and decombolution of XRD of heat-treated BZ coal.
[Reproduced with permission from Watanabe, I. et al., Energy Fuels, 16, 21 (2002)]
1.4 Various Analytical Methods 23
Figures 1.15 and 1.16 illustrate the smoothing and decombolution treatment procedures
of XRD of BZ and UF coal, respectively. The XRD patterns of the coals were smoothed by
the removal of the peaks derived from mineral matter, and the residual peak was divided
into two portions: the one peak at 20 ~ (called the ?'-band), the other peak at 26 ~ (called the
to-band). The divided XRD profiles of the coals are shown in the right side of the figures at
the two different heating rates of 80 and 180 ~ This smoothing and decombolution
treatment of XRD profiles well reflect the differences in the two coals and the heat-treat-
ment conditions. In the case of BZ coal, the peak intensity at 20 ~ increased with heating
temperature, while the peak at 26 ~ decreased with heating temperature in the temperature
I 180 ~ ]
2000 I26
~ ~ . 800 , ,
~'~ 1500 . . . . . . . . . . . . . . ~ ~ 600
400 ~
.,.~ 1000 > " - k - " ~ ~ , \ \ k . _ ~ 350 ~ ~ 400 Ii~
..------~ \\,v-7----~ 300 oc ~-
0
~ \k,Z~ 250 ~ E 200
~= 500 -x...._...~~ 200 ~
Raw 0 ,
0 15 20 25 30
10 20 30 40
2 0 (deg)
2 0 (deg) [ 80 ~
I
3 ? 3
~ 2.5 o 2.5
.g 2 ~ 2
1.5 1.5
Orig. 200 250 300 350 400 Orig. 200 250 300 350 400
Holding temperature (~ Holding temperature (~
range of 250 to 400 ~ indicating that the loosening of noncovalent bonds may contribute
to the disordering of the aggregate structure of the lower rank coal. In contrast, for UF
coal, the peak at 20 ~ was weakened while the peak at 26 ~ was intensified by the heat-treat-
ment, especially around 400 ~ suggesting that the change in the aggregate structure of UF
coal may be directed to the more stable form through the partial fusibility and the initial
coking reactions.
Figure 1.17 shows a comparison of the relative intensity of 126/12o of the two coals cal-
culated based on the treated XRD profiles as illustrated in Fig. 1.15 and 1.16. The relative
intensity of UF coal started to increase around 300 ~ from ca. 2.5 to ca. 3.5, while the
126/12o value of BZ coal did not change so much in this temperature range. This shows that
the change in the aggregate structure of BZ coal was not reflected in this semi-quantitative
XRD evaluation, although the partial fusibility and initial coking reaction of UF coal were
successfully evaluated. It can be said that the XRD profiles of the coals essentially repre-
sent the ordering extent of the aromatic plane stacking and the coking reactivity, naturally
leading to better sensitivity to the higher ranking coals.
1.4.5 Pyrolysis Gas Chromatography/Mass Spectrometry
Analysis of coal and their constitutent maceral fractions by FTIR and solid-state 13C-NMR
provides overall information concerning the aliphatic and aromatic moieties present and
their diagenetic changes. A more detailed molecular characterization, however, cannot be
achieved by these spectroscopic techniques. Analytical pyrolysis techniques such as pyrol-
ysis gas chromatography/mass spectrometry (py-GC/MS) or pyrolysis mass spectrometry
(py-MS), along with multivariate data analysis techniques, are powerful analytical tools for
simultaneously investigating coal structure and reactivity of organic materials (Meuzelaar
et al., 1984c). During pyrolysis, complex mixtures are released from coal due to distilla-
tion, desorption and thermal degradation. The amount and chemical nature of these prod-
ucts depend on both the chemical composition and structure of the coal and on the heating
conditions. Consequently, the results of a chemical analysis of these devolatilized coal
products provide information on the original structure of the coal. The mass spectrometric
technique in coal characterization was reviewed previously in Smith and Smoot (1990).
Most widely accepted structure models of coal are derived primarily from structures re-
lated to lignified plant tissue or vitrinite, but recent mass spectrometric studies indicate that
contributions from other coal precursors such as algae, plant cuticles and resins may have
been systematically underestimated in the literature (Meuzelaar et al., 1992). Py-MS tech-
niques have the advantages of minimal sample preparation techniques, high sensitivity,
specificity, analysis speed and good compatibility with computerized data reduction meth-
ods. The patterns in py-MS contain information concerning rank, depositional environ-
ment, maceral content and reactivity. Multivariate analysis techniques have correlated py-
MS data to explain up to 80% of the total variance described by as many as 25 coal parame-
ters obtained by conventional coal characterization techniques, including ultimate and prox-
imate analysis, petrology and mineral analysis (Meuzelaar et al., 1984a; Metcalf et al.,
1987). Consequently, py-MS data has been used to model many coal processes (Meuzelaar
et al., 1987).
chemical ionization (CI) and field ionization (FI)), (3) mass spectrometer type (e.g., inte-
grated, time-resolved and tandem mass spectrometry), (4) MS operating modes (e.g.,
quadrupole, ion trap detector and magnetic sector instruments) and (5) chemometric data
analysis techniques (e.g., factor analysis, discriminant analysis and canonical correlation
analysis (Meuzelaar et al., 1984c). An overview of the development of py-MS techniques
can be found in Meuzelaar et al. (1982).
The pyrolysis mass spectrometry method most commonly used to study coal samples is
based on the filament Curie-point pyrolysis principle in which a thin ferromagnetic probe
coated with microgram quantities of the coal sample is inductively heated to its well-de-
fined Curie-point temperature or to the transition temperature between the ferromagnetic
and paramagnetic states. The samples are pyrolyzed directly in front of the electron ioniza-
tion source of a quadrupole mass spectrometer. Low-voltage electron ionization (LVEI) is
often used to limit the fragmentation of the ions. The wire may heat up to its Curie-point
temperature in a time as short as 100 ms or as long as 5 s, depending on the strength of the
field. Typical heating rates vary between 102 and 104 K/s. The final temperatures of the
wire will stabilize to within a few degrees of the Curie-point temperature 631 K (358 ~
for Ni, 1043 K (770 ~ for Fe and 1421 K (1128 ~ for Co and intermediate values for al-
loys. Most analytical pyrolysis experiments are performed under vacuum py-MS condi-
tions to minimize secondary reactions and mass transfer effects. The mass spectrometer is
repetitively scanned over the desired mass range at high speed and the resulting spectra are
either summed in integrated mode or sequentially recorded in time-resolved mode
(Meuzelaar et al., 1984c). However, due to residual E1 fragmentation and the discrimina-
tion effects of higher-mass ions in quadrupole mass spectrometers used in Curie-point py-
MS, a mass range of only up to m/z 400 is usually detected. For a detailed description of
this technique, see Meuzelaar et al. (1982) and Snyder et al. (1987).
Two disadvantages of the Curie-point mass spectrometer analysis of pyrolysis products
are (a) some of the mass peaks are due to fragment ions and (b) the mass range covers only
the lower part of the overall molecular-weight range of the coal pyrolysis precuts. These
problems can be overcome through use of direct probe py-FIMS (Haider and Schulten
1985; Schulten et al., 1987, 1989a, b). Field ionization is a soft ionization mode which
causes minimal ion fragmentation producing almost exclusively molecular ions in a mass
range up to m/z 1000 (Schulten et al., 1989a). Py-FIMS then provides molecular-weight
distribution information of coal derived tars over a broader mass range. Hence, FI mass
spectra contain a large amount of information about the structure of coal. Most py-FIMS
studies on coal have been performed by Schulten et al. (1987, 1988, 1989a). In py-FIMS,
small samples (100-400/.tm) are placed in quartz crucibles and heated quasi-linearly from
320 to 1070 K (50-800 ~ with very low heating rates programmed from 0.2 to 10K/s
with pyrolysis products analyzed in the mass range from m/z 70 to 2100 m/z (Schulten et
al., 1987).
Figure 1.18 shows the changes in the number-average molecular weight (Mn) and tem-
perature-total ion intensity (TII) profile of Beulah-Zap lignite by time-resolved py-FIMS.
Three different stage of devolatilization can be observed. The first step is shown by the
maximum of 480 daltons at 510 K and is interpreted to indicate the desorption and
volatilization of coal constituents (Simmleit et al., 1992). The second stage is found in the
decrease to a minimum of 180 daltons at 690 K, indicating the thermal degradation of the
lignite macromolecular network. At temperatures over 730 K, the average molecular
weight increases to 280 daltons showing higher-molecular-weight pyrolysis products re-
leased from the sample. Similar patterns are observed for other coals.
26 1 Methods of Classification and Characterization of Coal
500 500
400 400
300 = 300
0
b
100 100
0 0
270 32/0 42/0 5+0 670 7+0 8+0 270 370 470 570 670 770 870
Temperature (K) Temperature (K)
Fig. 1.18 (a) Variation of number-average molecular weight of pyrolyzates during heating of Beulah-Zap lignite.
(b) Temperature-total ion intensity (TII) profile of Beulah-Zap lignite. Heating rate 100 K/min.
[Reproduced with permission from Simmleit, N. et al., Advances in Coal Spectroscopy, 306, Plenum
Press (1992)]
Ii
40 II
i i
i i
i i C
i i
35 i
i
i
m/z 110
b i i
i /
' m./z 424 i
/
30
a
o 25
m/z 544
: 20 [~
/
.,..~
r.~
.= 15 d
~= ./JFli , mZ 94
10
- / 90
5
i
-----// .... , . . . . . . . . ,. -,
100 200 300 400 500 600
Tamperature (~
Fig. 1.19 Field ionization signal thermograms for deconvoluted coal components of Beulah-Zap lignite showing
the devolatilization of components a-e during heating. Heating rate 100 k/min.
[Reproduced with permission from Simmleit, N. et al., Advances in Coal Spectroscopy, 314, Plenum
Press (1992)]
(a) Lignite
~ Dihydroxybenzenes
""-" or methoxyphenols
n
0 , ~" , , ,
ti,., , , , , , ,
,,,
, , , ,
..................
, , | ,
,
/I I\ nes
I.I
. . .... .......
Benzenes
5
(c) Low volatile bituminous coal
4
1
k Phenanthrenesor anthracenes
0 . - ~ ~ l l ...............
pyrolysis products from the Pocahontas # 3 low-volatile bituminous coal are aromatic and
aliphatic hydrocarbons with little oxygen-containing compounds. Generally, aliphatic and
aromatic oxygen-containing compounds decrease with increasing rank, with increasing
amounts of aromatic compounds.
Curie-point py-MS was used to characterize a set of 40 coals in Tromp et al. (1988).
Their results indicate that the classification of coal samples according to rank order is possi-
ble. The number and nature of pyrolysis products are characteristic of the coal's rank, par-
ticularly oxygen functionality of aromatic compounds. Tromp et al. (1988) also indicate
that the composition of pyrolysis products as a function of rank also gives fundamental
knowledge on the metamorphism of coal precursor materials and on the development of the
coal molecular structure of coals during maturation. Evidence of a mobile phase in coal is
also presented from Curie-point py-MS data.
1.4 Various Analytical Methods 29
150
TII (103 counts)
,.,,..
120
90 (a)
60
~, 30
o
0 100 260 300 400 560 660 760 860
o 2500
0 m/z 401-900
2000
1500 (b)
o
1000
m/z 2 0 1 - 4 ~ ] t/~
500
50-200__
0 100 200 300 400 500 600 700 800
Tamperature (~
122
[.., 1.0 [.., 1.0
~, 0.8 ~, 0.8
r~ r~
= 0.6 0.6
=- 0.4
9~D 9=-
9
0.4
> >
=9 0.2 9 0.2
"~
9
o ~, o
100 200 300 400 500 600 700 800 100 200 300 400 500 600 700 800
m/z m/z
Fig. 1.21 Time-resolved py-FIMS analysis of Pittsburgh coal during low-temperature, low-heating-rate de-
volatilization. (a) Temperature profile of total ion intensity. (b) Temperature profiles of integrated ion
intensities representing simulated differental thermogravimetric curves for the molecular ranges indicat-
ed. (c) Integrated py-FIMS of the low-temperature component which is released during heating of coal
in the first temperature interval crosshatched in (a). (d) Integrated py-FIMS of the high-temperature
component which is released during heating of coal in the second temperature interval crosshatched in
(a). [Reproduced with permission from Simmleit, N. et al., Advances in Coal Spectroscopy 330,
Plenum Press (1992)]
The low-temperature components derived from the foregoing analysis of ANL coals
were further studies in Meuzelaar et al. (1989) and Yun et al. (1991a, b) to verify the exis-
tence of a "mobile phase" as seen by py-MS. The presence of the low-temperature hump
(below 670 K) of the Pittsburgh coal TII in Fig. 1.21 (a) appears to explain 25-30% of the
integrated spectrum. Studies with Curie-point py-FIMS also demonstrated this low-temper-
ature hump, which consisted of alky-substituted aromatic (benzenes, naphthalenes and
phenanthrenes) and hydroaromatic (tetralins)compounds (Chakravarty et al., 1988). These
compounds were thought to be the bitumen components of bituminous coal. Comparison
of the low-temperature component, or the mobile phase component and the high-tempera-
30 1 Methodsof Classification and Characterization of Coal
~ Benzenes
~~e~henanthreneSn~
.0
!
1.0
(b) TG/MS
.~.
~ .5
"\
.0
1.0- ._--.-.
L
60 80 100 120 140 160 180 200 m/z
Fig. 1.22 Mass spectra of Pittsburgh coal in the overlapping region (m/z 50-200) by (a) Curie-point, (b) TG fur-
nace and (c) direct probe pyrolysis methods. [Reproducedwith permission from Huai, H. et al., Prepr.
Am. chem. Soc., Div. Fuel Chem, 35, 821,822 (1990)]
ture component, or bulk pyrolyzate, in Figs. 1.21(c), 1.2 l(d) reveals that even with similar
average molecular weights, there is a difference in the molecular-weight distributions and
compositions.
A comparative study on several coal samples by several different py-MS techniques
was also performed as shown in Fig. 1.22 (Huai, et al., 1990). Considering the wide range
of experimental conditions used with regard to sample size (25 p m to 5 mg) and heating
rate (1000 to 25 K/s), the three techniques produce remarkably similar mass spectral pat-
terns in the mass range m/z 50-200. On the other hand, the py-FIMS technique produces
less fragmentation with higher-molecular-weight ions. The variation in average molecular
1.4 Various Analytical Methods 31
weight observed by the different py-MS techniques is due to the type of mass spectrometer
used. Only detecting components below m/z 240, the Curie-point py-MS and TG-MS sys-
tems only record 10-40% of the pyrolysis products as compared to py-FIMS. Thus py-
FIMS appears to provide the most complete and detailed information on coal pyrolysis tars,
although Curie-point py-MS and TG/MS methods provide more reliable information on rel-
atively light gaseous pyrolysis products. Currently, short-column py-GC/MS is capable of
providing detailed information on pyrolysis compounds up to m/z 400, or two-thirds of total
tar.
~ /R RuO4
R-COOH
~ R ~ RuO4
HOOC" R "COOH
HOOC COOH
Fig. 1.23 Oxidationof various aromatic compounds using RuO4. [Reproducedwith permission from ed. Iino,
M.; Murata. S. et al., Primary and Higher Order Structures of Coal and Their Influence on Coal
Reactivity-Final Report on "Research for the Future" Coal Research Project-, 5 (2001)]
Drawbacks of this analysis are (1) difficulty in getting quantitatively lower carboxylic
acids due to higher solubility in water and higher volatility, and (2) poor carbon recovery.
As to the former problem, Nomura and co-workers proposed the use of an ion chromato-
graph for the analysis of lower carboxylic acids (Murata et a1.,1994), because of its advan-
tage like an easy experimental workup without any derivatization of the acids produced,
which was used in Strausz's work (Mojelsky et al., 1992). As to the latter, they achieved
the carbon balance in more than 85%, by improving the method of recovering water-soluble
products especially polycarboxylic acid derivatives with a careful workup. The details of
product workup are shown in Fig. 1.24.
Nomura et al. (2001) employed this oxidation reaction (called RICO, Ruthenium Ion
Catalyzed Oxidation) for analysis of two bituminous coals, Goonyella (GN) and Witbank
(WB), two brown coals, South Banko (SB) and Yallourn (YL). The RICO products were
separated into five portions, CO2, lower carboxylic acids, organic-soluble products, water-
32
Coal
RuC13. nHeO
NalO4
CCI4-CH3CN-H20
40 ~ 24-48 h, N2
Extraction
1st run
co2J 2nd run
with 5%
NaOH aq CH2C12, H20
Lower
carboxylic I
Filtration Organic
1
Aqueous
acids (C~-C5)
residue phase phase
Analysis with ion
Elemental Removal
chromatography
analysis of water
GC, GC-MS Solid
Elemental analysis esterification with
13C_NMR CH2N2 followed by
extraction with ether
.... L
Esterified Residne
water soluble
fractions Elemental
GC, GC-MS analysis
Elemental analysis
~3C_NMR
F i g . 1 . 2 5 Yield of aliphatic mono- and dicarboxylic acid from RICO of sample coals.
[Reproduced with permission from ed. Iino, M.' Murata. S., et al., Primary and Higher Order Structures of Coal and
Their Influence on Coal Reactivity-Final Report on "Research for the Future" Coal Research Project-, 6 (2001)]
1.4 Various Analytical Methods 33
9 = COOH
Fig. 1.26 Detected various carboxylic acids. [Reproduced with permission from ed. Iino, M.; Murata. S. et al.,
Primary and Higher Order Structures of Coal and Their Influence on Coal Reactivity-Final Report on
"Research for the Future" Coal Research Project-, 6 (2001)]
'ia)
9 I ]'q.De
@o
n--37
o
/
it [iL~ ,,li,ldllllil~ ~,
I
300
I
400
J
500
I
I
1i i i ,t
600
I
,,ta
" "
,,.,,., ,,., ,,..
700
I
,,, ,l ,,
m/z
Fig. 1.27 FD-MS spectra for water-(a) and organic-soluble (b) fractions of RICO of SB coal.
[Reproduced with permission from ed. Iino, M.; Murata. S. et al., Primary and Higher Order
Structures of Coal and Their Influence on Coal Reactivity-Final Report on "Research for the
Future" Coal Research Project-, 7 (2001)]
soluble products, and insoluble residue, as shown in Fig. 1.24. Summation yield of these
fractions ranged from 86% to 101%, this being more excellent than the results of previous
studies by Stock and Tse (1983). Yields of aliphatic mono- and dicarboxylic acids in or-
ganic-solubles are summarized in Fig. 1.25. As described in the earlier papers, the yield of
carboxylic acids decayed drastically as elongation of alkyl chains. The yield of shorter
chains or bridges (<-C5) from the sample coals was similar to each other, while the order of
the yields of acids with longer chains (=>C6) obeyed the following sequence: YL > SB
WB > GN, this meaning that longer chains or bridges were rich in low-rank coals. GC
analysis of products in the water-soluble portion was conducted, the major components de-
tected being aliphatic di- and tricarboxylic acids with 6-8 carbons and benzene polycar-
34 1 Methods of Classification and Characterization of Coal
The formation of these aromatic polycarboxylic acids indicates the presence of a highly
condensed polycyclic aromatic cluster; for example, RICO reaction of coronene afforded
benzene tetra- and hexacarboxylic acids, and biphenylhexacarboxylic acid. Organic phase
fraction was also analyzed by FD-MS. Fig. 1.27 (b) clearly indicates that the main compo-
nents were methyl esters of aliphatic monocarboxylic acids.
small and can usually be ignored. However, there are 1H with mass number 1, 2H with
mass number 2, and 3H with mass number 3 in atomic number 1, hydrogen, in which the
masses are largely different, rendering the isotope effect large. According to a report on the
H/D isotope in the reaction of hydrogen atoms with olefins at room temperature, the isotope
effect for kD/kH can be simplified as kD/kI-i=(mH/mD) 1/2, where kD and kH represent the ab-
solute rate constants of reaction of hydrogen and deuterium atoms with simple olefins at
room temperature; mH and mD are the masses of H and D, respectively (Ishikawa and Sato,
1979). A similarly simplified relationship can also be applied to the H/T system. When the
reaction of coal with hydrogen or tritium is performed at higher temperatures, however, the
isotope effect (kD]kH) is much smaller than the value, (mH]mT)1/2-- 0.58, since the dynamic
isotope effect decreases with increasing reaction temperature (Melander. 1960).
As units of radioactivity, Ci, Bq, dps (disintegration per second) and dpm (disintegra-
tion per minute) are used. 1 Ci indicates that 3.7 • 10 l~ of a radioactive nucleus disinte-
grate per second. 1 Ci is equal to 3.7 • 10 l~ Bq or 3.7 • 10 l~ dps. The number of disinte-
gration per unit time is defined as AN where ~ is the disintegration constant and N is the
number of a radioactive nucleus. As A, -- 0.693/T (T, half life time), AN is described as fol-
lows:
The number of hydrogen atoms in 1 g of water is 6.68 • 1022. This amount of tritium is
sufficient to trace in a chemical reaction. Even if this amount of tritium is added to hydro-
gen, tritium does not affect the reaction chemically. Therefore, when tritium is added to a
sample, the chemical behavior of hydrogen in the sample can be traced by measuring the
radioactivity.
Twenty years ago in coal science, deuterium was used as the tracer to elucidate the be-
havior of hydrogen. In the complicated system with coal, however, it was difficult to trace
hydrogen in coal by NMR or a mass spectroscopy because coal includes various com-
pounds and the solubility of coal to solvent is not so high. However, Kabe and coworkers
(Kabe et al., 1983) initially introduced a tritium tracer method into the research on coal liq-
uefaction to elucidate the reaction mechanisms. Details for this subuject are described in
Chapter 4. In the tritium tracer method, hydrogen or tritium in a solid, liquid or gas sample
is oxidized into water or tritiated water, which is measured with a liquid scintillation
counter. Since all hydrogen in coal is converted into water, this method enabled us to de-
termine hydrogen accurately. In this section, recent approaches to elucidate the coal struc-
ture, especially the determination of functional groups, are described. In these studies, the
reactions of coal with [3H]H20, [3H]H2, and [3H]organic solvent were investigated.
(Ishihara et al., 1999, 2000, 2001, 2002a, b, c; Qian et al., 1997).
36 1 Methods of Classification and Characterization of Coal
mesh). The samples of Wandoan were ground to under 150 mesh particles and dried for 3
h at 100 ~ under 10 -4 Torr. The analytical data of coals are shown in Table 1.10.
Tritiated water was purchased from the Japan Isotope Association (185 MBq/ml) and dilut-
ed with water to 106 dpm/ml. Phenol, 1-naphthol, toluidine, indole, and phenanthrene
(guaranteed reagent) were used without further purification. Commercially available deu-
terium oxide was used (D ~ 99.99%). All scintillator solvents for the measurement of ra-
dioactivity were also commercially available reagents.
One reaction procedure is as follows. One gram of coal and 1 g tritiated water (initial
radioactivity 106 dpm) were added into a 25-ml Pyrex glass reactor. After the mixture was
degassed in vacuum via three freeze-pump-thaw cycles, the reactor was immersed into an
oil bath and the reaction mixture was stirred with a magnetic stirrer. The reaction tempera-
tures were 50 ~ and 100 ~ and the reaction times were 1-24 h. After the reaction, the reac-
tion mixture was separated into tritiated water and coal with a vacuum line ( ~ 10 -4 T o r r ) .
Every tritiated water sample (ca. 0.4 g) was dissolved into 14 ml of a scintillator solvent
(Monophase S) and the radioactivity of the obtained solution was measured with a liquid
scintillation counter. After drying in vacuum at 50 ~ or 100 ~ for 7 h, tritiated coal was
oxidized by an automatic sample combustion system into tritiated water to measure its ra-
dioactivity.
The reactions of model compounds with tritiated water were performed in a similar
way. After the reaction, the tritiated model compounds were isolated by filtration or by dis-
tillation. The radioactivity of the recovered water was measured in a way similar to that de-
scribed above. The radioactivity of the model compound was measured by adding it into a
scintillator solvent (Instafluor for nonpolar compounds, Permafluor for polar compounds).
In order to elucidate the isotope effect of hydrogen, deuterium oxide was also used in
the exchange reaction with coals instead of the tritiated water in the batch reactor. The ratio
of deuterium to hydrogen in water before and after reaction was measured with 1H-NMR.
The other reaction procedure is as follows. In order to estimate the behavior of hydro-
gen in the exchange reaction at higher temperature, a pulse method using a tritium tracer
was developed. The hydrogen exchange was carried out using a glass column reactor in a
gas chromatograph (GC) equipped with thermal conductivity detector (TCD), as shown in
Fig. 1.29. About 0.5 g of coal sample was packed into the column (i.d. 4 mm) and was
fixed with quartz wool at both ends of the column. The flow rate of nitrogen as a carrier
gas was 10 ml/min. After the column was heated to the desired reaction temperature and
held for 4 h, a pulse of tritiated water (2-10 ml) was introduced with a microsyringe every
30 min. During the reaction, tritiated water was detected by TCD and recovered by bub-
bling through a scintillator solvent (Monophase S, 6 ml) at the outlet of the GC. Then, 8 ml
38 1 Methods of Classification and Characterization of Coal
Microsyringe
(Tritiated water 2-10/.d/pulse)
Recorder
I i
TCD
Carrier gas
Nitrogen
10 ml/min ~ i I~;l~lt.i~.'l III111111111111
n I] las
I/////1
I/////1
Fig. 1.29 Flow chart of hydrogen exchange reaction of coal with tritiated water in a pulse flow reactor.
[From Qian, W. et. al., Energy Fuels, 11, 1289 (1997)]
M o n o p h a s e was added into the sample and its radioactivity was measured by liquid scintil-
lation counter. Several tritiated water pulses were introduced into the coal-packed column
until the radioactivity of the r e c o v e r e d pulse a p p r o a c h e d that of the introduced pulse.
Similar to the case using the batch reactor, the radioactivities of recovered tritiated water
and coal were measured after the reaction.
Calculation of hydrogen exchange ratio (HER) is calculated as follows. Table 1.11
shows a typical e x a m p l e of the experimental data obtained in the e x c h a n g e reaction of
Illinois # 6 coal with tritiated water at 100 ~ for different reaction times. The HER was
obtained and calculated on the basis of such data. The H E R is the ratio of exchangeable
hydrogen in coal (Hex) to the total amount of hydrogen in an original coal (Hcoa0. The HER
between coal and water was calculated on the basis of Eq. (1.30).
H E R -- Hex/Hcoal (1.30)
1.5 TritiumTracer Methods for Coal Characterization 39
Ocoal w a s calculated using the analytical data presented in Tables 1.10 and 1.11.
The
amount of hydrogen exchanged between water and coal (Hex) represents the amount of the
exchangeable hydrogen in coal and was calculated on the basis of Eq. (1.31)
A. H y d r o g e n E x c h a n g e Reaction in a B a t c h R e a c t o r
Figure 1.30 shows the change in hydrogen exchange ratio (HER) of Illinois #6 coal at 100
~ with reaction time. The coal used was as received or dried in vacuum before reaction.
The HER of the coal as received gradually increased and approached 7.8% after 4 h while
the HER of the dried coal took longer (6 h) to reach a similar constant value. Although this
delay may be due to the effect of the internal diffusion of water in the coal, the effect of the
internal diffusion of water could be neglected when the reaction time was sufficiently long
( > 6 h).
The effect of exchange temperature on the HER of coal is shown in Fig. 1.31.
Although more time was required to reach the equilibrium of hydrogen exchange at 50 ~
10 i 100 ~
8 k Q
9
~. 6
~z 4
2 o As received
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Reaction time (h)
Fig. 1.30 Hydrogenexchange ratio in the exchange reaction of Illinois #6 coal with tritiated water.
[From Qian, W. et. al., Energy Fuels, 11, 1290 (1997)]
than at 100 ~ the HER even at 50 ~ after reaction for 12 h also reached 7.8%, which is
as well representative of the HER as the 100 ~ experiment. Thus, it is likely that for the
reaction, the rate-limiting step is not the speed of the exchange reaction between water and
coal but the diffusion of water into the coal. That is, temperature in fact affects the diffu-
sion rate of the water into coal.
When the hydrogen exchange with tritiated water was carried out using other coals,
40 1 Methods of Classification and Characterization of Coal
10
= 4
2 9 100~
II 50 ~
0 , I , I , t , ~ , ~ , t , ~ , i , i , I , I , 1~2 ,
0 1 2 3 4 5 6 7 8 9 10 11 13
Reaction time (h)
similar results were obtained. It was observed that the HERs for all coals in reaction for
over 6 h at 100 ~ ended toward a constant value independent of drying of the coal before
reaction. The constant value of the HER for each coal was obtained for over 6 h of reaction
time and presented in Table 1.12.
Table 1.12 Hydrogen Exchange Ratio of Coal with Water (%, 100 ~ Batch Reactor)
Coal ND WA IL UF POC
In order to investigate the accuracy of the tritium tracer method in the hydrogen ex-
change reaction, the exchange reaction of coal with deuterium oxide was also carried out at
100 ~ The concentration of protons in recovered water after reaction was determined by
1H-NMR. HER was calculated from the increase in the amount of proton in recovered wa-
ter (deuterium oxide) before and after reaction. These results are also presented in Table
1.12. Compared with the results obtained with tritiated water, no significant difference be-
tween the two sets of HERs was observed except for POC coal. In the case of POC coal,
the value for the deuterium-NMR method was higher than that for the tritium tracer
method, suggesting that there is a larger margin of error for the deuterium-NMR method.
very little hydrogen in phenanthrene was exchanged with water, the aromatic hydrogen in
coal was assumed not to be exchanged with the hydrogen in water. In contrast, all het-
eroatom compounds could readily exchange hydrogen with water and the HERs of these
compounds were approximately the same as the ratios of hydrogen in functional groups to
total hydrogen derived from the stoichiometry of the model compounds. Acid- and base-
catalyzed deuterium incorporation into aromatic nucleus has been reported in the literature
(Thomas, 1971; Ingold et al., 1936; Calf and Garnett, 1973). Ingold et al. showed that the
reaction of phenol with deuterium oxide in the presence of NaOH exchanged three nuclear
hydrogen atoms in an aromatic ring after 3 0 - 4 0 days at 100 ~ This does not agree with
the results mentimed above. This is not unexpected since in the present study, neutral water
was used and the reaction time was not so long. In a recent study on the protium-deuterium
exchange of heteroaromatics with D 2 0 in the neutral condition, Werstiuk and Ju (1989) re-
ported that the hydrogen in the aromatic ring of phenol was scarcely exchanged by deuteri-
um at 165 ~ for 24 h. This is in good agreement with the present results. Therefore, it is
considered that hydrogen in the functional groups of coal such as hydroxy and carboxylic
acid is rapidly exchanged through the proton exchange between water and coal, and that the
effect of temperature is less for this type of ion-exchange reaction because the rate of ion-
exchange reaction is generally very rapid. Hence, it is suggested that the HER of coal at
lower temperatures represents the amount of hydrogen in the functional groups of coal.
10000
ooo
6ooo
~ 4000 ~/,
000
0 I i i i i i
1 2 3 4 5 6 7 8 9 10
Number of pulse
Fig. 1.32 Radioactivity of introduced pulse of tritiated water. (Illinois #6 coal, 200 ~
[From Qian, W. et. al., Energy Fuels, 11, 1291 (1997)]
30
100 ~
25
20
#.
15
ud
10
5 -
i i i I i i i i
0
70 75 80 85 90 95 100
Carbon (%)
Fig. 1.33 Hydrogen exchange ratio in the exchange reaction of coal with tritiated water in a batch or pulse
reactor. 9 Batch reactor, IP Pulse reactor. [From Qian, W. et. al., Energy Fuels, 11, 1291 (1997)]
1.5 Tritium Tracer Methods for Coal Characterization 43
HER were determined from the difference in the radioactivity between the introduced and
recovered pulse or from the radioactivity of the tritium incorporated into coal obtained by
combustion of coal. The results are presented in Table 1.14. Similar results were obtained
with other coals and these are also presented in Table 1.14. HERs of Beulah-Zap,
Wandoan, Upper Freeport, and Pocahontas coals at 100 ~ were 16.9, 6.4, 2.0, and 0.8%,
respectively. The results using the pulse reactor at 100 ~ were compared with those using
the batch reactor in Fig. 1.33. There is no significant difference in HERs between the two
types of reactors, although the results obtained in the pulse reactor are slightly lower than
those obtained in the batch reactor. It is well known that the diffusion rate of material in a
flow system is much faster than that in a batch system. Thus, the exchange reaction in the
pulse reactor could approach an equilibrium state although the reaction time was much
shorter than that in the batch reactor. This further indicates that the hydrogen exchange rate
with water is very rapid and that the diffusion rate is the limiting step of the hydrogen ex-
change. In addition, the consistency of data between the two reactors also shows that the
extent of hydrogen exchange into the wall of the glass reactor in the batch method is negli-
gible in the present study.
30
20
10
I I i i
ND WA IL UF POC
Coal rank
low ,, high
Fig. 1.34 Comparison of hydrogen exchange ratio with content of hydroxy group.
9 Measured hydrogen exchange ratio
[] Calculated from oxygen content
[] Calculated from OH content (Attar et al., 1982)
[] Calculated from OH content (Solomon et al., 1991)
[From Qian, W. et. al., Energy Fuels, 11, 1292 (1997)]
30
25
20
m
- ~ II
15 B
10
_ ~ t. t,
0 i i i I:1:1 ~ I i I i I i I ,
0 50 100 150 200 250 300 350
Temperature (~
al groups was exchanged with hydrogen in water. It was proposed that, in the hydrogen ex-
change reaction of coal and coal-related compounds with tritiated water, aromatic hydrogen
in an aromatic ring substituted by such a functional group as hydroxy group in coal would
be exchangeable (Ishihara et al., 1993). It was shown that only hydrogen in the hydroxy
group was exchangeable at 100 ~ while hydrogen at the para or ortho position of 1-naph-
thol became exchangeable at 300 ~ Therefore, in this study using the pulse flow reactor,
1.5 Tritium Tracer Methods for Coal Characterization 45
it also appears that a part of the hydrogen in aromatic rings substituted by functional groups
as well as the hydrogen in the functional groups was exchanged at 300 ~
Similar to the previous section, four kinds of coals of the Argonne Premium Coal
Sample Program ( < 100 mesh), and Wandoan coal ( < 150 mesh) were used as raw materi-
als. Coal samples were dried under vacuum at 120 ~ for 1 h. When the reaction was per-
formed at 200, 250 or 300 ~ the reaction of coal with tritiated water was performed in a
batch type stainless tube reactor (i.d. 8mm, length 12 cm). The reactor was put into a fur-
nace heated to the expected temperature. After the reaction, water was removed and the
separated coal was dried under vacuum at 120 ~ for 1 h. After drying, the coal sample
was oxidized to water using the combustion system to measure its radioactivity with a liq-
uid scintillation counter. In order to determine the hydrogen in functional groups, the hy-
drogen exchange reactions of the tritiated coal and water were performed at 100 ~ for 24 h
in a glass reactor. After the reaction, suction filtration was performed and tritiated coal was
washed with hot water. Further, the separated coal was dried under vacuum at 120 ~ for 1
h. After drying, the coal sample was oxidized by a method similar to the above using the
combustion system to measure its radioactivity with the liquid scintillation counter.
The reactions of model compounds with deuterium oxide were performed in a similar
way. After the reaction, the deuterated model compounds were isolated by filtration or by
distillation. The deuterated position of model compounds was identified by 1H and 13C-
NMR.
46 1 Methods of Classification and Characterization of Coal
2O
O
O
15
O ........................ O---
5
I I
00 100 200 300 400
Reaction time (min)
0 " HER-Total O 9HER-OH (3H-Coalc=:,H20)
Fig. 1.36 Effect of reaction time on hydrogen exchange ratio of Illinois #6 coal at 250 ~
[From Ishihara, A. et. al., Energy Fuels, 16, 34 (2002)]
Table 1.15 Hydrogen Exchange Ratios of Coals with Tritiated Water in a Batch Reactor
ring.
Using total HER at 300 ~ (HER- Total300oc) and HER of functional groups at 100 ~
(HER-OH10o oc), the ratio of the amount of aromatic hydrogen to the amount of functional
groups was calculated according to Eq. (1.33).
AR-H/OH -- (HER-Total300 oc -- HER-OH100 oc)/HER-OH100 oc (1.33)
The aromatic hydrogen in a ring with a carboxy group hardly exchanges with water,
which presents later in the reaction of the model compound. Therefore, in the case of ND
coal, half the value of the HER of the functional group was used in the denominator of Eq.
(1.33). The ratios of the amount of exchangeable aromatic hydrogen to the amount of func-
tional group (phenolic OH group) were 2.7, 2.3, 0.9 and 1.7, for ND, IL, UF, and POC, re-
spectively. The hypothetical positions exchanged in the model structures are shown in Fig.
1.37. For ND coal there may be the structure I with a single ring and one hydroxy group
and three exchangeable ortho and para hydrogens. In IL coal, there may be a two-ring sys-
tem where one ring has a hydroxy group, and hydrogen atoms at ortho and para postions
are exchangeable and can be exchanged with tritiated water under high pressure of water at
elevated temperatures (Structures II and III). The aromatic hydrogen in the other ring
without a hydroxy group does not exchange. For higher rank coals such as UF and POC
coals, there may be two-or three-ring structures (Structures IV, V and VI). In structures
IV, V and Vl, one hydrogen atom is exchangeable.
In order to investigate in detail what kind of hydrogen in coal can be exchanged, the re-
actions of a coal model compound with deuterium oxide were performed. Aniline, anisole,
benzoic acid, ethyl benzoate were used as model compounds. The reaction was performed
OH
~OH
T T IL
T T ND (POC)
I II
~ ~
T
III
T IL
(POC) ~ IV
T
UF
POC
OT
UF
POC POC
V VI
Fig. 1.37 Exchanged positions in model structures. [From Ishihara, A. et. al., Energy Fuels, 16, 35 (2002)]
at 300 ~ for 180 min. Deuterium in the compound after the reaction was determined by
NMR. Hydrogen at ortho and para positions of aniline was exchanged with deuterium ox-
ide and the reaction almost reached the equilibrium state. No aromatic hydrogen ex-
changed with deuterium oxide in anisole. Although a trace amount of deuterium seemed to
48 1 Methods of Classification and Characterization of Coal
pKa
OH O- (25 ~
+ H§ 9.14
OH O-
O O+ H+ 9.82
~2 HN~
H+ 4.65
OH O-
0-O+ .+
O- 0 0
0 .0- -0
0 -T~o.O T .0"
o o i Oi
0 T~ -0
o o OT
iT W
Fig. 1.40 Mechanism of hydrogen exchange between phenol and tritiated water.
[From Ishihara, A. et. al., Energy Fuels, 16, 34 (2002)]
be introduced into the ortho position of benzoic acid, the exact amount could not be traced
and only the signal of ortho carbon in N M R became slightly small.
The electron densities of the model c o m p o u n d s were calculated using W i n M O P A C
molecular orbital calculation software. The data for phenol, aniline and anisole are shown
in Fig. 1.38. The electron densities of ortho and para positions of phenol and aniline were
1.5 TritiumTracer Methods for Coal Characterization 49
higher than those of other positions. However, the electron densities of the ortho and para
positions of anisole were also higher than those of other positions and very close to the val-
ues for phenol and aniline. Therefore, the results obtained in the present study cannot be
explained by this electron density of model compounds.
The common feature of phenol and aniline is that they can dissociate to a proton and a
counter anion as shown in Fig. 1.39. The anion structures may be related to the hydrogen
exchange with water. As a model compound phenol is used to explain the exchange mech-
anisms (Fig. 1.40). Phenol dissociates to a proton and a phenoxide. This phenoxide anion
has these resonance structures. These structures may contribute the hydrogen exchange at
the ortho and para positions. In the reaction of coal with tritiated water, it is believed that
exchange of aromatic hydrogen through these routes may occur.
1.5.3 Catalytic Hydrogen Exchange Reaction of Coal with Tritiated
Gaseous Hydrogen
Coal has a complex structure that includes various aromatics and functional groups. To
convert coal into useful fuels and chemicals by thermolysis and hydrogenation, it is impor-
tant to elucidate its structure (Solomon et al., 1991; Kelemen et al., 1990). The reactivity of
hydrogen in coal reflects the coal structure since each hydrogen in aromatic, aliphatic, and
functional groups with heteroatoms, etc., has a different nature. Therefore, the determina-
tion of the reactivity of hydrogen in coal provides significant information regarding its
structure.
There are in principle two methods to investigate the chemical structure of coal. One
includes destruction of the coal macromolecules into representative fragments to deduce the
initial structure of the coal from the structure identified from the fragments (Murata et al.,
1994). The other is the direct nondestructive characterization of coal in its original form by
solid state spectroscopic methods (Haenel, 1992; Solomon et al., 1990; Martin et al., 1988).
Useful methods to measure hydrogen in coal employ isotopes such as deuterium and tritium
tracers (Franz and Camaioni, 1981b; Cronauer et al., 1982; Wilson et al., 1984; Collin and
Wilson, 1983; Skowronski et al., 1984; Kabe et al., 1990d, 1991a, b, Ishihara et al., 1995).
In early works, these tracers were used to measure the amounts of hydrogen transferred in
coal liquefaction. Although the method includes destruction of the original structure of
coal, the reactivity of hydrogen in coal with gaseous hydrogen and hydrogen in solvent can
be estimated. A deuterium tracer was effective to trace reactive sites in coal and coal model
compounds. However, there were few examples which enabled quantitative analysis of the
hydrogen mobility in coal because of the poor solubility of coal products and the difficulty
of quantification of the deuterium tracer (Franz and Camaioni, 1981b; Cronauer et al.,
1982; Wilson et al., 1984; Collin and Wilson, 1983; Skowronski et al., 1984). In contrast,
it has been reported that tritium tracer techniques were effective to trace quantitatively hy-
drogen in coal liquefaction (Kabe et al., 1990d, 1991a, b; Ishihara et al., 1995). These
works showed that quantitative analysis of hydrogen mobility of coal could be achieved
through the hydrogen exchange reactions among coal, gas phase and solvent as well as hy-
drogen addition. The authors were interested in the direct nondestructive determination of
hydrogen in coal in its original form. As described above, the tritium tracer methods are ef-
fective to determine hydrogen in the functional group of coal through the reaction of coal
with tritiated water in a pulse flow reactor or a batch reactor (Ishihara et al., 1993; Qian et
al., 1997).
In this section, the hydrogen exchange of coal with tritiated gaseous hydrogen in the
presence of a catalyst to estimate the mobility of hydrogen in coal is described. In the ex-
50 1 Methodsof Classificationand Characterizationof Coal
periment, a pulse of tritiated hydrogen was introduced into the reactor. A hydrogen atom is
generated on the catalyst and hydrogen exchange occurs without destruction of the coal
structure. After the hydrogen exchange with tritiated gaseous hydrogen, the reaction of the
tritiated coal and water was carried out to remove tritium in functional groups and to obtain
the information for the position of hydrogen exchanged in coal.
[3H]H2
Qi ~ ~ Recovery
I --1
"
Scintillator1
] Recoder ~ ] N cell r-'--
t ,
I
, ~ _ . . . . . . . . .
Radioanalyzer
Recorder
~
al Recovery
Reactor
N2or H2
Fig. 1.41 Schematicdiagramof experimentalapparatus. [Reproducedwith permissionfrom Kabe,T. et al.,
Fuel, 79, 312, Elsevier(2000)]
1.5 Tritium Tracer Methods for Coal Characterization 51
reactor was heated to the desired reaction temperature (200, 250 and 300 ~ and held for 2
h to dry the coal, a pulse of tritiated gaseous hydrogen (6420 dpm/ml) was introduced into
the coal-catalyst bed using a 6-way valve with a high-pressure gas sampler tube (9.62 ml)
every 30 min. During reaction, effluent tritiated gaseous hydrogen was detected by TCD.
The radioactivity of tritiated gaseous hydrogen recovered from the reactor, i.e., unreacted
[3H]H2, was directly monitored with a radioanalyzer. Several tritiated gaseous hydrogen
pulses were introduced into the reactor until the radioactivities of recovered pulse ap-
proached that of the introduced pulse. After reaction, the tritiated coal was oxidized by an
automatic sample combustion system into tritiated water to measure radioactivity. Every
tritiated water sample was dissolved into 14 ml of Monophase S, a commercial scintillator
solvent and measured with a liquid scintillation counter.
In order to predict locations Of exchangeable hydrogen in coal, hydrogen exchange re-
actions of coal that reacted with tritiated gaseous hydrogen, i.e., tritiated coal, and water
were performed. The tritiated coal and water were added into a glass reactor. The reactor
was immersed into an oil bath with stirring. The reaction was performed at 100 ~ for 24 h.
After the reaction, suction filtration was performed and tritiated coal was washed with hot
water. Further, the separated coal was dried under vacuum ( < 10 -4 Torr) at 120 ~ for 1 h.
After the coal was dried, it was oxidized by a method similar to the above using the com-
bustion of coal.
The hydrogen exchange ratio (HER) described here means the ratio of the amount of
hydrogen exchanged in coal (Hex) to the total amount of hydrogen in an original coal (Hcoal).
The HER between coal and gaseous hydrogen was calculated based on Eq. (1.34).
Hcoal was calculated using the analytical data in Table 1.10. The amount of hydrogen ex-
changed between gaseous hydrogen and coal (Hex) was calculated using Eq. (1.35).
3000
2500 -
_
2000
0
~9 1500
.=
O
1000
500
0 i i
1 2 3 4 5 6 7
Number of pulse
Fig. 1.42 Variation in radioactivity of introduced pulse of tritiated gaseous hydrogen
(Pocahontas coal, 250 ~ with 1%-Pt/A1203).
9 Introduced pulse C) Recovered pulse
[Reproduced with permission from Kabe, T. et al., Fuel, 79, 313, Elsevier (2000)]
and the coal. It can be assumed that the hydrogen exchange between the tritiated gaseous
h y d r o g e n and the coal r e a c h e d e q u i l i b r i u m after the i n t r o d u c t i o n of the fifth pulse.
According to Eqs. (1.34) and (1.35), the amount of hydrogen exchanged and the H E R are
determined from the difference in the radioactivity between the introduced and recovered
pulse or from the radioactivity of tritium incorporated into coal obtained by the combustion
of coal. HERs of IL # 6 coal with tritiated gaseous hydrogen calculated by the two methods
are compared in Fig. 1.43. The H E R increased with increasing temperature. At all temper-
atures, the H E R derived from the balance of radioactivity in the tritiated gaseous hydrogen
measured by the radioanalyzer was consistent with the H E R derived from the conversion of
40
30
20
10-
0 i t i
150 200 250 300 350
Reaction temperature (~
Fig. 1.43 Hydrogenexchange ratio of Ilinois #6 coal with [3H]H2.
(3 Derived from combustion of coal
9 Derived from balance of [3H]H2 in radioanalyzer
[Reproduced with permission from Kabe, T. et al., Fuel, 79, 313,
Elsevier (2000)]
1.5 Tritium Tracer Methods for Coal Characterization 53
301
25
20
15
10
0
2OO 250 3OO
Reaction temperature (~
Fig. 1.44 Effect of reaction temperature on hydrogen exchange ratio of Illinois #. 6 coal.
m: HER-Total I1: HER-OH ([3H]Coalc=>H20) U]: HER-OH (Coalc=>[3H]H20)
[Reproduced with permission from Kabe, T. et al., Fuel, 79, 314, Elsevier (2000)]
tritiated coal into tritiated water the radioactivity of which is measured by liquid scintilla-
tion counter.
It has been already reported that hydrogen in the hydroxy group present in coal ex-
changes with water at 100 ~ Therefore, in order to understand the behavior of hydrogen
of functional groups in coals in the hydrogen exchange reaction between coal and gaseous
hydrogen, the hydrogen exchange reactions between the tritiated coal, which was obtained
in the reaction with tritiated gaseous hydrogen, and water were performed at 100 ~ for 24
h in a batch reactor. The HER between tritiated IL coal and water is shown with the HER
between coal and tritiated gaseous hydrogen in Fig. 1.44. HERs increase with increasing
temperature. However, HERs between tritiated coal and water (HERs removed with water)
increase only slightly with temperature. At 200 ~ the HER with tritiated gaseous hydro-
gen is very similar to the HER with water, indicating that most of the former HER corre-
sponds to the hydrogen exchange of the functional group at this temperature. At 300 ~
the latter HER is very similar to the HER between coal and tritiated water described in the
previous section (Qian et al., 1997). This suggests that the hydrogen of the functional
group in coal can easily be exchanged with gaseous hydrogen in the presence of a Pt cata-
lyst.
50
40
~" 30
~Z 20
10
ND ND
I I
WA IL
II
UF
II
POC
No catalyst
Coal rank
Low High
Fig. 1.45 Hydrogen exchange ratio of coal with tritiated gaseous hydrogen at 250 ~ [Reproduced with
permission from Kabe, T. et al., Fuel, 79, 314, Elsevier (2000)]
60
ND WAIL UF POC
50
40
30
20-
10-
0 I I I I
70 75 80 85 90 95
Carbon (%)
Fig. 1.46 Effect of coal rank on hydrogen exchange ratio at several temperatures.
I1:200 ~ O" 250 ~ A: 300 ~ [Reproduced with permission
from Kabe, T. et al., Fuel, 79, 314, Elsevier (2000)]
the HERs between tritiated coal and water are very similar to those between coals and triti-
ated water described above. The results show that in all cases the hydrogen of functional
groups such as the hydroxy group in coals is almost completely exchanged with tritiated
gaseous hydrogen at 300 ~ The difference between hydrogen of coal exchanged with
gaseous hydrogen and hydrogen of functional groups is larger for lower rank coals. This
shows that, besides the hydrogen in the functional group, the hydrogen of aromatics with
functional groups or hydrogen at the a position of alkyl groups may be related to this ex-
change reaction because the low rank coals include large amounts of functional groups and
alkyl groups.
In the hydrogen exchange of model compounds using the deuterium tracer, Benjamin et
al. (1982) studied the hydrogen exchange reaction of a group of aromatic compounds in re-
1.5 Tritium Tracer Methods for Coal Characterization 55
60
50
40
30
20
10
ND WA IL UF POC
Fig. 1.47 Hydrogen exchange ratio of coal with tritiated gaseous hydrogen at 300 ~
m: HER (Coalc:~[~H]H20) N: HER([3H]Coalc=>H20) [--]:HER (Coalc=~[3H]H20)
[Reproduced with permission from Kabe, T. et al., Fuel, 79, 314, Elsevier (2000)]
cycle solvents with diphenylmethane-D2 (Ph2CD2), deuterated pyrene o r D2 gas under lique-
faction conditions by assuming that the reactivity toward hydrogen exchange is related to
the hydrogen shuttling. They reported that methyl-substituted aromatics such as methyl-
naphthalene and toluene undergo extensive exchange reactions while nonsubstituted aro-
matics such as naphthalene, biphenyl and diphenyl ether show little observable exchange
with three deuterated reagents. They concluded that the methyl substituted aromatic and
hydroaromatic compounds in the recycle solvent make the most important contribution to
the hydrogen shuttling and hydrogen transfer. However, the detailed mechanisms and loca-
tion of the hydrogen exchange are not discussed.
The amount of hydrogen in coal exchanged with tritiated gaseous hydrogen is larger for
lower rank coals which have larger amounts of alkyl groups or heteroatom functional
groups. Possible mechanisms for hydrogen exchange of coal with tritiated gaseous hydro-
gen in the presence of a catalyst are illustrated using model compounds in Fig. 1.48. Since
hydrogen in low rank coal is more labile in this reaction, hydrogens in alkyl groups, func-
tional groups with a heteroatom and aromatics with those functional groups may be related
to the hydrogen exchange. When toluene is used as a model compound, a benzyl radical is
formed in the reaction with a tritium radical (or hydrogen atom). The benzyl radical reacts
with the tritium radical to form tritiated toluene. The benzyl radical may be stabilized at
equilibrium with a methylphenyl radical. The methylphenyl radical reacts with the tritium
radical to form the other tritiated toluene which is tritiated at the ortho or para position.
When phenol is used as a model compound, a phenoxyl radical is formed in the reaction
with the tritium radical (or hydrogen atom). The phenoxyl radical reacts with the tritium
radical to form the tritiated phenol. The phenoxyl radical may be stabilized at equilibrium
with the hydroxyphenyl radical. The hydroxyphenyl radical reacts with the tritium radical
to form the other tritiated phenol which is tritiated at the ortho or para position.
56 1 Methods of Classification and Characterization of Coal
CH3 CH2 ~
CH2 9 CH2T
OH Oo
+ T9 ~ + HT
O9 OT
+ T2 ~ + T~
O9 OH OH
'or G
OH OH OH OH
or + T2 or + T9
9 T
Fig. 1.48 Hydrogen exchange reaction of coal with tritiated gaseous hydrogen in the presence of catalysts.
[Reproduced with permission from Kabe, T. et al., Fuel, 79, 315, Elsevier (2000)]
coal structure since each hydrogen in the aromatics, aliphatics, functional groups with het-
eroatoms, etc. has a different nature. Thus, the determination of the reactivities of hydro-
gen in coal will provide significant information regarding the coal structure.
Useful methods to measure hydrogen in coal utilize isotopes such as deuterium and tri-
tium tracers (Ishihara et al., 1995; Franz and Camaioni, 1981b; Cronauer et al., 1982;
Wilson et al., 1984; Collin and Wilson, 1983; Skowronski et al., 1984; Kabe et al., 1990,
1991a, b). It has been reported that tritium tracer techniques were effective in tracing quan-
titatively hydrogen in coal liquefaction (Ishihara et al., 1995; Kabe et al., 1990, 1991a, b).
In these works, it was shown that quantitative analysis of the hydrogen mobility of coal
could be conducted through the hydrogen exchange reactions among coal, gas phase and
solvent as well as hydrogen addition. Recently, we reported that the tritium tracer methods
are effective for determining hydrogen in the functional group of coal through the reaction
of coal with tritiated water and tritiated gaseous hydrogen in a pulse flow reactor as well as
a batch reactor (Ishihara et al., 1993; Qian et al., 1997; Kabe et al., 1999, 2000).
In the reaction with tritiated water, it was assumed that protons were related to the hy-
drogen exchange of functional groups in coals (Ishihara et al., 1993; Qian et al., 1997). In
the reaction with tritiated gaseous hydrogen, a Pt/A1203 catalyst was used to generate hy-
drogen radicals (Kabe et al., 1999, 2000). In these studies, the direct nondestructive deter-
mination of hydrogen in coal in its original form was attained. However, the effect of the
particle size of coal on the hydrogen exchange reaction in these works was not into account.
If the particle size affects the hydrogen exchange, an exchangeable hydrogen such as
the hydrogen of a functional group will be located on the inside of the coal structure which
is difficult for the hydrogen atom or proton to approach. In contrast, if the particle size
does not affect the hydrogen exchange, exchangeable hydrogen will always be located on
not only the exterior surface but also on the interior surface of the coal structure or the
cross-linking position in the macromolecular network structure of coal (Takanohashi et al.,
1999a; Nakamura et al., 1995) which is easy for the hydrogen atom or proton to approach
even in coal of large particle size. The latter case may suggest that there is no pore-like
structure which is difficult for the hydrogen atom or proton to approach and that a coal par-
ticle breaks into smaller pieces by separation from the large layer structure of the aromatic
group or the breaking of cross-links in the macromolecular network structure of coal.
In this section, the hydrogen exchange of coal with tritiated gaseous hydrogen in the
presence of a catalyst is described to confirm the effect of particle size of coal on the hydro-
gen exchange. In the experiment, a pulse of tritiated hydrogen was introduced into a reac-
tor containing coal and the catalyst. Initially, to determine the amount of the Pt/A1203 cata-
lyst, the effect of the amount of the catalyst on the hydrogen exchange was investigated.
The hydrogen atom is generated on the catalyst, and the hydrogen exchange occurs without
destruction of the coal structure. After the hydrogen exchange with tritiated gaseous hydro-
gen, the reaction of the tritiated coal and water was carried out to remove the tritium from
the functional groups and to obtain information on the position of the hydrogen exchanged
in coal. The hydrogen exchange reaction of coal with tritiated water was also carried out
using a batch reactor system to confirm the effect of particle size of coal, and the results
were compared with the results of experiments using tritiated gaseous hydrogen.
58 1 Methods of Classification and Characterization of Coal
40 -
~ 3o
<
10
0
25 45 75 106 150 212 300 425 600 850 1000
(b)
35
30
- 100 mesh - 2 0 mesh
25
20
o 15
E
6( c )
50
- 100 mesh - 2 0 mesh
II
40 ~
< 20
10
Initially the hydrogen exchange reaction between coal and tritiated gaseous hydrogen
was performed using the pulse flow reactor. A pulse of tritiated gaseous hydrogen with a
constant radioactivity (16800 counts/pulse) was introduced into IL coal with 50 mg of 1%-
Pt/A1203 at 250 ~ every 30 min. Fig. 1.50 shows the change in radioactivity of tritiated
gaseous hydrogen in a recovered pulse with the n u m b e r of introduced pulse. After the first
pulse was introduced, the radioactivity of the recovered pulse was only 1900 counts. This
indicates that some tritium in the tritiated gaseous h y d r o g e n was incorporated into coal.
Further, the radioactivity in the recovered pulse increased with the n u m b e r of pulses intro-
18000
16000
14000
= 12000
o
o
10000
>
o 8000
o
6000
4000
2000
0
1 4 7 10 13 16 19 22 25
Number of pulse
Fig. 1.50 Variation in radioactivity of introduced pulse of tritiated gaseous hydrogen.
(Illnois # 6, 250 ~ - 100 mesh, 1%-Pt/A1203 50 mg)
[From Ishihara, A. et al., Energy Fuels, 14, 708, (2000)]
30
25 9
lO
. . . . . . . . . . . . . . . . . . . . . . . h. A ........
5 9
0 ) I I I I
0 20 40 60 80 100 120
Amount of catalyst (mg)
Fig. 1.51 Effect of amount of the catalyst on hydrogen exchange ratio of Illinois # 6 coal
with 3%-Pt/A1203 at 250 ~
C): HER-Total (Coalc=>[3H]H2) O: HER ([3H]Coalc=~H20)--- : HER (Coalcz~[3H]H20)
[From Ishihara, A. et al., Energy Fuels, 14, 709, (2000)]
60 1 Methods of Classification and Characterization of Coal
duced and approached a constant value (16800 counts) at the 24th pulse. In contrast, the
amount of recovered gaseous hydrogen, which was monitored by the TCD at the outlet of
the reactor, remained approximately constant for every pulse introduced. This indicates
that the decrease in the radioactivity of the introduced pulse could not be attributed to the
adsorption/desorption of gaseous hydrogen in the coal but to the hydrogen exchange be-
tween the tritiated gaseous hydrogen and the coal. The hydrogen exchange between the tri-
tiated gaseous hydrogen and the coal can be assumed to have reached equilibrium after the
introduction of the 24th pulse. According to equations described in the preceding section,
the amount of hydrogen exchanged and the HER could be determined from the difference
in the radioactivity between the introduced and recovered pulse or from the radioactivity of
the tritium incorporated into coal obtained by the combustion of the coal. The HER derived
from the balance of radioactivity in the tritiated gaseous hydrogen measured by the radioan-
alyzer was consistent with the HER derived from the conversion of tritiated coal into tritiat-
ed water the radioactivity of which was measured by a liquid scintillation counter.
The effect of the amount of Pt/A1203 catalyst on the HER between Illinois # 6 coal and
tritiated gaseous hydrogen at 250 ~ was investigated. As shown in Fig. 1.51, the HER of
coal increased linearly with increasing amount of catalyst to 25 mg, reached about 23% and
leveled off over 25 mg of catalyst. Further, the hydrogen exchange reaction between tritiat-
ed coal and water was performed in a batch reactor and its HER was also plotted in Fig.
1.51. The HER between tritiated coal and water also increased with increasing amount of
catalyst to 20 mg, reached approximately 8% and leveled off at over 20 mg of the catalyst.
This value agrees with the HER obtained in the hydrogen exchange of coal with tritiated
water which represents all the hydrogen in the functional groups in the coal (vide infra).
This indicates that the hydrogen of functional groups such as the hydroxy group in coals
was more easily exchanged than other hydrogen present in coal.
50
45
42
40
~- 30
20 19 ~ 19
17
10
Fig. 1.52 Change in hydrogen exchange ratio with particle size of Beulah-Zap at 250 ~
m: HER-Total (Coalc=>[3H]H2); m: HER ([3H]Coalc:~H20); Fl: HER (Coalc=~[3H]H20)
[From Ishihara, A. et al., Energy Fuels, 14, 710, (2000)]
1.5 Tritium Tracer Methods for Coal Characterization 61
gaseous hydrogen were performed. Two ND coal samples of different particle size, under
100 mesh and under 20 mesh, were used. Fig. 1.52 shows the effect of the particle size of
ND coal on the HER in the presence of lwt%-Pt/A1203 catalyst at 250 ~ As shown in the
Fig. 1.52, HERs between coal and gaseous hydrogen for under 100 mesh and under 20-
mesh samples were about 45% and 42%, respectively. When these tritiated coal samples
were reacted with water, tritiums in the functional groups were removed and the HERs cor-
responding to hydrogen exchanged in the functional groups could be estimated to be 17%
for the under 100-mesh and 19% for the under 20-mesh samples, respectively. Further, the
percentage of functional groups can be estimated by the hydrogen exchange reaction of coal
and tritiated water at 100 ~ and HERs estimated for under 100-mesh and under 20-mesh
samples were 19% and 17%, respectively. These results show that in the reaction between
coal and tritiated gaseous hydrogen almost all hydrogen in the functional groups exchanged
with gaseous hydrogen in the presence of the catalyst at 250 ~ and that the particle size of
coal showed no significant effects on the HER between coal and hydrogen or between coal
and water.
Similar results were obtained for IL and POC coals. In these cases, two kinds of sam-
ples of different particle size, under 100 mesh and under 20 mesh, and two kinds of cata-
lysts with different metal loading, l wt%-Pt/A1203 and 3wt%-Pt/A1203, were used. In the
reactions between coals and tritiated gaseous hydrogen at 250 ~ almost all the hydrogen
in the functional groups exchanged with gaseous hydrogen in the presence of the catalyst.
As shown in Figs. 1.53 and 1.54, very little if any significant effect of particle size of coal
on the HER between coal and hydrogen or between coal and water was observed. The
HERs between IL coal and tritiated gaseous hydrogen decreased only slightly with increas-
ing particle size of coal. This may be due to the fact that there is a similar distribution of
particle size in the two Argonne samples of under 100-mesh and under 20-mesh used. As
shown in Fig. 1.49a, b and c, the under 20-mesh sample includes a significant amount of
under 100 mesh particles for each coal. Therefore the effect of particle size on the HER
was further investigated at 250 ~ using Wandoan coal of under 250-mesh, 250-150-mesh
25
20
20
18
16 ~ 16
~, 15
7.8
7.1
Fig. 1.53 Comparison of hydrogen exchange ratio with par-ticle size of Illinois #6 at 250 ~
9 : HER-Total (Coalr 3H]H2); I1: HER ([ 3H]Coalc:~H20); D: HER (Coalc=>[3H]H20)
[From Ishihara, A. et al., Energy Fuels, 14, 709, (2000)]
62 1 Methods of Classification and Characterization of Coal
t
II
4.6
4.1
3.9
3.6
~3
2.4
7Z
1.9 ~1.9
2
0
- 100 mesh -20 mesh - 100 mesh -20 mesh
1%-Pt/m1203 50 mg 3%-Pt/AI203 50 mg
Fig. 1.54 Comparison of hydrogen exchange ratio with particle size of Pocahontas at 250 ~
I1: HER-Total (Coalc=~[3H]H2); I1: HER ([3H]Coalc:~H20); [-1:HER (Coalc=~[3H]HzO)
[From Ishihara, A. et al., Energy Fuels, 14, 710, (2000)]
and 2 0 - 2 2 - m e s h samples. Fig. 1.55 shows the results from the reactions of W A coal. It
was shown that particle size affects only slightly the hydrogen exchange reaction of coal
with tritiated gaseous hydrogen. The H E R between coal and tritiated gaseous hydrogen de-
creased only slightly with increasing particle size. Further, there was no significant effect
of particle size in the reaction of coal with tritiated water. These results suggest that most
of the hydrogen exchanged with gaseous hydrogen or water, e.g. hydrogen in functional
groups and a - c a r b o n of an aromatic ring, is located not only on the exterior surface but also
on the interior surface of coal, where hydrogen atom generated in the presence of the cata-
lyst or the proton generated from water can be easily approached, even when the particle
size of coal is larger. The exterior and interior surfaces of coal may represent the bulk en-
25
19
17.4
16
15
lO
0
-250 mesh 250-150 mesh 20-22 mesh
(-58 ~tm) (58 ~tm-96/,tm) (841-800 Ftm)
Fig. 1.55 Variation in hydrogen exchange ratio with particle size of Wandoan Coal at 250 ~
I1: HER-Total (Coalc=>[3H]H:); I1: HER ([3H]Coalc=~H20); IS]:HER (Coalc=>[3H]H20)
[From Ishihara, A. et al., Energy Fuels, 14, 710, (2000)]
1.5 Tritium Tracer Methods for Coal Characterization 63
velope particle surface and the interior surface may be accessible to the liquid or gas phase,
respectively. The result shows that there is very little if any exchangeable hydrogen that is
not on the exterior and interior surfaces of coal.
C. Reaction M e c h a n i s m s
The amounts of hydrogen exchanged in coal with gaseous hydrogen or water were lower
for the higher rank coals. Further, almost all hydrogen of functional groups in coals was
exchanged with gaseous hydrogen at 250 ~ in the presence of a catalyst because the HER
between tritiated coal and water was very close to the HER between coal and tritiated wa-
ter. It is proposed that the hydrogen in functional groups and the a-carbon of an aromatic
ring is primarily exchanged with gaseous hydrogen in the presence of a catalyst. Possible
mechanisms for the hydrogen exchange of coal with tritiated gaseous hydrogen in the pres-
ence of a catalyst are shown in Fig. 1.48, above. The reaction of coal with water proceeds
as shown in Fig. 1.40. Since hydrogen in the low rank coal is more labile in this reaction,
hydrogens in functional groups with a heteroatom and an a-carbon in the aromatic ring may
be related to the hydrogen exchange. Because there was very little if any significant effect
of particle size in the reaction of coal with tritiated gaseous hydrogen and tritiated water, it
is suggested that most of the exchangeable hydrogens in functional groups and the a-car-
bon of an aromatic ring are located on not only the exterior surface but also the interior sur-
face of coal or in the cross-linking position in the macromolecular network structure of coal
(Takanohashi et a1.,1999; Nakamura et al., 1995) where the hydrogen atom generated in the
presence of the catalyst or the proton generated from water can approach easily, even when
the coal is of larger particle size. The hypothetical linkage structure of coal is illustrated in
Fig. 1.56. Further it is suggested that there is no pore-like structure where it is difficult for
a hydrogen atom or a proton to approach and that the process, whereby a coal particle
breaks into smaller pieces involves a separation from the larger layer structure of the aro-
matic group or the breaking of cross-links such as hydrogen bonds in the macromolecular
network structure of coal, as shown in Fig. 1.56.
hydrogenbondingNx
~ @o~ --ro.H
CT3 ~ I U O T
Tritiated coal
Fig. 1.56 Hypothetical coal structure. A larger particle of coal, smaller particles of coal and tritiated coal.
[From Ishihara, A. et al., Energy Fuels, 14, 711, (2000)]
64 1 Methods of Classification and Characterization of Coal
case because there is no solvent in the system. However, hydrogen radicals may access ex-
changeable hydrogens passing over aromatic ring planes or aromatic hydrogens which may
play a role like a railway carrying the radicals.
Further, although the hydrogen exchange of hydrogen in aromatic rings may occur, it
will proceed only to very small extent. In Buchanan's experiments, the aromatic spacers
were inert while spacer molecules containing benzylic C-H bonds were reactive in their
radical relay reaction. On the other hand, HER reached the constant value with the addition
of 20 mg of the catalyst, and the further addition of the catalyst did not increase the HER.
These results suggest that although there is a large amount of aromatic hydrogen in coal,
the hydrogen exchange reaction of most aromatic hydrogen with hydrogen radicals may be
difficult to observe at least in the range 200-300 ~
In summary, the hydrogen exchange reactions of Argonne premium coals and
Wandoan coal with tritiated gaseous hydrogen and tritiated water were carried out to inves-
tigate the effect of the particle size of coal on the hydrogen exchange reaction. Coals of
several particle sizes ( < 100 mesh and < 20 mesh for Argonne Premium coals, < 250
mesh, 250-150 mesh, and 20-22 mesh for Wandoan coal) were used and it was found that
the particle size of coal only slightly affected the hydrogen exchange ratio of coal in both
cases of tritiated gaseous hydrogen and water. These results suggest that most of the hydro-
gen exchanged with gaseous hydrogen or water, i.e., hydrogen in functional groups, is lo-
cated on not only the exterior surface but also on the interior surface of coal, where hydro-
gen atoms generated in the presence of the catalyst or proton generated from water can easi-
ly approach, even in the case of coal of larger particle size.
1.5.5 E l u c i d a t i o n o f C a t a l y t i c H y d r o g e n E x c h a n g e R a t e in C o a l u n d e r
Reductive Atmosphere
The hydrogen exchange reactions of coal with tritiated water and tritiated gaseous hydrogen
have been described in the preceding sections. In these investigations, the amount of hy-
drogen in the functional groups in coal was determined and the mobility of hydrogen in
coal elucidated. Further, in these hydrogen exchange reactions between coal and tritiated
gaseous hydrogen using a flow reactor, a pulse of tritiated hydrogen molecule in N2 cartier
gas was activated on a Pt/A1203 catalyst to give hydrogen radicals which exchanged with
hydrogen in coal without hydrogen addition in the range 200-300 ~ (Kabe et al., 2000;
Ishihara et al., 2000).
In this section, the hydrogen exchange reaction between coal and tritiated gaseous hy-
drogen in the presence of a catalyst is examined to determine the hydrogen exchange rate of
coal. Three experiments were performed using a flow reactor to control temperature and to
determine precisely the mobility of hydrogen in coal. Initially a pulse of [3H]H2 was intro-
duced into a coal in H2 carrier gas at several temperatures, and from the pulse delay ob-
served the hydrogen exchange between the coal and [3H]H2 was estimated. Then, the
amount of hydrogen exchanged was determined from the radioactivity introduced into coals
in a N2 carrier gas. The tritiated coal formed in this reaction was reacted with gaseous hy-
drogen, which was introduced as the carrier gas at constant temperature. The hydrogen ex-
change rate was estimated from the release rate of [3H]H2. Finally, the reaction of tritiated
coal with gaseous hydrogen was performed during heat treatment and the change in the tri-
tium concentration was traced at the outlet of the reactor to estimate the behavior of hydro-
gen in coal.
In the estimation of the hydrogen exchange rate of coal with a pulse of tritiated gaseous
hydrogen using a flow reactor, the procedure is almost the same as that presented in the pre-
66 1 Methodsof Classification and Characterizationof Coal
ceding section, 1.53A. Coal samples used were Argonne Premium Coal Samples, Illinois
# 6 (IL), Upper Freeport (UF), and Pochahontas # 3 (POC). A tritium pulse tracer reaction
system was performed using the pulse flow reactor same as shown in Fig. 1.41. About 0.4
g of coal and 0.05 g catalyst were mixed in Ar gas and packed into the reactor. The flow
rate of carrier gas (H2) was 5 ml/min and the pressure was 15 kg/cm 2. The reactor was
heated and kept at the reaction temperature (100-250 ~ for 2 h to remove water. After
that, the tritiated H2 was introduced into a coal bed using a 6-way valve and reacted with
the coal. The radioactivity of [3H]H2 recovered from the reactor was directly monitored
with a radioanalyzer. The average residence time was calculated from the shape of the tri-
tium pulse. For a blank test, quartz sand was used and its pulse delay was compared with
those obtained for coal.
In Eq. (1.37), tritium is introduced into exchangeable hydrogen in coal then released again
from coal during time (Zcoal- Zbla,k) /2. Z~oaland Zblankare the average residence time in coal
and quartz sand, respectively. It was assumed that the rate of tritium incorporation into coal
was the same as the rate of tritium release from coal and that tritium was introduced into the
exchangeable hydrogen (Hex) or released from it during time (Z~oa~- "rblank)/2.
In another description of the hydrogen exchange reaction, the rate constant of tritium
release was estimated from the first-order plot of the radioactivity of tritium released in the
reaction of tritiated coal with gaseous hydrogen. The straight line of the first-order plot can
be represented using Eq. (1.38).
lnY-- - kt + lnZ, Y-- Z e -kt (1.38)
In Eq. (1.38), Y is the radioactivity of tritium released during a unit time (counts/min), k the
a rate constant of tritium release (l/min), t the reaction time (min), and Z the initial rate of
tritium release (counts/min). Further, the integral of Eq. (1.38) from 0 to ~ gives the total
radioactivity introduced into coal Rcoa~(counts) (Eq. 1.39). From Eqs. (1.37) and (1.39), the
initial rate of tritium release (Z) can be converted to the hydrogen exchange rate using the
amount of hydrogen Z • HgadRgas (Eq. 1.40).
5000
0 IO0~
~, 4000 II 150 ~
O A 200 ~
3000
9 250 ~
.~ 2000
1000
0 L
0 5 10 15 20 25 30
Time (min)
Fig. 1.57 Change in radioactivity of introduced pulse of tritiated gaseous hydrogen. (Illinois #6 Coal with 1%-
Pt/A1203) [Reproduced with permission from Ishihara, A. et al., Fuel, 81, 1411, Elsevier (2002)]
[Reproduced with permission from Ishihara, A. et al., Fuel, 81, 1411, Elsevier (2002)]
68 1 Methods of Classification and Characterization of Coal
exchanged in coal was calculated from the amount of tritium incorporated into coal, which
corresponds to the shaded area in Fig. 1.58, and the results are shown in Table 1.17. The
amount of hydrogen exchanged increased with increasing temperature for every type of
coal and decreased with increasing coal rank at every temperature except for POC at 200
and 250 ~ Using both values of the amount of hydrogen exchanged in Table 1.17 and the
residence time of tritium in Table 1.16, the apparent rate of hydrogen exchange reaction in
coal was estimated using Eq. 1.37 and the results are listed in Table 1.18. In the case of IL
coal, the apparent rate increased to a value about 1.4 times larger with a rise from 200 ~ to
250 ~ This is related to the fact that the amount of hydrogen exchanged increased to a val-
18000
16000
14000
o 12000
10000
8000
6000
4000
2000
0
1 2 3 4 5 6 7 8 9 10
Number of pulse ( - - )
IL 9.82 X 10 - 4 1.41 X 10 -3
UF 7.62 X 10 - 4 5.75 X 10 - 4 7.44 X 10 - 4
POC 9.19 X 10 - 4 8.16 X 10 - 4 5.89 X 10 - 4
[Reproduced with permission from Ishihara, A. et al., Fuel, 81, 1412, Elsevier
(2002)]
1.5 Tritium Tracer Methods for Coal Characterization 69
ue that was about 3.4 times larger. In contrast, the apparent reaction rates for UF and POC
coals decreased slightly. This shows that for these coals the increase in the amounts of hy-
drogen exchanged was only about 1.3 and 1.2 times, respectively, and that the effect of the
amounts of hydrogen exchanged was small. Further, it can be considered that less reactive
hydrogen, which did not affect the apparent rate at 200 ~ began to participate in the hy-
drogen exchange reaction at 250 ~ and decreased the apparent reaction rate at this temper-
ature. With a rise from 250 ~ to 300 ~ a similar decrease was observed for POC coal
while the apparent rate for UF coal increased because the amount of hydrogen exchanged
increased to a value that was about 1.9 times larger.
1200
O
~
9 400 _ O0 %0 D m%
,oo O ,-D_x3
CCr' CroO ~ '
200
D , , o o -~Y.IP,,P..~
I
Or.
0 1000 2000 3000 4000
Time (s)
Fig. 1.59 Variation in the amount of tritium released from IL coal under a reductive atmosphere.
[Reproduced with permission from Ishihara, A. et al., Fuel, 81, 1412, Elsevier (2002)]
70 1 Methods of Classification and Characterization of Coal
7 O 200 ~
6 9 250 ~
_= 5
2 J I I
0 1000 2000 3000 4000
Time (s)
lnY= kt + lnA; Y: Counts of radioactivity
Fig. 1.60 First-order plot of radioactivity released from Illinois #6 coal.
[Reproduced with permission from Ishihara, A. et al., Fuel, 81, 1412, Elsevier (2002)]
gen for lower rank IL coal decreased because a larger a m o u n t of less reactive h y d r o g e n was
related to the h y d r o g e n e x c h a n g e , c o m p a r e d with the cases of U F and P O C coal.
Using the a m o u n t of h y d r o g e n e x c h a n g e d in Table 1.17 and the rate constant of tritium
release in Table "1.19, the apparent rate of h y d r o g e n e x c h a n g e was estimated and listed in
Table 1.20. C o m p a r e d with the apparent rates in Table 1.18, which were calculated in the
e x p e r i m e n t using a pulse of [3H]H2 in H2 carrier gas, the apparent rates in Table 1.20 were
several times to one order of m a g n i t u d e lower. In the pulse m e t h o d , where m u c h smaller
a m o u n t of tritium is reacted with coal within a short time, tritium is difficult to introduce
into less reactive h y d r o g e n in coal, the effect of which is difficult to see in the apparent rate.
In contrast, w h e n coal is initially tritiated, tritium is introduced until the exchange reac-
tion reaches equilibrium and therefore less reactive h y d r o g e n also tritiated. In this case,
w h e n the carrier gas N2 is replaced by H2, the reaction of tritium in less reactive hydrogen is
observed to be very slow since there is a c o m p a r a t i v e l y large a m o u n t of tritiated less reac-
tive hydrogen. As a result, the apparent reaction rate b e c o m e s small. H o w e v e r , the trend in
the variation of the values in Table 1.20 is very similar to that in Table 1.18. For example,
Table 1.19 Rate Constant of Tritium Release in Hydrogen Exchange Reaction of
Coal with Gaseous Hydrogen (1/min)
Coal 200 ~ 250 ~ 300 ~
IL 4.2 X 10-2 3.0 X 10-2 m
UF 1.9 X 10-~ 1.3 X 10-I 1.5 X 10-~
POC 1.9 X 10 -~ 1.4 X 10-! 8.4 X 10-2
[Reproduced with permission from Ishihara, A. et al., Fuel, 81, 1413, Elsevier (2002)]
[Reproduced with permission from Ishihara, A. et al., Fuel, 81, 1413, Elsevier (2002)]
1.5 Tritium Tracer Methods for Coal Characterization 71
700
i
i
"~ 300
-.,,e.
O No
200
100
Fig. 1.61 Variation in the amount of radioactivity released from IL #6 coal under a reductive
atmosphere at 3 ~ [Reproduced with permission from Ishihara, A. et al., Fuel, 81,
1413, Elsevier (2002)]
with a rise from 200 ~ to 250 ~ the reaction rate for IL coal increased while those for UF
and POC coals slightly decreased. This shows that the increase in the amount of hydrogen
exchanged affected the increase in the reaction rate in the case of IL coal while for UF and
POC coals the participation of less reactive hydrogen in the exchange reaction began, slow-
ing the rate. With a rise from 250 ~ to 300 ~ a similar tendency was observed for POC
coal, and the apparent rate decreased while the increase in the amount of hydrogen ex-
changed for UF coal increased the apparent rate.
The above results indicate that less reactive hydrogen clearly participated in the ex-
change reaction under a reductive atmosphere where H2 is used as a carrier gas. To exam-
ine this in detail, the reaction of tritiated coal with gaseous hydrogen (carrier gas) was per-
formed during heat treatment at a heating rate of 3 ~ and the change in the radioactivi-
ty of tritium with reaction time was traced. The heat treatment was performed to up to 300
~ for samples tritiated at 200 and 250 ~ and to up to 380 ~ for a sample tritiated at 300
~ respectively. After that the maximum temperature was maintained for each case for
several hours. As shown in Fig. 1.61, the number of peaks changed depending on tempera-
ture where the sample was tritiated. When the samples were tritiated at 200, 250, and 300
~ one, two and three peaks were observed respectively. These results show that there are
at least three kinds of hydrogen of different reactivity. It has been reported that when a
sample was tritiated at 200 ~ most of the tritium was introduced into functional groups
(Kabe et al., 2000). This result shows that the hydrogen in functional groups is the easiest
to exchange among all hydrogen in coal. Therefore, with a sample that is tritiated at 200
~ it is likely that the hydrogen of functional groups is the main participant in the exchange
reaction. Further, it has been reported that in reactions involving radicals hydrogen at the
benzyl position is easy to exchange (Ishihara et al., 1995, 1999; Collin and Wilson, 1983;
Skowronski et al., 1984). For example, a hydrogen of tetralin was easiest to exchange
among a, fl and aromatic hydrogen of tetralin in the hydrogen exchange reaction between
deuterated tetralin and coal. Therefore, tritiated at 250 ~ in the experiment, hydrogen at
the benzyl position as well as functional groups seemed to participate in the reaction.
Tritiated at 300 ~ even aromatic hydrogen may participate in the reaction. In this last ex-
72 1 Methods of Classification and Characterization of Coal
actions were performed under the conditions 200-300 ~ and 5-360 min. After the reac-
tion, coal and tetralin were separated under vacuum at 200 ~ for 2 h and the tritiated coal
was washed with n-hexane, dried and oxidized by an automatic sample combustion system
into tritiated water to measure its radioactivity. Every tetralin sample was dissolved in a
scintillator solvent and measured with a liquid scintillation counter. The tetralin sample
was also analyzed by a gas chromatography equipped with FID. In order to predict the lo-
cation of exchangeable hydrogen in coal, the hydrogen exchange reactions of coal that re-
acted with tritiated tetralin, i.e., tritiated coal, with water were performed.
Calculation of Hydrogen Transfer Ratio: Hydrogen transfer includes both hydrogen ad-
dition and hydrogen exchange. The hydrogen transfer ratio (HTR) means the ratio of the
amount of hydrogen transferred into coal (Htr) to the total amount of hydrogen in an origi-
nal coal (Hcoa0. HTR between coal and tetralin was calculated using Eq. (1.41):
20
9 TotalHTR
II HTR with-OH [Coal]
15
lO
m
5
0 , I I I
0 60 120 180 240 300 360 420
Reaction time (min)
Fig. 1.62 HTRsof Illinois #6 coal with [3H]Tetralin at 300 ~
[From Ishihara, A. et al., Prepr. ACS. Div. Fuel Chem. 44 (3), 660 (1999)]
74 1 Methodsof Classification and Characterization of Coal
gen exchange between hydrogen in the functional group of coal and tetralin increased with
reaction time and reached a constant value at 180 min similar to the total HTR.
In Fig. 1.63, W A coal was used instead of IL coal. Total HTR also increased with re-
action time and reached constant value at 180 min. After the hydrogen transfer reaction,
the hydrogen exchange between the tritiated coal sample and water was also performed to
remove tritium in the functional group of coal. As shown in Fig. 1.63, HTR corresponding
to the hydrogen exchange of functional group in coal with tetralin also increased with reac-
tion time and reached a constant value at 180 min. Since these kinds of reactions reached
equilibrium after 180 min, the reactions of other coals were performed for 180 min.
Figure 1.64 shows the change in HTR of coal with reaction temperature at 180 min.
Although HTR was observed for each coal at 200 or 250 ~ this may be due to the sorption
of the tetralin molecule into coal as well as the hydrogen exchange. HTR increased re-
20
9 Total HTR
[.., 10
=: Q
5
m
I I I I I I
00 60 120 180 240 300 360
Reaction time (min)
Fig. 1.63 HTRs of Wandoan coal with [3H]tetralin at 300 ~
16
[] Beulah Zap
14
9 Wandoan
12 9 Illinois #6
10 9 Upper Freeport
9 Pocahontas
[..
6 -
4-
2-
0 I I I
markably in most coals with increasing temperature from 250 ~ to 300 ~ while HTR for
POC coal increased only slightly. Total HTRs for all coals and HTRs corresponding to hy-
drogen exchange of the hydrogen of the functional group in coal with tritiated tetralin are
listed in Table 1.21. The result shows that at lower temperatures of 200 and 250 ~ hydro-
gen exchanges between hydrogen in the functional group and tetralin were very low. These
results were significantly different from those for the reaction of coal with tritiated gaseous
hydrogen, where most of the hydrogen in functional groups exchanged with tritiated
gaseous hydrogen at the same lower temperature in the presence of the catalyst (Kabe et al.,
2000). However, HTR for functional group of coal with tritiated tetralin also increased re-
markably with a rise from 250 ~ to 300 ~
Figure 1.65 shows the variation in tetralin conversion to naphthalene with temperature
at 180 min. The trend of increase in tetralin conversion with temperature was similar to
that for HTR. In the cases of lower rank coals ND and WA, tetralin conversion increased
with increasing temperature from 200 to 300 ~ In the cases of higher rank coals IL, UF
and POC, tetralin conversions were very low at temperatures 200 and 250 ~ However,
these increased remarkably with a rise in temperature from 250 to 300 ~ The result sug-
gests that, since lower rank coals ND and WA generated larger amounts of radical species
even at lower temperatures, tetralin conversion to naphthalene, i.e., hydrogen addition into
coal, was enhanced. For the higher rank coals, however, it seems that large amounts of rad-
Table 1.21 HTRs of Coals with Tritiated Tetralin
a Hydrogen transfer ratio of functional groups such as hydroxy group which was determined by the reaction of tri-
tiated coal with water at 100 ~ [From Ishihara, A. et al., Prepr. ACS. Div. Fuel Chem. 44 (3), 660 (1999)]
10
[] Beulah Zap
/
8 - 9 Wandoan ~,
A Illinois # 6 Ld~~
//
"~
o 6 - 9 Upper Freeport / /
J/ w
o O Po
._.= 4 -
[-
2-
0 T
Fig. 1.65 Effect of temperature on tetralin conversion for the reaction of coals with tetralin for 3h.
[From Ishihara, A. et al., Prepr. ACS. Div. Fuel Chem. 44 (3), 661 (1999)]
76 1 Methods of Classification and Characterization of Coal
25
[] Total HTR of coal with [3H]Tetralinat 300 ~ for 3h
[] HTR of-OH [Coal] with [3H]Tetralinat 300 ~ for 3h
20
[] HTR of-OH [Coal] with [3H]H20 at 100 ~ for 6h
15
[.,
10
__z_.~ I
ND WA IL UF POC
Coal rank
Fig. 1.66 Comparion of HTRs of coal with tetralin and water.
[From Ishihara, A. et al., Prepr. ACS. Div. Fuel Chem. 44 (3), 661 (1999)]
e~0
>, 2
Ii~!~i
d::
Q
=1
o 1
E
<
ND WA IL UF POC
Coal rank
icals were generated at 300 ~ and increased the conversion of tetralin. Tetralin conversion
for IL coal was highest at 300 ~ and this m a y be due to the catalysis by a larger a m o u n t of
pyrite included in IL coal.
In Fig. 1.66, total H T R of coal with tritiated tetralin and H T R c o r r e s p o n d i n g to hydro-
gen e x c h a n g e for the functional group in the reaction with tritiated tetralin at 300 ~ and
360 m i n were c o m p a r e d with the H T R c o r r e s p o n d i n g to h y d r o g e n e x c h a n g e b e t w e e n coal
and tritiated water at 100 ~ w h i c h is described above. Except for N D coal, H T R corre-
sponding to h y d r o g e n e x c h a n g e for the functional group of coal in the reaction with tetralin
was very similar to the H T R of the functional group b e t w e e n coal and tritiated water. The
1.5 Tritium Tracer Methods for Coal Characterization 77
results shows that at 300 ~ most of the hydrogen in the functional group of coal exchanged
with tritiated tetralin. In the reaction of coal with tritiated tetralin, HTR decreased in the or-
der IL > UF > WA > ND > POC, unlike the reaction of coal with tritiated water where
HTR decreased with increasing coal rank (Kabe et al., 2000). The decomposition of the
coal structure in lower rank coals at 300 ~ may affect the hydrogen transfer.
Figure 1.67 shows the variation in the amounts of hydrogen transferred to coal with
coal rank. The bar on the left shows the total amount of hydrogen transferred from tetralin
to coal, calculated by the amount of tritium transferred in coal. The bar in the middle
shows the amount of hydrogen addition calculated by the conversion of tetralin to naphtha-
lene. The bar on the fight side shows the net amount of hydrogen exchanged, indicating the
difference between the amount of hydrogen transferred and amount of hydrogen addition.
When this net amount of hydrogen exchanged was compared with the HTR of the function-
al group in Fig. 1.40, the net amounts of hydrogen exchanged for lower rank coals such as
ND, WA and IL correspond to the hydrogen exchange of the functional group. In contrast,
for higher rank coals such as UF and POC, the hydrogen exchange of the functional group
was only part of the net hydrogen exchange because the HTR of the functional group was
much less than the total HTR of coal, and the net amount of hydrogen exchange is relative-
ly large.
On the other hand, the result also shows that for lower rank coals hydrogen radicals
generated from tetralin combined with radicals formed by the decomposition of lower rank
coals, and that as a result hydrogen addition occurred. However, for higher rank coals, be-
cause the amount of radicals formed by the decomposition of coal is less than in the case of
lower rank coals, the hydrogen radical generated from tetralin contributes to the hydrogen
exchange rather than to hydrogen addition.
10
~, 6
I 9
m
m
m
I I I I
0 90 180 270 360 450
Reaction time (min)
a Hydrogen exchange ratio of functional groups such as hydroxy group which was determined by the reaction of
tritiated coal with water at 100 ~ [Reproduced with permission from Ishihara, A. et al., Prospects for Coal
Science in the 21th Century, 1,204, Shanxi Science & Technology Press (1999)]
10
f..rd
4 3.4
ND WA IL UF POC
Coal rank
Low High
Fig. 1.69 HERs of coals with tritiated toluene at 300 ~ for 180 min without a catalyst.
[Reproduced with permission from Ishihara, A. et al., Prospects for Coal Science in the 2 l th Century,
1,205, Shanxi Science & Technology Press (1999)]
1.3 'lritium Tracer Methods for Coal Characterization 79
hydrogen at the same lower temperature in the presence of a catalyst (Kabe et al., 2000;
Ishihara et al., 2000).
HERs of various coals with tritiated toluene at 300 ~ and 180 min are compared in
Fig. 1.69. The dotted line shows the HER value of IL #6 coal with tritiated toluene at 200
~ which may be due to the sorption of toluene into coal. In the reaction of coal with triti-
ated toluene, the HER decreased in the order IL > UF > ND > POC > WA, which was
very similar to the cases with tritiated tetralin except for WA coal, but different from that of
the reaction of coal with tritiated gaseous hydrogen or tritiated water where HER decreased
with increasing coal rank (Kabe et al., 2000; Ishihara et al., 2000). Decomposition of the
coal structure or the functional group may occur in the lower rank coals at 300 ~ decreas-
ing the hydrogen exchange in the cases of toluene and tetralin.
2.1 Introduction
Coal is an aggregate of heterogeneous substances composed of organic and inorganic mate-
rials. The organic materials are derived mainly from plant remains which have undergone
various degrees of decomposition in the peat swamps and physical and chemical alteration
after burial. The bulk of all coal deposits were formed in a peat swamp environment where
different types of vegetation flourished, reflecting primarily conditions of climate, water
level, and water chemistry of the swamp. In addition, evolution in the plant kingdom
played an important part in the makeup of the plant source material in the swamp, ultimate-
ly affecting the petrographic composition of the coal. Coals forming in a peat swamp envi-
ronment are essentially autochthonous or in situ in origin. On the other hand, some peat
and even coals may be reworked and redeposited in a fluvial system and are called au-
tochthonous in origin. Coal deposits derived from extensive accumulation of driftwood
also belong to this category.
Optically homogeneous discrete organic material in coal is called maceral (derived
from Latin macerare, to macerate, to separate). There are three major groups of macerals:
the vitrinite, liptinites (exinite), and inertinite groups. Inorganic materials in coal consist
primarily of mineral matter, chiefly clay minerals, quartz, carbonates, sulfides, and sulfates,
as well as many other substances in very small quantities (see Chapter 1). The total bulk of
inorganic constituents in coal range from a few percent to more than 50%. If the inorganic
constituent is > 50%, it is classified as carbonaceous shale. The threshold between coal
and noncoal is difficult to define, although Schopf (1956) has classified aggregates contain-
ing organic material amounting > 50 wt% and 70 vol% as coal.
Its genesis alone suggests that coal is a nonhomogeneous material. Coal, being an or-
ganic sedimentary rock, is composed of fossilized plant remains, which, in analogy to min-
erals, are called macerals, and of mineral occlusions. Macerals show distinctive areas un-
der the microscope and are differentiated into three major groups. Vitrinite is the most
prevalent group, accounting for 80%, and is believed to be derived from woody plant mate-
rial (mainly lignin). Exinite (alternatively called liptinites) is said to have developed from
lipids and waxy plant substances. Char formed by prehistoric pyrolysis, e.g., during wood
fires, is suggested as a possible origin of inertinites, the third maceral group. Fig. 2.1 shows
a microscopic view of a bituminous coal (Franck and Knop, 1979).
It is generally agreed that coal is predominantly of vegetal origin (Van Krevelen,
198 l a; Berkovitz, 1985; Grainger and Gibson, 1981, Haenel, 1992). Fig. 2.2 shows a very
simplified representation of coal genesis. First, the debris of sinking swampy forests
formed large peat deposits (biochemical phase) via bacteria action. In the later geochemi-
cal phase, which extended over several hundred million years, the peat underwent coalifica-
tion under the influence of pressure and temperature caused by overlying sediment, forming
82 2 Chemical and Macromolecular Structure of Coal
Fig. 2.1 Microscopic view of a bituminous coal. (Franck and Knop, 1979) [Reproduced with permission from
Franck, H.G. and Knop. A., Kohleverding-Chemie und Technologie, 14, Springer (1979)]
Swamp
lignite and thereafter, with increasing coalification, bituminous coal and anthracite. Coal is
therefore regarded to be an organic sediment.
The elementary composition changes with increasing coal rank (Table 2.1). The car-
bon content, amounting to roughly 55% in peat, increases to more than 92 wt% in an-
thracite, whereas hydrogen, initially at 10 wt%, drops to below 3 wt%, and oxygen, initially
35 wt%, to finally 2 wt%. Very little sulfur and nitrogen (only a few percent) are present,
and the change in their concentration with coalification is less significant. The proportion
of aromatic carbon atoms Car/Ctot, 0.5 in lignite, increases to above 0.95 in anthracite.
The changing elementary composition is also reflected in the hydrogen/carbon ratio, which
2.2 Chemical Structure of Coal 83
Coalification
Plant ~ Peat -- Lignite ~ Bituminous coal -- Anthracite ..... ,,- Graphite
C (%) 55 70 8 0 - 90 92
H(%) 10 8-5 6-4 3
0(%) 35 25 10-5 2
Car/Ctot --0.5 0.6 ----- 0.95
H/C 1 ~- 0.5
[Reproduced with permission from Haenel, M.W., Fuel, 71, 1212, Elsevier (1992)]
lyst, of extract (bitumen) and residue (fumin), which were obtained in a benzene extraction
of a coal at 260 ~ were carried out under the same condition (Biggs and Weller, 1937).
When the liquid product obtained in the pretreating above was fractionated, the similar dis-
tribution of molecular weight, the boiling point, a refractive index, hydrogen-carbon ratio,
etc. were observed for the oil originated from both the extract and the residue.
Taken together, these lines of evidence leave little doubt to consider coal to be the ag-
gregate of the material which is of similar structure. Therefore, it becomes possible that
coal is dealt averagely and statistically even in the form of mixture.
As mentioned above, we should think that the structure of the coal is not of the particu-
lar chemical constitution in the viewpoint of organic chemistry, but is a polymer of a cer-
tain molecular weight via polymerization of the monomer of similar structure. Generally, it
is the corresponded opinion that the basic frame structure of the coal, which is equivalent to
the m o n o m e r in polymer, is made of ring-structure based on the statistical, physical and
chemical research result mentioned above. Though there are two types of ring structures,
an aromatic ring and fat ring structure, the former is the subject in coal. The ring structure
of the coal is thought to be the condensed ring structure in which fat ring structure linked to
aromatic ring. The ring structure, which contains oxygen besides such a carbon frame ring
also exist.
The size of the ring determined using the statistical, physical and chemical method is
summarized in Table 2.2 (Kimura and Fujii, 1984). Though the number (size) of the ring
shows different value by the method, it is reasonable to consider that the average number of
the ring in coal with 80-85 C% is between 4 and 5. Aromaticity (fa), which is used to esti-
mate the amount of aromatic ring in coal as an index, and is determined by various meth-
ods, is 1.00 in the case of the graphite constructed from the aromatic tings completely. Fig.
2.3 (Whitehurst, 1978) shows the value Offa, which was determined using a solid 13C-NMR
spectrum. Values offa for the coal of 8 0 - 8 5 C% are 0.6-0.7, meaning that 6 0 - 7 0 % of car-
bon in the coal are aromatic one. It is concluded that an aromatic ring is main in the frame
structure of the coal, and aliphatic carbon containing aliphatic ring is cross-link to the aro-
matic ring. Further, the ratios of carbon in aromatic ring show an increase tendency with
increasing carbon content in the coal.
Table 2.2 Numbersof Ring in Coal Determined by Various Methods
Method C (%) Reference
80 85 90 95
Statistical and physical methods
X-ray 4- 5 4- 5 7- 30- Hirsch, 1954
X-ray >4 >4 >4 - Nelson,1954
Density 15 17 24 26 Dryden,1958, 1962
Refractive index 6 9 16 31 van Krevelen and Chermin, 1954
Combustion heat 4- 9 10 18 Dryden,1958, 1962
Combustion heat - peri 3.3 - 4.3 peri 5.3 - 5.8 - Dryden,1958, 1962
cata 3.8 - 5.6 cata 9.5 Dryden, 1958, 1962
IR - 6.8 6.8 - Dryden,1958, 1962
NMR - 3.2 - 4.5 4.3 - 5.5 - Dryden,1958, 1962
Magnetic coefficient 2 3 5 - Hondaand Ouchi, 1957
Chemical methods
Hydrocracking 6 a Schuhmacher, et al., 1956
Hydrocracking 2-6 Le Claire, 1941
Hydrocracking 1 - 5b(Ave.3) Sakabe, 1961
Oxidative decomposition 1-3 Entel, 1954
Oxidative decomposition 1-3 Montgomery et al., 1956
aSample of Coal 986.5C% 9 b Sample of Coal" 84C% [Reproducedwith permission from Kimura, H. and Fujii, O.,
Chemistry and Industry of Coal, 174, Sankyo Pub. (1984)]
2.2 Chemical Structure of Coal 85
1O0 Anthracite o
90
80
60
50
Bituminous
Fuming acid o
o=/S o
40 ~ o
Subbituminous coal
30
o Wood
20 . . . . . . . . . .
45 50 55 60 65 70 75 80 85 90 95
C (maf%)
Fig. 2.3 Values off, in coal determined by means of solid NMR. [Reproduced with permission from
Whitehurst, D.D., Organic Chemistry of Coal, 71, 8 (1978)]
O O ~.. iO
O
~ ~ O
H
C135H9709NS
Fig. 2.4 Structure model of coal. [Reproduced with permission from Fuchs, W. and Sandohoff, A.G.,
Ind. Eng. Chem. 34, 570 (1942)]
86 2 Chemical and Macromolecular Structure of Coal
Given (1960) proposed a model from the experiment of the dehydroaromatization and
infrared spectrum. Then, the 9, 10-dihydroanthracene type is changed to the 9, 10-dihy-
drophenanthrene type, as shown in Fig. 2.5. The feature of this model is that aromatic nu-
clear mutuality as the naphthene ring can specially fit the 9, 10-dihydrophenanthrene type
combination.
In 1966, Kurogawa et al. (1966) proposed a model, as shows in Fig. 2.6. The feature of
the model is different from the model until former times by the thing which the basic struc-
H
HO OH O/ 0
9 H2 ./~ ~ CH2/~/,~ _.cn. ~ / ~ C H , H2 H2
H 2 ~ H E ~ N ~ ~ c ,,j(i i ~ C H ~ H H H 2 H2
2C'CHE~~H 2
H 3 C / ~ CH2 H2n ~ 2 H2
Fig. 2.5 Structure model of bituminous coal. [Reproduced with permission from Given, P.H.,
Fuel, 39, 150, Elsevier (1960)]
H CH~ H
OH
I H
~///~""'~9~
.?, ~.,-c- c.-
~ %c.,c ~ / 7 ~~c
' c~.
H--
~C. /CH~
] O
~, II i
CH 1,C~ /C~ CH2 H
'.c~S.). CH~..~.... dH
"V/~ C'c~,
"~" c k~\\W ~ COHIcgC
~., /c I,cH,.H.
/ \ I OH
[ Small ~ CH2 / H H C~2 H2
~ molecules ] C[__ CH_ CH_ .Cr162
\ / I Ha'///Za..a'~/Z~H
H ~ / OH CH2 ~ ~ "~C~"
' '
.C. CH ~ / ~ H H
H~~C'e/~H H H~~~H ~
HCW@a -o. "c.,
CH.,
H IH H CH2
CH,-- (7I HC~/~,/'~,C-CH-dH--
CH,
_c.-U/~C..b. C=H,o.~O,.6R - 3
CH2
Fig. 2.6 Structure model of Yubari coal. [Reproduced with permission from Kurogawa, M. et al.,
Coal and Coal Chemistry, 213, Nikkan Kogyo Shimbunsha (1963)]
2.2 Chemical Structure of Coal 87
ture unit which is mainly made of the aromatic ring is combined each other with methylene
and ether etc., and the combination between the units is not as strong as that in other models
and small molecules are also involved in frame structure.
A widely known model of a bituminous coal proposed by Wiser (1975) is shown in
Fig. 2.7, which is illustrated by more flexible linkages of ether, sulfide, and carbon-carbon
bridges together with numerous functional groups. This model has been shown to be rea-
H H
6 ~ H
H 6
H" ~ " ~1 H 6 N CH3 H
~"'r c~ o ----k.~~,. ~~ ~r c
. cr=o T _c...
H OH NH2 I I I .1. S __"&'O
H H-C -H H
H-C-H H H H I H- , ,~",.~CH3
I I I I II I "iN" ~..-. i
H H ~ 3- 3L /.3 u__ H O H H
~- Tt ~" Y~ Y,--,H " T~V " ( Y
TT i \ d/ ~ ~ r~.,......"~.~tJ
' -""
H- C-H t .~/ 1-12 112 C H I II II "l
I H ] I ....t.~ ~ /3,
H C H H~S'~H H-C-H rr ~ ~ "H
-I- ~ "'o I ," " .. ~
.-c-. " , " -~"r-~-r~ ~
:~ H C" T T "d \ / \ /
I I I II I r-r __"_C"__'ta / \ / \
,_.,,.J~/1,,~/J.~ I ~ ~ , , .H H ' ' H H H H H
i
.
i
~
H-C-H Hz H2 H-(~- H
H2 ~ OH
H-C- H H - C - CH3
CtS H-(~- H H2H-(~- H H2
ll3"O'~ ! O H2 ~ " Q , . ~ O*~'~.. IH2
"~_ , SH L. iL..1 2. U
CH3 H - o . ' H H2 ~ "IN" ",if
H-C- H H2
sonable for explaining some pyrolytic behavior of coals by introducing labile bridges into
the structure (Davidson, 1982). The coal structure proposed by Solomon (1981a) shown in
Fig. 2.8 is based on data from FTIR, NMR, elemental analysis, gel permeation chromatog-
raphy, and thermal decomposition data for a high-volatile bituminous Pittsburgh seam coal.
This model attempts to describe the pyrolytic reactivity of a coal by containing linkages
which are chemically reactive under pyrolysis conditions to produce gases and tars. The
sizes of the aromatic clusters were estimated from the gel permeation chromatography of
coal tars. The relationship of the aliphatic and hydroaromatic constituents in the structure
to the behavior of coal during pyrolysis was also specifically addressed. Tar yields predict-
ed from this model are close to experimental results.
Shinn (1984) proposes a more complex reactive model of coal structure. Shinn's mod-
el of a vitrinite-rich high-volatile bituminous coal is shown in Fig. 2.9 with a molecular
weight of 10,000. The model was based on detailed chemical analyses of both coal and
products from various liquefaction schemes in terms of elemental distribution, aromaticity,
functional group chemistry, and reactivity. This model was assembled from the knowledge
of reaction chemistry which reflects the amount of various structures in coal, the functional
groups that connect these structures, and how these functional groups react during liquefac-
tion. The model considers the mobile phase in coal, two-stage liquefaction yields, the
three-dimensional nature of coal, maceral chemistry, the geochemical origins of coal, and
response to pyrolysis and gasification reaction chemistry and aliphatic constituents.
A model describing the pyrolysis and hydropyrolysis behavior of a brown coal pro-
HO
OH
A~.,,~ OH
HBC OH
HO
HO
OH
CH3
z,,~_ OH
HO
OH
m HO
.o ou
. t . .L .t. OH OH
posed by Huttinger and Michenfelder (1987) is shown in Fig. 2.10. This model was devel-
oped from results of elemental analyses, pyrolysis experiments, titration studies, and ex-
trapolation of literature data for similar coals. One, two, and three aromatic tings and long-
chain aliphatic moieties identified from pyrolysis products are predominant features of the
structural unit. Note the large number of inorganic cations replacing hydrogen in phenolic
and carboxylate groups. The important point in this model is that approximately 10% of
the H atoms are replaced by cations, thus compensating for the hydrogen deficiency en-
countered in constructing low rank coal models. Polymeric models for low rank coals relat-
ed to conversion are reviewed by McMillen et al. (1990).
OH OH ONa
0,,_ OH ~ A ~'~ -C.~ OH OH
0 0,, OH C HOT('- ~ "]"('--"~"]"""~0 .L .,k . . . . .
H3C ~ OH KO "C" ~ ~__...~,,,,/ ~ (~""~y g g V ~ "C15H3,
Fig. 2.10 Model of a brown coal, comprising C270H240N351090. [Reproducedwith permission from Htittinger,
K.J. and Michenfelder, A.W., Fuel, 66, 1165,Elsevier (1987)]
Though some chemical constitution models of the coal have been introduced, these are
so far completely shown as two-dimensional figures. These molecular structures, while
providing much helpful information on the chemical nature of coal, do not provide data on
the three-dimensional structures or the intercluster interactions that provide the basis for
many of the physical properties of coal. For example, the glassy nature of coal, the glass-
to-rubber transition that coals go through when heated, and the nature of coal-solvent inter-
actions are not explained adequately by knowledge of the coal molecular structure alone.
To obtain a more complete picture of the physical characteristics of coal related to these
molecular structures, Spiro (1981) constructed three-dimensional representations of the
Given (1960), Wiser (1975) and Solomon (1981) structures using space-filling physical
models.
Spiro (1981) suggests that three-dimensional space-filling models of coal structure
should be considered along with other parameters. Of the four structural models studied,
only the Solomon model could be built without alteration. The others contained sterically
inaccessible moieties. Based on the molecular models, a mechanism for the thermal de-
composition and plasticity was proposed by Spiro (1981), which takes into account the
thermolysis of aliphatic, alicyclic, and hydroaromatic groups which protrude from the aryl
planes and act as spacers and lubricants allowing the parallel aryl planes to flow in two di-
mensions.
Spiro and Kosky (1982) propose space-filling models for low-, intermediate- and high-
rank coal molecules shown in Fig. 2.11. The molecules are designed to conform to experi-
mentally determined parameters such as chemical composition, aromaticity, and ring index.
The low rank coal appears fluffy, porous, and random, with most interior atoms exposed as
surface. The intermediate rank coal model is more fiat and oriented, with closed pores and
90 2 Chemicaland MacromolecularStructure of Coal
fewer noncoplanar protrusions due to aliphatic, alicyclic, and hydroaromatic moieties. The
high rank coal model is of higher order with graphitic domains. Physiochemical properties
with respect to the coal models are further discussed in Spiro and Kosky (1982).
HH
H {
H20-HO 0 [-0 0
H (
H <
.o
HI
<"
H H H
. . . . -
H ' H
H H
tt H
tt
H"
Because of steric problems, most of the molecular structures had to be altered some-
what before they could be constructed. Spriro's models provided insight into the interclus-
ter interactions and the degree of ring alignment (stacking) that might occur in bituminous
coal. These observations led to a plausible explanation for coal plasticity based on sliding
of ring systems, facilitated by small aliphatic fragments released in the early stages of py-
rolysis reactions. A limitation of these space-filling model studies was that the energetics
of the various structures and structural conformations could not be determined. Thus, the
relative probability of the various three-dimensional structures was not established.
Recently, with the development of molecular modeling software (Fruhbeis et al., 1987),
it has become possible not only to visualize molecular structures in three dimensions (on
the computer screen), but also to calculate energetically favorable structural conformations
using molecular mechanics and molecular dynamics methods. Molecular modeling tech-
Table 2.3 Molecular Parameters and Calculated Energies for Bituminous Coal Models
Type of coal "low rank . . . . bituminous" PSOC 170 Vitrinite-rich high-vol bituminous
No. of atoms 192 393 396 1311
Molecular weight 1492 2967 3020 9956
Car/Ctot 0.66 0.70 0.74 0.71
Har/(Har+Hal) 0.21 0.28 0.40 0.34
Weight fraction
C 0.820 0.782 0.823 0.789
H 0.053 0.059 0.056 0.057
O 0.107 0.113 0.090 0.119
N 0.019 0.014 0.009 0.015
S 0.032 0.021 0.019
Normalized formula C looH7709.8N2.0 C 100H900lo.9N1.6S1.6 C looH8108.2Nl.oS1.o ClooH87011.3N1.75o.9
Normolized energy of minimized
structure (kcal/atom) 2.07 1.78 1.75 1.65
[Reproduced with permission from Carlson, G.A., Energy Fuels, 6, 775 (1992)]
92 2 Chemicaland MacromolecularStructure of Coal
niques are being used widely today to provide insight into the structure, properties and in-
teractions. The principle of the molecular dynamics method has been published by Burkert
and Allinger (1986). The examples to apply the molecular dynamics method in the con-
struction of three-dimensional model is introduced below.
Carlson (1992) computes the model structure of bituminous coal using the force field
methodology DREIDING, which is a very general force field that can be used for a large
number of atom types. The three-dimensional models of four model structures proposed by
Given, Wiser, Solomon and Shinn models for bituminous coal were constructed using BI-
OGRAF, a molecule design software program allowing construction, visualization and en-
ergy calculations. The minimum energy of the physical structure was calculated using mol-
ecular dynamics calculation and energy minimization after the model construction.
Computer-modeling results for the Shinn, Wiser and modified Solomon models showed
similar structural folding because of van der Waals interactions or the driving force of a
noncovalent bond such as the hydrogen bond. In contrast, the Given model, because of the
stiff linkages between molecular clusters, was unable to fold into a compact structure even
when energy minimization was finished because aromatic structures combine forcefully
with each other by the methylene bridge. The results of the calculation are shown in Table
2.3 (Carlson, 1992). The most important difference is the number of aromatic hydrogens
for four kinds of model structures. The value in the Given model is lower than those in the
other models and is probably wrong when it is compared with a recent FT-IR result
(Solomon, 1979). It is suggested that the Wiser, Solomon and Shinn models for the coal
structure are more appropriate models because of their energy stability and the flexibility of
the structure.
As mentioned, Carlson's report was introduced easily. The molecular structure of the
coal is very important to appreciate reactivity of the coal. The molecular dynamics tech-
nique can be applied to the models of coals of different rank of coalification. Information
about the aggregation between the molecules in the coal can be obtained from these struc-
ture models.
Cross-links
i ~ - - -
29 n ~
,v, "x~_. ", ~'--,, .,~:'---------~
. - - . . , ,,,/,,----
,""%~-----r ~ 'Liquid structure'
,-"'b-..~-""'b~,. \ ~ "%, ~ / Y ; 20 nm (89% C)
"~,----~.'~.~--,
' ' ,~_.._._
"~r~ j r]08nm
"
Single layer ,,
+ amorphou~_ _-" '~''~" 0~8 nm
material
11,.
Fig. 2.12 Model of coal ranks. [Reproduced with permission from Hirsch, P.B.,
Proc. Roy. Soc., A., 226, 157 (1954)]
94 2 Chemicaland MacromolecularStructure of Coal
tection has not been assessed, although Hirsch pointed out the re-Jr stacking in his model, as
shown in Fig. 2.12. This figure illustrates Hirsch's classic model of coal ranking (Cartz and
Hirsch, 1960). Fewer and smaller aromatic rings, more alkyl- and oxygen-containing
groups, and larger molecular weights are characteristic of lower rank coals.
The present model for coal structure consists of clusters of aromatic and hydroaromatic
systems that are cross-linked such that domains exist which undergo rapid reorientational
motion (Larsen, 1988a). Grant et al. (1989) have described the macromolecular structure of
coal in terms of a Bethe lattice, of coordination number sigma + 1, in which the aromatic
clusters are connected by both labile and char (stable) bridges. Solomon and coworkers
have used an alternative description of the coal lattice (1988a). Lattice models have been
quite successful in modeling coal devolatilization behavior (Grant et al., 1989; Fletcher et
al., 1990a, 1992a; Solomon et al., 1993).
In addition to covalent cross-links, many hydrogen bonds between clusters constitute
branch points and, if not broken, will affect Me. Hydrogen bonding between clusters is
thought to be responsible for the brittle and rock-like character of bituminous coals and
plays an important role in coal structure (Nishioka and Larsen, 1990a). Strongly basic sol-
vents, such as pyridine, will break these hydrogen bonds, replacing the coal-coal hydrogen
bond with coal-solvent bonds, and the coal macromolecules relax, achieving a lower free-
energy state (Ouchi et al., 1989; Nishioka and Larsen, 1990b). This will reduce the cross-
link density of the coal and make coals swell much more than nonpolar solvents (Quinga
and Larsen, 1988; Hall et al., 1990).
Peppas and Lucht (1984) proposed a structure in which chain entanglements can also
play a large role. There is a sudden decrease in swelling observed for coals above 86% C.
This could be the result of possible physical stacking of the polynuclear aromatic structures
or an increase in coordination number due to additional cross-linking above 86% C, as ob-
served by the coalification break discussed above (Quinga and Larsen, 1988).
Me plays a major role in the mechanical properties of coal. Coals as mined are glassy
and strained, locked into this configuration during the high pressures of coalification by
noncovalent cross-links (Brenner, 1985). When the hydrogen bonds are removed by the ad-
dition of a basic solvent, the coal becomes rubbery with the macromolecular chains having
high mobility (Green et al., 1982; Brenner, 1985). Coal reverts to a plastic phase when the
solvent is dried.
The transition from a glass to an elastic phase can also be induced by heating. Whether
the thermal transition to a mobile phase takes place before a coal thermally decomposes is
an important question. This glass transition temperature, Tg, has been measured for a num-
ber of coals and is shown to be around 60 K (Lucht et al., 1987; Sakurovs et al., 1987;
Yti~m et al., 1991). Tg values for coals are significantly reduced in the presence of basic
solvents. Below Tg, there is limited mobility of chain segments with diffusion effects very
low and cage effects very large. Above Tg, diffusion rates increase by 102-104 as segmental
mobility increases (Larsen, 1988a, b).
The glass transition temperatures were shown to be relatively independent of heating
rate in the range 5 to 40 K/min by Ytirtim et al. (1991). It has been shown that softening in
coal can occur without covalent bond cleavage, demonstrating that covalent bond cleavage
is not necessary for either coal softening or glass-to-rubber transition. Larsen (1988b), Yun
et al. (1991a, b) and Yiirtim et al. (1991) suggest that early tar evolution at about 600 K
could be due in part to the early release of the bitumen or mobile-phase compositions by
diffusion through the relaxed lattice network as the cross-link density of the network is not
change significantly during the early stages of evolution. Thus, the observed tars may not
2.3 MacromolecularStructure of Coal 93
onding I
Fig. 2.13 Two-phasemodel of a bituminous coal, macromolecular(shaded) and molecular (M). [Reproduced
with permission from Marzec, A., Fuel Process. Technol., 14, 41, (1986)]
have been covalently bonded to the network, but just trapped by slow diffusion rates and
strong noncovalent bonds as shown in Fig. 2.13.
Nishioka (1992) provides persuasive arguments that the bulk coal does not consist of
the traditional covalently cross-linked macromolecular network with associated molecular
phase. Based on solvent swelling and solvent extraction experiments, Nishioka (1992) ar-
gues that the rigid lattice network could consist of substantial quantities of material held to-
gether by relatively strong physical interactions (e.g., hydrogen bonds and relatively strong
interactions, such as ionic, charge transfer, and re-Jr secondary interactions caused by poly
functional groups).
The type and characteristic of the second phase or component of coal is not as well de-
fined as the macromolecular network, and has been the subject of extensive debate (Given
et al., 1986; Williams et al., 1987; Derbyshire et al., 1989a, c; Nishioka, 1992), "Mobile
components" observed in pulsed NMR studies have been shown to be complex, and their
relationship to a trapped molecular phase has not yet been determined (Kamienski et al.,
1987; Derbyshire et al., 1989c; Jurkiewicz et al., 1989).
Two views are presented concerning the "mobile protons" observed by 1H NNR spec-
troscopy in coal (Derbyshire et al., 1989c; Marzec and Schulten, 1989; Nishioka and
Gorbaty, 1990). The first is that these protons can be attributed to molecules that are free to
rotate in cages of the macromolecular network, and the second is that the mobile protons
are associated with fragments of the macromolecular network that can rotate due to single
C-C or C-O bond linking such fragments to the network. The mobile phase studied by
Marzec and Schulten (1989) by py-FIMS has a wide structural diversity that account for the
wide range of rotational mobilities of components in the mobile phase.
A general definition of the mobile phase is material, which is readily liberated by mild
thermal treatment (Derbyshire et al., 1989a). These components are richer in hydrogen and
significantly more aliphatic than the macromolecular network, which is connected each oth-
er by stronger bonds that require more severe conditions and higher hydrogen consumption
to decompose.
Based on the geochemistry of coal, Given et al. (1986) suggest that there should be a
component of small highly mobile molecules originally presented in the peat or formed dur-
ing coalification clathrated inside most coal macerals, which should be mobile. This mole-
96 2 Chemicaland MacromolecularStructure of Coal
cular component of coal would be trapped by the forming macromolecular phase and it
would require considerable activation energy to let them escape. Vahrman (1972) conclud-
ed that relatively small molecules in coal represent a considerably larger fraction of coal
than had been realized. Klotzkin (1985) supports the view that coal molecules are castrated
in pores of a macromolecular network. However, as shown by Nishioka (1989) and
Nishioka and Gorbaty (1990), some of these small molecules could be constituents of
macromolecules which are released during mild heat treatment.
Low molecular weight compounds, such as n-alkanes, naphthalene, phenanthrene, and
alkylpentacyclic hydrocarbons, were produced by very mild heating (Nishioka and Larsen,
1990a). This result implies that some of the low molecular weight compounds found in sol-
vent extracts are components of the macromolecular network and are not always trapped in
the coal. For a high volatile bituminous coal (Illinois # 6), it was concluded that most of the
n-alkanes and PAH are not physically trapped inside networks and can be extracted by con-
ventional Soxhlet extraction during heat treatment (Nishioka and Gorbaty, 1990).
Following an extensive solvent extraction procedure, Stock and Wang (1990) carded out a
series of deuteridegradation reactions on the extraction residues of Illinois #6, West
Virginia Upper Kittannig and Millemerran coals. They concluded that the alkanes from
Illinois # 6 coal were virtually all extractable with no evidence of covalently bonded, large
linear alkanes in the extracted coal residue. In the West Virginia coal residue, the C16-C18
alkanes removed by deuteriodegradation appear to be essentially all trapped in the matrix
while the higher molecular weight alkanes appear to have been covalently bound to the coal
lattice. The alkanes derived from the Millemerran extracted coal appear to be covalently
bonded to the insoluble component of the coal. These data demonstrate that small amounts
of unbonded alkanes can escape persistent extraction efforts in some coals and that some
coal macromolecules may exist with large quantities of covalently bonded alkyl groups.
The alkanes found by Youtcheff et al. (1983) consisted mostly of n-alkanes, C12-C35,
but also contained many biomarkers such as pristane, phytane, and diterpanes, and a com-
plex homologous series of hopanes and moretanes, C27-15C35. Given (1984b) concludes
that much of the mobile phase is structurally unrelated to the network and is an independent
component derived from molecules in the original peat. This is true for the biomarkers and
their aromatized equivalents found in extracts of coal. The alkyl aromatics may be the re-
mains from carotenoids, consistent with the view that the extractable material is largely de-
rived from lipids. Allan and Larter (1983) compared the structural features of the mobile
phase and the macromolecular network. Distributions of alkyl aromatic compounds were
similar in vitrinite and liptinite from the same coal, implying that some mobile phase hy-
drocarbons can diffuse from one maceral to another during coalification. Nip et al. (1992)
reached a similar conclusion.
Given (1984b) stated that minor amounts of polymeric material, which probably de-
rived from polysaccharides, and were not cross-linked into the network, are probably incor-
porated into the network consistent with van Krevelen's ideas on coalification. The pres-
ence of the mobile phase in vitrinite is responsible for a major part of the maceral's sec-
ondary fluorescence (Lin et al., 1986, 1987; Davis et al., 1990, 1991). Fluorophoric struc-
tures of the mobile phase include aromatics, polar and a small proportion of alkanes. Radke
et al. (1982, 1984) resolved a number of methyl homologues of naphthalene and phenan-
threne, which were strongly controlled by rank. Many of the aromatics found by Radke et
al. (1982, 1984) possibly represent fragments of the macromolecular network.
The association of smaller mobile-phase molecules with the macromolecular network
is expected to be varied (Derbyshire et al., 1989a). These associations may involve both
2.3 MacromolecularStructure of Coal 97
physical and chemical forces, including covalent bonds, dispersion forces, hydrogen bond-
ing, and physical entrapment. However, it is doubtful whether any technique can distin-
guish between these different modes of attachment to establish the precise associations of
the mobile phase (Derbyshire et al., 1989a). Whatever their original association in the coal,
the existence of a thermally extractable bitumen-like fraction, which is chemically distinct
from the remaining coal components, evolves at much lower temperatures than the main
pyrolysis event associated with network breakdown (Williams et al., 1987; Yun et al.,
199 l a, b). In coals of bituminous and higher rank, the nature of this thermally extractable
fraction is consistent with that of a natural pyrolyzate formed by catagenic processes during
the "oil formation window" of maturation, while in low rank coals, various types of bio-
marker molecules are the important constituents (Chang et al., 1988, 1992., Yun et al.,
1991a; Carlson et al., 1992). The development of the mobile phase during devolatilization
has been shown to have an important effect on coal's thermoplasticity and hydrogen dona-
tion reactions (Neavel, 1982; Fong et al., 1986; O'Brien el al., 1987; Lynch et al., 1988).
Marzec (1985, 1986) envisions the molecular phase attached to the macromolecular
netwotk through electron donor-acceptor interactions (see Fig. 2.13). This view has recent-
ly been emphasized by Nishioka (1992). Marzec (1986) proposes that for low rank coals
( < 82% C), the molecular phase may represent 30-50% by weight of a whole coal organic
matter, and even more for some lignites. Youtcheff et al. (1983) showed that the additional
release of alkanes in liquefaction was six to eight times greater than could be obtained from
extraction, and concluded that these alkanes were physically trapped inside the macromole-
cular network and released only on disruption of the network. These materials are likely
residues from the bitumization process during coalification. However, Nishioka and
Gorbaty (1990) have shown that the large amounts of small neutral molecules are not tight-
ly occluded inside the coal network, but that the "mobile phase" consists largely of transla-
tionally restricted components with substantial mobility that are part of the macromolecular
structure.
A careful study by Jurkiewicz et al. (1989) on the APCS Utah Blind Canyon coal was
carried out to reexamine the nature of the mobile protons in coal. Using ~H CRAMPS and
1H CRAMPS dipolar dephasing experiments, the aromatic and aliphatic protons were re-
solved in the native coal, pyridine-d5 swollen coal, and pyridine-d5 saturated extraction
residue. The data provide a rough correlation between proton mobilities as determined by
dipolar dephasing rates and the extractable components of coal. These authors rationalize
the data on proton mobility in terms of a molecular-macromolecular structure with domains
of coal structure with different mobilities. Lorentzian time constants for dipolar dephasing
rates are associated qualitatively with the translational mobility implied by extractability.
The data on extraction residue are consistent with a rigid macromolecular network. De la
Rosa et al. (1992) have carefully examined the proton relaxation data in all of the Argonne
Premium Coal Samples and note a non-exponential of the spin lattice relaxation time, /'1,
which increases as the rank decreases. Proton relation in the rotating frame, Tip, follows a
similar trend to that noted for T1 in the same suite of coals. Although the mechanisms of
longitudinal relaxation in coals are not well understood, Dela Rosa et al. (1992) state
... it seems reasonable to suppose that the non-exponential relaxation in coals is
associated with lack of facile spin diffusion between domains of differing T1. These
domains, which could correspond to different macerals, can vary with respect to
concentrations and mobilities of hydrogen nuclei and to concentrations of para-
magnetic centers.
The important point to be made from the NMR data is that isolated domains do exist.
98 2 Chemicaland Macromolecular Structure of Coal
Differences in translational and/or rotational motion within the coal material or heterogene-
ity due to different macerals or a nonuniform distribution of paramagnetic centers could ex-
plain these results.
Two-component system:
9 Macromlecularthree-dimensional cross-linked
network (immobile phase)
9 Multitude of relatively small molecules with
varying structures embeddedtherein (mobile
phase, 10-50% share)
nite structure should be representative of the substitution pattern of lignin, such as the pres-
ence of ortho-dihydroxy substituents. Meuzelaar et al. (1984a, b) and Winans et al. (1987,
1988a) suggest that polycondensed aromatic species do not dominate the aromatic structure
except in very high rank coals and inertinites. Lignin does not contain polycyclic aromat-
ics; thus, it is expected that the amount of polycyclic aromatics in vitrinites gradually in-
creases with rank. A good model should show most of the OH as hydrogen bonded.
Structural features such as dibenzyl ethers and thioethers should indicate how all of these
structures vary with rank. Solum et al. (1989a) have demonstrated this functional group
dependence on rank. The structure must be based on data for a wide enough sampling of
coals to permit differences due to differing geochemistry and geological history to be built
into the model (Given, 1984b). Even though no such model has yet been reported in the lit-
erature, current working coal structural molecular models are helpful in understanding and
modeling the nature and reactivity of coals (Lazarov and Marinov, 1987). Synthetic coal
models have proven useful in interpreting coal results; however, there is danger in extrapo-
lating behavior of small molecules to that of solid coal (Winans et al., 1990; Larsen, 1990).
Relationships between coal structure and its reaction processes are reviewed in Given
(1984a), Larsen et al. (1986), and Larsen (1988b).
One of the problems of assigning a structure to coal is the assignment of a structure to a
mixture. However, coal can be described in terms of the following structural parameters as
described in Davidson (1980, 1982): size distribution of the macromolecules, degree of
cross-linking, type of cross-links and linkages, carbon aromaticity, average size of the con-
densed aromatic units, number of hydroxy groups and other functional groups of oxygen,
nitrogen, and sulfur, hydroaromatic and other aliphatic structures, and scissile, or easily
split, bridging structure (see also Given, 1984b).
A good review of the molecular structure and chemistry of coal is given in Davidson
(1980, 1982). A number of significant attempts have been made to construct an average or
representative macromolecular structure of coal. In general, the information used for con-
structing a coal structure are molecular weight, size, linkages between aromatic units, aro-
matic-to-aliphatic carbon ratio and elemental composition, as well as experimental results
from chemical and thermal reactions. Nishioka and Larsen (1990a) state that the most fa-
vored geometry of aromatic tings in coal involves the edge of one of the rings approaching
the face of the other in a T or herringbone stacking configuration. Carlson (1991) and
Nishioka (1992) have also pointed out the importance of considering the nonbonded inter-
actions in the various representations of coal structures.
Most coal structure studies are concerned with characterizing the structure of vitrinite.
The chemical structure of the other macerals has not yet been extensively studied. The vit-
rinite maceral is regarded as a mixture of at least two different kinds of components: a
macromolecular phase connected by cross-links with a molecular phase inside the macro-
molecular network as shown in Fig. 2.13 (Marzec, 1985, 1986). It has also been shown that
some of the small molecules usually attributed to the molecular phase may consist of highly
mobile constituents that are part of the macromolecular structure (Nishioka and Gobaty,
1990). If coal is a complex mixture of macromolecules which hold a guest phase or a com-
ponent of smaller molecules, there will be a great deal of structural diversity within one
sample, and the significance of information averaged over the whole mixture will be diffi-
cult to evaluate (Given, 1984b).
Derbyshire (1991) and Faulon et al. (1992) have reviewed the macromolecular struc-
ture of vitrinite and considered the chemical composition of the structural elements, the in-
terconnecting bonding and their spatial arrangement. Vitrinite is described as being made
100 2 Chemicaland Macromolecular Structure of Coal
up of relatively low molecular weight structural units connected by various types of bonds.
The units are made up of one, two, three, or more cyclic aromatic and hydroaromatic and
heterocyclic carbon structural units connected by covalent bonds, including alkyl, etheric,
oxygen, and sulfur bridges, as well as nonbonded interactions such as hydrogen bonds and
van der Waals forces (Derbyshire, 1991). Sakurovs et al. (1989) found two distinct types of
fusible materials in coal. One type is aliphatic-rich and is associated with liptinitic macer-
als and the other is aromatic-rich associated with vitrinite. The aliphatic-rich material has
enhanced stability with increasing coalification rank attributed to progressive covalent
cross-linking which also makes the material nonextractable. Fusion in aromatic-rich vitri-
nite macerals is the basis for thermoplasticity (Sakurovs et al., 1989). In lower ranked
coals, fusion is inhibited due to high covalent cross-link density. With further coalification,
these cross-links are degraded and, as a consequence, there is consolidation of the aromatic
units into microdomains or micelles with increasing graphite-like order (Sakurovs et al.,
1989). These microdomains are small and poorly ordered but become larger, more ordered,
and increasingly more stable with increasing rank so that their fusion temperature rises.
The thermoplasticity ceases in anthracites due to the stability of these structures, which in-
hibit fusion below temperatures of pyrolytic decomposition (Sakurovs et al., 1989). This
stability is greatly enhanced due to the rapid growth of the aromatic units. Inertinites, how-
ever, are generally more oxygenated and aromatic than vitrinites. According to Sakurovs et
al. (1989), thermal stability and resistance to pyridine destabilization is related to a greater
covalent cross-link density at all stages of coalification.
2.3.3 A d v a n c e s in S t u d i e s o n M a c r o m o l e c u l a r N e t w o r k o f C o a l
As mentioned above, there various interactions occur in coal, giving rise to the macromole-
cular structure of coal via either the covalent bond or the noncovalent bond. Based on ex-
perimental results, Aida et al. (1991) proposed a coal structure model and relative distribu-
tion of bonding interactions to be responsible for forming cross-linking the network struc-
ture in coal, as shown in Fig. 2.15. Recently, some new concepts regarding primary (mole-
cular) and higher order (supramolecular or macromolecular) structures of coal have been
proposed (Iino, 2002; Sanada, 2002). Coal-derived materials such as coal extracts and liq-
100
Covalent bonding
.o
'~ ~ Van der Waals bonding/
..~
uefaction products are known to readily associate in organic solvents. Hydrogen bonds and
the interactions between aromatic rings are considered to be the main associative interac-
tions, although charge transfer and ionic interactions may possibly contribute to the associa-
tions. As coal chemists well know, association behavior is also related to coal network
structure, i.e., covalent network or noncovalent (molecular associated) network. Coal
swells in organic solvents and a swollen coal shows elastic behavior, so it is sure that coal
has a kind of network structure. Although covalently connected giant network structures
are often assumed, recent works on solvent extractions suggest that at least for some bitu-
minous coals, a noncovalently connected network is a better model than a covalent one.
Probably the true figure lies somewhere between the two models, where the proportion of
covalent and nonocovalent network junctions depends on the kind of coal (Iino, 2002). In
this section, the newest advances in studies on the macromolecular structure of coal through
various phenomena characterizing coal are introduced.
A. Solvent Extraction
Polar solvents such as pyridine, THF and DMF have been reputed to be quite effective for
the liberation of noncovalent bond interactions, e.g., hydrogen bonds, aromatic plane stack-
ing and electrostatic interactions, in the coal macromolecular network (Suuberg et al., 1993;
Otake and Suuberg, 1997, 1998; Hall and Larsen, 1993; Cody et al., 1992; Takanohashi et
al., 2002). The solvent extraction method is widely used to investigate the characteristic of
this network in coal.
Figure 2.16 shows the relations between the amount of pyridine extraction and the C
content of the coal. It is observed that a maximum value occurs around the 86C% (Osawa
70 m
I)
I"1
60 - ii
ii
, ,
II
t/ i
50 - i i
i i
.~
/
0
~ 40
0
I) I)
30-
O
<
20-
10-
0
70 75 80 85 90 95
C (%, daf)
Fig. 2.16 Relationship between the amount extracted by pyridine and carbon content in coal.
[Reproduced with permission from Osawa. Y. et al., Kogyokagaku Zasshi, 73, 2214 (1970)]
102 2 Chemical and Macromolecular Structure of Coal
1400
A
1200 cpzx
.~ 1000 -
0
_~ 800-
m
0 O~y' OO4) A
O
600-
400 -
200 J t i t
70 75 80 85 90 95
C (%, daf)
Fig. 2.17 Relationship between the molecular weight of extract by pyridine and carbon content in coal.
[Reproduced with permission from Osawa. Y. et al., Kogyokagaku zasshi, 73, 2215 (1970)]
et al., 1979). Further, the molecular weight of the extract increases with increase of extract,
as shown in Fig. 2.17. A similar tendency was observed even if chloroform was substituted
for pyridine as the extraction solvent.
Krevelen (1965) and Sasaki et al. (Sanada and Honda 1966) tried to explain the extrac-
tion mechanism of the coal by a statistical method using the solubility parameter and the
degree of swelling, etc. In these analytical techniques, it is considered that the coalification
process is a condensation polymerization and a statistical method developed by Flory and
Stockmayer (Flory, 1953) for a condensation polymerization was utilized. Treatment of
coal as a condensation polymerization also leads to a minimum in the cross-link density at
ca. 86% carbon.
Krichko and Gagarin (1990) investigated the conversion of the chemical species around
~ 1t I I I i I t
@@ 9
0.50 9e
~ *~ 9
E 9o
0.45 ~o 9
~ ~ 9
"~ 00 9
9O OO
0.40
0000 9
t I J I , 9)0 I t
70 80 90
Carbon content (wt%, daf)
Fig. 2.18 Relationship between association parameter and carbon content in coal. [Reproduced with permission
from Krichko, A.A. and Gagarin, S.G., Fuel, 69, 888, Elsevier (1990)]
2.3 M a c r o m o l e c u l a r S t r u c t u r e of C o a l 103
100 _ ~ ~ II _ -t2"5
~- 80 -
, ~ 2.0
60 ~
o=
.,.-,
40 ' .=,
~ 1.5
20 q
0
0
/ , , , I ,
20
, , I ,
40
, , I ,
60
, , I ,
80
, , 1
1 O0
1.0
NMP (vol%)
aggregate structure of Upper Freeport coal can be related by mild heat treatments. In the
following section, the nature of the aggregate structure of Upper Freeport coal, which was
investigated using solvent extraction, is introduced in detail.
The extraction yields of Upper Freeport coal against CS2 vol% in the mixed solvent are
shown in Fig. 2.19 (Takanohashi et al., 1999b). The extraction yields with NMP or CS2
alone were 18% and 3%, respectively, while their mixed solvent gave higher extraction
yields and showed a significant synergistic effect. The maximum yield was obtained at
about 1:1 volume ratio (8:5, mole ratio) of CS2 and NMP, as in the case of other bituminous
coals. The effect of the solvent composition of the mixed solvent on the swelling ratio is
also shown in Fig. 2.19. Here, the residue from the 1:1 CS2-NMP mixed solvent extraction
was used for all the measurements of the swelling ratios for various solvent compositions.
The same trends were obtained for both extraction yield and swelling ratio.
The DSC thermograms of the residues obtained from the CSz/NMP mixed solvent ex-
tractions of Upper Freeport coal are shown in Fig. 2.20 (Takanohashi et al., 1998). In some
0.06 . . . . w
oo4 . o08
0.02 0.04
=
~Z --0.02o
--0.04 ,,l,,,,I,,,,
i (a) CS2: N M P = 0:100
I .... I,,,,I,,,,l .... I ....
]
1
--0.04
,,I .
!e) CS,2:NMp~,70:30
. . . . . .
1
0 100 200 300 400 0 100 200 300 400
0.15 0.1 .... I 'l .... I .... 1 .... I .... , .... I ....
0.05 0
0
--0.05 --0.05
--0.1 ~ - , , , i , , , , l .... J .... l .... 1 --0.1
0 1O0 200 300 400 0 1O0 200 300 400
0.05 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.04 ..... I .... I .... I .... I .... w .... I'"': I
-I
0 0
-0.05
--0.04
'~ --0.1
:= 0.08
--0.15 '
0 100 200 300 400 0 100 200 300 400
0.1 ,.. 0. 2 I''' '1'' ''1''''1''''1''''1''''1'';~
0.05 . . . . . . 0. )8
0
o 0. 14
--0.05
0
--0.1 ,j( , .... N -lOO:0
....
--0.15 r,,,,i .... l,,,,i,,,,i .... i .... I .... i .... 1 --0. t4
0 100 200 300 400 0 100 200 300 400
Temperature (~ Temperature (~
Fig. 2.20 DSC thermograms of the extraction residues with different extractrion yields from Upper
Freeport coal: (--) first scan, (. . . . ) second scan, ( - - - ) third scan.
Extraction yield: (a) 18%, (b) 43%, (c) 62%, (d) 66%, (e) 50%, (f) 35%, (g) 20%, and (h) 3%
[Reproduced with permission from Takanohashi, T. et al., Energy Fuels, 13, 509 (1999)]
2.3 Macromolecular Structure of Coal 105
thermograms, a broad peak was observed at 200~ ~ due to the NMP retained in the
sample. It is noted that, for the residues with relatively low extraction yields of 18 (a), 20
(g), 3 (h) wt% (daf) respectively, a peak similar to that seen for raw coal was observed at
350 ~ showing that the peak still remained after extractions with low extraction yields.
The peak also disappeared on the second and third scans, similar to the behavior in raw
coal. By contrast, in the case of residues with high extraction yields, i.e., 44 (b), 63 (c), 67
(d), 51 (e), and 35 (f) wt % (daf), the peak around 350 ~ has already disappeared from the
first scan. These results suggest that the solvent extractions could relax the aggregate struc-
ture of Upper Freeport coal even under mild conditions, i.e., at room temperature, similar to
the heat treatment at 350 ~ In the case of low extraction yields (a, g, h), the relaxation of
the aggregate structure with limited extraction is not enough. This is consistent with the
observation that pyridine-Soxhlet extracted coal also showed the endothermic peak (Yun
and Suuberg, 1993). It has been reported that coals heated at 375~ ~ at which soft-
ening starts for caking coals, gave higher extraction yields with the CS2/NMP (1:1) mixed
solvent (Suuberg et al., 1994). The structural relaxation may be related to the softening of
coals. The softening temperature is 373 ~ for Upper Freeport, showing that after the ap-
pearance of the endothemic peak it starts to soften. Pittsburgh # 8 coal, which also gives a
fluidity similar to that of Upper Freeport coal, gave no sharp peak before the initial soften-
ing temperature (381 ~ Thus, at present there is no clear relationship between the soften-
ing temperature and the endothermic peak.
In the CS2/NMP mixed solvent extraction, no significant reactions between the solvents
and the coals have been suggested from the characterization of the extracts and residues ob-
tained and the extraction mechanisms (Iino et al., 1988). Figs. 2.19 and 2.20 suggest that
the solvent compositions, which gave high swelling ratios, gave high extraction yields, and
also resulted in disappearance of the endothemic peak. This means that relaxation of the
aggregate structure by the dissolution of noncovalent bonds is most likely responsible for
the peak. The high extraction yields of Upper Freeport coal have been attributed to the dis-
sociation of noncovalent bonds such as Jr-Tr interactions and charge-transfer interactions
(Liu et al., 1993; Ishizuka et al., 1993; Takanohashi et al., 1995). The relaxation of the ag-
gregates by heating or solvent is considered to be the most likely mechanism for the en-
dothermic peak at 350 ~ A better solvent may better relax the aggregate structure of coal
and dissolves extractable substances (Suuberg et al., 1994). The access of solvent into the
Upper Freeport coal requires that the coal be greatly swollen and the aggregate structure is
likely to be relaxed in the process. In recent years, it has been proposed that the diffusion
of solvent into the high rank coals may be retarded by intermolecular interactions in the
coals, such as stacking among aromatic tings (Takanohashi et al., 2003). It has been sug-
gested that CS2 may be able to diffuse in the micropores of Upper Freeport coal because of
its straight shape and high affinity with higher rank coals. However, investigating the es-
sential role of C52 in the CS2/NMP mixed solvent extraction remains a challenge for the fu-
ture.
B. Solvent Swelling
Solvent swelling is considered to be one of the most reasonable phenomena reflecting the
characteristics of the cross-linking structure in polymeric substances. A macromolecular
network will absorb solvents and swell without dissolving. The amount of swelling is con-
trolled by the cross-link density and by the magnitude of the interactions between the sol-
vent and the network. Thus, measurements of the amount of swelling and of the interac-
tions between the solvent and the macromolecules can be used to calculate cross-link densi-
106 2 Chemical and Macromolecular Structure of Coal
ties. As noted above, coals are complex macromolecular solids of polymericity. Similarly,
the measurement for the swelling of a coal should be related to the covalent cross-links in
the coal, and should give information concerning the macromolecular structure of the coal.
Of the techniques used to define the macromolecular structure of various coals, solvent
swelling is thus the one most often used (Quinga and Larsen, 1988). The equilibrium
swelling value (Q-value) and swelling speed (V-value) are generally used to evaluate the
swelling process (Green et al., 1982; Larsen, 1988a; Quinga and Larsen, 1988). The degree
of swelling of a coal by solvents, the swelling value, Q, is the ratio of the volume of
swollen coal to the original coal volume at equilibrium with the swelling solvent.
The solvent swelling of coals has been reviewed previously (Quinga and Larsen, 1988).
The arguments, which support a cross-linked macromolecular network, other than the fact
that coals are not soluble in any solvent, are that coals swell and expand by as much as
250% when brought in contact with the appropriate solvent. The degree of reversibility is
also consistent with a covalently cross-linked structure and rules out entanglements as the
sole associative force (Quinga and Larsen, 1988). Coals also display viscoelastic proper-
ties, which is consistent with a network structure. The physical structure of coal is re-
viewed by Gorbaty et al. (1986).
In a study of cross-linking reactions during coal conversion by Solomon et al. (1990e),
the swelling behavior of the Argonne coals was measured to investigate cross-linking reac-
tions during pyrolysis. The volumetric swelling value, Q, for the Argonne coals reveal that
higher rank coals do not swell to the same degree as the lower rank coals. The high
swelling value for the Blind Canyon coal agrees with the observations that this coal has a
low cross-link density (Orendt et al., 1992). These data also qualitatively agree with the
fraction of intact bridges and the number of bridges and loops calculated by Orendt et al.
(1992).
Table 2.4 Summary of Swelling Experiments of Coals
Coals Qcoaa. b
Upper Freeport (mvb) 1.32 (1.32)
Wyodak (subC) 2.70 (2.73)
Illinois #6 (HvCb) 2.53 (2.53, 2.53)
Pittsburgh (hvAb) 2.30 (2.33, 2.26)
Pocahontas # 3 (lvb) 1.12 (1.12)
Blind Canyon (hvBb) 2.74 (2.76, 2.66, 2.74)
Lewiston-Stockton (hvAb) 2.18 (2.35, 2.18)
Beulah-Zap (ligA) 2.71 (2.71)
aVolumetric swelling ratio for coal, i.e., the ratio of the volume of swollen coal
to original coal volume at equilibrium with the swelling solvent.
bInterpolated (duplicate runs)
[Reproduced with permission from Norinaga, K. et al., Energy Fuels, 14,
1249 (2000)]
The most fundamental feature of a polymer network is the number of average molecu-
lar weight between cross-links or junction points ( Me ). This quantity is a measure of the
"mesh size" of the network and, to a first approximation, it determines the physical proper-
ties of the network. It is possible to determine Me from mechanical properties or solvent-
swelling measurements (Flory, 1953; Treloar, 1975; Lucht and Peppas, 1981, 1987b; Green
et al., 1982; Peppas and Lucht, 1984). At equilibrium, the driving force for penetration of
the network exactly balances the elastic restoring force of the network by the solvent. Me
is calculated from an equation, which is derived by equating theoretical expressions for the
elastic chemical potential. There are two terms involved in the free energy of dissolution of
2.3 Macromolecular Structure of Coal 107
2.5 I I I
2.0
1.5
3
A
1.0
8 9 10 11
t~ (cal]cm3) t/2
Fig. 2.21 Swelling ratio (Qv.dry)for native Ilinois #6 coal (11), the pyridine-extracted coal (O), the oxygen-methy-
lated unextracted coal (C)), and the pyridine-extracted oxygen-acetylated coal (A) as a function of the
Hildebrand solubility parameter (d;) of the swelling solvent. The swelling solvents are (1) n-pentane, (2)
n-heptane, (3) methylcyclohexane, (4) cyclohexane, (5) o-xylene, (6) toluene, (7) benzene, (8) tetralin,
(9) naphthalene, (10) carbon disulfide, (11) biphenyl.
[Reproduced with permission from Larsen, J.W., J. Org. Chem, 50, 4731 (1985)]
the solvent in the coal. The first is entropic; the concentration difference between the bulk
solvent and the interior of the coal. The second is any interaction between the solvent and
the coal, usually given by the Flory Z parameter. Except for one now being developed
(Painter et al., 1988), the statistical treatments that have been used to calculate the number-
average chain length between the solvent and the coal is random (Kovac, 1978; Lucht and
Peppas, 1987). It is explicitly assumed that there are no specific one-to-one interactions
such as hydrogen bonds or charge-transfer complexes formed between the solvents and
structural units in the coal. Since coals contain significant concentrations of hydroxy
groups and interact strongly with hydrogen-bonding acceptor (Szeliza and Marzec, 1983),
swelling measurements using hydrogen-bonding solvents cannot be interpreted by this the-
ory. However, Painter et al. (1990) concluded that the solubility parameters for coal are
not accurately determined from swelling measurements because of free volume effects, and
they are developing a more comprehensive model.
The first to apply solvent swelling to coals were Sanada and Honda (1966), who mea-
sured the pyridine swelling of a series of coals and obtained Z values from the concentra-
tion dependence of osmotic pressure of pyridine solutions of coal extracts. Me was calcu-
lated using the Flory-Rehner equation. The next application of this technique was by Kirov
and co-workers, who swelled three coals in 17 solvents and used a regular solution ap-
proach to calculate c (Kirov et al., 1968). Recently, Larsen et al. developed this theory and
investigated the macromolecular structure of coal. Fig. 2.21 shows the swelling in a set of
nonhydrogen-bonding solvents of native Illinois # 6 coal, the same coal extracted with pyri-
dine, the methyl ether of the coal and the acetylated coal. The swelling ratio is plotted as a
function of the Hildebrand solubility parameter of the solvent (t~s). The swelling ratio
(Qv.dry) is the ratio of the volume of the swollen coal to that of the unswollen sample. A
maximum in swelling occurs when the macromolecular network and a swelling solvent
108 2 Chemicaland MacromolecularStructure of Coal
have the same value, and swelling decreases as the 6 values diverge as the solvent-network
interactions become less favorable. When all of the hydrogen bonds are removed by de-
rivatizing the hydroxy groups, maximum swelling is obtained. The interactions with hydro-
gen-bonding solvents were also investigated in the swelling of coal with hydrogen-bond ac-
ceptors such as pyridine, acetone, etc. They proposed that the excess swelling, i.e., the dif-
ference between the swelling by the solvent and that by nonpolar solvent of the same 6, be-
comes greater when the heat of hydrogen bonding between a solvent and p-fluorophenol
calculated by Arnett et al. (1970) is higher. Further, the range of hydrogen-bond enthalpies
in an unextracted bituminous coal was investigated by measuring the ratio of the swelling
of the coal to the swelling of the acetylated coal. A solvent which breaks all of the hydro-
gen bonds in a coal will swell the coal and the acetylated coal to the same extent; the ratio
will be one because there are no hydrogen bonds in the acetylated coal. As solvent hydro-
gen-bonding ability decreases, the ratio will drop, as shown in Fig. 2.22. Solvents having
hydrogen-bond strength with p-fluorophenol below about 5 kcal/mol cause little excess
swelling. Pyridine breaks nearly all of the hydrogen bonds in the coal and a solvent, which
formed 8.5 kcal/mol hydrogen bonds strength to phenol, would break all of the coal-coal
hydrogen bonds. Further, the number average molecular weight between branch points
( M e ) was calculated using the Flory-Rehner equation and the Kovac equation (Kovac,
1978), although there is some problem regarding the confidence of the method, especially
in the case of lignite. The low rank coals are highly oxygenated and have much more po-
larity than bituminous coals. Because of this, lignite coal does not follow the regular solu-
tion theory. This makes it impossible to calculate Z for coal-solvent pairs (Larsen and
Shawver, 1990). Thus, at best, solvent swelling provides qualitative information for com-
parison purposes. As described in Chapter 1, the mole fraction of intact bridges and the
number of bridges and loops determined from a 13C NMR analysis also provide an estimate
of the cross-linking of a coal or char. Assumptions in the calculation of these bridges and
loops are given in Chapter 1. The cross-linking as estimated by NMR provides more quan-
titative data compared to solvent swelling studies.
1.0 I I I / I
/
~ Pyridine
_ DMS~ _
0.8
100
:
60
40
20
0 , ~ ....
Fig. 2.23 Transverse relaxation signals for pyridine swollen BL coal. (Norinaga et al., 2000)
FID: free induction decay, S/C: Ratio of solvent to coal mass.
[Reproduced with permission from Norinaga, K. et al., Energy Fuels, 14, 1248 (2000)]
C. 1H-NMR
The phenomenon of solvent-induced swelling of coal has long been used to characterize the
macromolecular structure of coal. In particular, the Flory-Rehner theory and variants there-
of used to estimate the molecular weight between cross-link points have been frequently
employed to relate the macromolecular network parameters to the degree of swelling in a
good solvents (Sanada and Honda, 1966; Kirov et al. 1968; Green et al., 1982; Nelson,
1983; Larsen et al., 1985; Lucht and Peppas, 1987). The F-R theory tacitly assumes that
the deformation of the elementary chains of the network is affined, down to the molecular
level. Thus any observed macroscopic deformation (e.g., osmotic dilation) of a given sam-
ple is assumed to correspond linearly with a change in the statistical distribution of chain
lengths of the coal macromolecule (Flory, 1953). In other words, the coal macromolecule
is assumed to dilate uniformly at the segmental scale when we relate the macroscopic
swelling to molecular characteristics such as the cross-link density. Recent experiments ex-
ploring ~H-NMR transverse relaxation characteristics (proton spin diffusion), reveal that the
coal hydrogen in the pyridine-swollen state must be divided into two groups: hydrogen with
relaxation characteristic of solids and hydrogen with relaxation characteristic of liquids
(Jurkiewicz et al., 1981, 1982, 190, 1993; Barton et al., 1984; Kamienski et al., 1987; Yang
et al., 1993, 1994). For example, Barton et al. (1984) have shown that up to 60% of coal
hydrogen becomes mobile when immersed in deuterated pyridine, while the remaining 40%
remains in rigid structures and is characterized by an NMR signal component that is similar
to the signal for the corresponding dry coal. These results appear to suggest that the osmot-
ically dilated coal is phase separated into domains that are solvent rich and solvent deplet-
ed.
Recently, Norinaga et al. (1999a, b; 2000) used proton diffusion measurements to de-
termine the scale of the heterogeneity in the phase separated structure of swollen coals.
Five coals of different ranks were osmotically dilated via saturation with deuterated pyri-
dine and were analyzed via ~H-NMR spectroscopy using a partially modified Goldman-
Shen pulse sequence and modeled using simple geometric models of a two-domain (phase)
110 2 Chemicaland Macromolecular Structure of Coal
system. These calculations indicated that the solvent-rich phase domains in the swollen
coals exist as discrete regions in the size range of 20 to 200 ~ , requiring that coal does not
swell affinely at the molecular scale but rather dilates non-uniformly leading to nanoscale
structural heterogeneities. These results place our understanding of the macromolecular
structure of coal in some jeopardy and certainly raise questions regarding the applicability
of any affine statistical thermodynamic model to describe macroscopic, osmotic strains of
coal measured by bulk solvent swelling (Norinaga et al., 2000).
Figure 2.23 shows the FID curves for the swollen Blind Canyon (BL) coal are drawn as
a function of decay time. Although the solvent swelling enhances the fraction of slowly de-
caying components, a portion of the coal hydrogen remains rigid. For a dipole coupled
rigid system such as dry coal, the time decay of the nuclear magnetization can be character-
ized by a Gaussian function. On the other hand, in a liquid or a liquid-like environment, the
magnetization decay is approximately an exponential function. Therefore, the observed
FID was assumed to be expressed by the following equation and was analyzed numerically
0.8 .
0.6 ~
I
It
O
0.4
0.2
0
0 1 2 3 4 5
S/C ( - )
Fig. 2.24 Change infMHwith S/C. [Reproducedwith permission
from Norinaga, K. et al., Energy Fuels, 14, 1248 (2000)]
Table 2.5 Resultsof Proton Longitudinal Relaxation Measurements for Blind Canyon Coal Swollen in
Deuterated Pyridine
S/C a T~p(ms) T~(ms)
Ti/b Tips
[Reproduced with permission from ed. Iino, M., Hayashi, J.-i. et al., Primary and Higher Order Structures of
Coal and Their Influence on Coal Reactivity-Final Report on "Research for the Future" Coal Research
Project-, 119 (2001)]
2.3 MacromolecularStructure of Coal 111
where l(t) and Ii(t) are the observed intensity at time t, and that attributed to component i,
respectively, and T2i is the transverse relaxation time of the i th component. The fractions
of hydrogen producing exponential decays, fMH, are plotted against S/C in Fig. 2.24. full in-
creased to 0.5 with increase in S/C. However, full remained at an almost constant value
above S/C--2.24, indicating that there exist solvent impenetrable regions in the swollen
coal even at S/C = 4.72. For the swollen coal samples, it is clear that there are domains
that do not swell and are not penetrated by solvent as reported previously. The phase struc-
tures of the swollen coal are separated into at least two phases, i.e., solvent rich (SR) and
solvent impervious phase (SI).
To examine whether the spin diffusion process is active in the swollen coal samples,
proton longitudinal relaxation was measured both in the laboratory and rotating frame.
Table 2.5 lists the results of 7'1 and T~p measurement for the swollen BL coal. T~ is com-
posed of one component while Tip can be analyzed by the sum of two exponential func-
tions. From these results, one can clearly resolve the effects of spin diffusion. T2 signals
are composed of three components without the effect of spin diffusion while Tip and 7'1
measurements are affected strongly by spin diffusion, and the number of the components
decreases from Tip to 7'1. The existence of at least two time constants for a rotating frame
longitudinal relaxation process, i.e., T~p, in a system means that spin diffusion processes
cannot effectively average the different dynamical properties of protons in different spatial
domains on the relevant time scale of the specific relaxation process. On the other hand, in
the time scale of 7'1 measurements, the distinctly separate spin systems were sufficiently av-
eraged by the spin diffusion. The scale of spatial heterogeneities of the swollen coals can
be estimated by evaluating the diffusive path length, i.e., the maximum linear scale over
which diffusion is effective. The Goldman-Shen pulse sequence was thus employed to
monitor the spin diffusion process (1966). The advantage of the Goldman-Shen experiment
is that the time for spin diffusion can be arbitrarily varied, and if this time is much less than
7'1, the analysis is straightforward. The Goldman-Shen experiment is a technique that puts
the separate spin systems at different spin temperatures and then samples them as a function
of time so that their approach to equilibrium can be followed. In Fig. 2.25 (Norinaga et al.,
2000), the recovery factor of the magnetization of SI phase, R(t), is plotted versus the
square root of time, t 1/2, for the solvent swollen BL coals. It should be noted that S/C has
almost no effect on the observed R(t). The time evolution of R(t) is analyzed by the diffu-
sion equation solved by Cheung and Gerstein (1966) to obtain information on the diffusive
path length, 1. Since spin diffusion is similar to the spin-spin relaxation process, it is ex-
pected to occur much faster in SI phase than in SR phase. The rate-determining step of the
spin diffusion would be the diffusion in the SR phase. Thus 1 would correspond to the size
of the SR phase. The solid curve in Fig. 2.25 represents the nonlinear least squares fits to
the data using the diffusion equation. The analytical fits give 1 to be 70, 160, and 250 ik for
one, two, and three dimensions, respectively. As shown in Fig. 2.26, the morphology of the
domain, i.e., the spatial dimension that is assumed to solve the diffusion equation, affects
the estimated 1 values. Hence information regarding the morphology of the domains is re-
quired for the precise evaluation of the domain size.
Recently, Kumagai and Tanabe (2001) observed the swelling behavior of coal in sol-
112 2 Chemical and Macromolecular Structure of Coal
0.8
0.6
I
0.4
v S/C =0.68
9 1.67
0.2 9 3.52
0.0 m I I I I I t-
O 5 10 15 20 25 30
t 1/2(msU2)
Fig. 2.25 Recovery of proton magnetization in SI phase as a function of t ~/2for the solvent-swollen BL coal.
Solid lines represent the best fit to the data using a diffusion model.
[Reproduced with permission from Norinaga, K. et al., Energy Fuels, 14, 1249 (2000)]
vents monitored by in situ 1H-NMR relaxation measurements. Slowly decaying tail of sol-
id-echo signals for Beulah Zap (BZ) and Upper Freeport (UF) coals in pyridine gradually
increase with swelling time, reflecting increase in the mobile c o m p o n e n t from dissociation
of the associated structure in coal with solvent. In the mixed solvent of pyridine and CS2 the
c o m p o n e n t s increase rapidly at an earlier stage of swelling, indicating that the mixed sol-
vent accelerates the dissociation of the associated structure of coal. Further, solid echo sig-
nals obtained from original and swollen coal were curve-fitted with one Gaussian and two
exponential functions, i.e., the rapidly decaying c o m p o n e n t at Tzc of ca. 20/.ts, the interme-
diate c o m p o n e n t having T2int of 30 t o 7 0 ].LS and the relatively slow decaying tail of the sig-
2.3 Macromolecular Structure of Coal 113
nals at T2mof 1200 to 2000 ~ts. The first component was considered to be related to the hy-
drogen in the cross-linking, molecular structure, i.e., immobile phase, and the third was
considered to be related to the hydrogen attached to liquid-like molecular structure, i.e.,
mobile phase. The results showed that the effects of the mixed solvent on the dissociation
of the associated structure in UF are distinct from that in BZ. The method gives us some
hint in quantifying the hydrogen behavior in the macromolecular structure, but further ef-
forts are necessary.
I
',,o
'0 0 9
[] []
0.1 I I I I I I I I I I I I I I i i
I
0.01 0.1
Q (~-1)
Fig. 2.27 Coherent scattering intensity vs. momentum vector for variably swollen BL residues in binary solvent
of benzene-pyridine. For clarity only the scattering curves for the dry particles and osmotically dilated
particles corresponding to 20, 40, and 100% pyridine solutions are shown. The dashed line: low Q
highlights the power law region of coherent scattering that dominates the dry particle scattering and is
present in all of the scattering curves independent of the extent of osmotic dilation. Growth of scattter-
ing intensity in the intermediate range correlates with the degree of osmotic dilaton and indicates the
formation of dense regions of scattering domains.
[Reproduced with permission from Norinaga, K. et al., Energy Fuels, 14, 1247 (2000)]
114 2 Chemicaland MacromolecularStructure of Coal
~xponents below - 3 , are not physically consistent with scattering off topologically rough
surfaces. Although in the present case such power law behavior reveals a complex and dis-
3rdered structure for the BL dry coal particles, it is not currently possible to be more specif-
ic regarding precisely what structure(s) are responsible for this complex scattering behav-
ior.
With the addition of solvent, resulting in the osmotic dilation of the coal particles, the
coherent scatting behavior changes significantly and systematically. First, there are subtle
changes in the power law region, i.e., the magnitude of d with the benzene:pyridine ratio.
More interesting, however, is the obvious growth of a scattering peak or shoulder evident in
the intermediate range, 0.01 to 0.3 A -1 (Fig. 2.27). Such scattering behavior is qualitatively
consistent with the development of a dense system of scattering regions that formed as a di-
rect consequence of osmotic dilation.
A likely physical explanation for this new scattering component is that it results from
phase separation accompanying osmotic dilation leading to the formation of solvent rich
and solvent poor domains as has been suggested to occur in solvent dilated coals (Norinaga
et al., 1999). The scattering behavior of the dry and osmotically dilated coals could be sim-
ulated using a number of different geometric models. Unfortunately, the SANS data pre-
sented above do not provide a unique solution for a structure through direct inversion of the
data. Efforts in this direction are being explored and will be presented separately. For the
present purposes, however, these data are sufficient to show that the previous interpretation
derived from the proton spin diffusion measurements is reasonable.
Notwithstanding the difficulties of directly simulating the scattering curves, a crude es-
timate of the mean domain interspacing length scale can be obtained by the following ap-
proach. First, it is assumed that the power law scattering exhibited by the dry coal particles
and persisting over the range of osmotic dilation constitutes a background reflecting the in-
trinsic structural and chemical complexity of coal. No attempt (at this point) has been made
to ascribe a specific structure to this background. The development of nanoscale dense
scattering domains and the observed scattering "peak" associated with osmotic dilation
(Fig. 2.27) is, therefore, considered a phenomenon independent of the source the power law
background. Whether this assumption is correct remains to be determined. However,
crude estimation of the mean domain interparticle spacing will not be affected significantly
by these assumptions.
Treating the low Q power law scattering as being essentially a background, the dry coal
powder data can be subtracted from each of the osmotically dilated scattering curves high-
lighting the progressive development of the osmotic swelling derived scattering peak (Fig.
2.28). It has long been recognized that a coarse approximation of the mean interparticle
scattering dimension can be derived by application of the Bragg relation whereby Qrnax =
2zrr-~ (r is interdomain distance), although it is acknowledged that the Bragg relationship
cannot be rigorously applied to such features in the low Q region (Guinier, 1963). The
peaks revealed in Fig. 2.28 are highly asymmetric with a peak at c a . 0.015/~-~ and a central
moment closer to 0.03/~-~. This would suggest a mean interdomain distance of about 200
to 400 A. With a more sophisticated analysis of the data (dependent, of course, on a clear
picture of the physical structure responsible for the power law scattering in the low Q re-
gion) this estimate will be refined. However, it is unlikely that such an estimate will devi-
ate appreciably from the rough estimate presented above.
The effect of S/C on the ~H-NMR relaxation characteristics was examined, full in-
creased to 0.5 with increase in S/C. However, fMH maintained an almost constant value
above S / C - 2.24, indicating that there exist solvent impenetrable regions in the swollen
2.3 M a c r o m o l e c u l a r S t r u c t u r e of Coal 115
0.5
B e n z e n e : Pyridine v o l u m e ratio
!.t ' .~
i...".
":"J, "
0 ! ! I f i i i i i i i i | i n 0 I - I i I
0.01 0.1 1
a (/~-1)
Fig. 2.28 R e s i d u a l scattering intensity vs. m o m e n t u m vector for variably s w o l l e n B L residues in binary solvent
of b e n z e n e - p y r i d i n e . [ R e p r o d u c e d w i t h p e r m i s s i o n f r o m N o r i n a g a , K. et al., E n e r g y F u e l s , 14, 1248
(2000)1
coal even at S/C = 4.72. Whereas the transverse relaxation characteristics revealed the ex-
istence of at least two distinct structural regions in the swollen coals, the measured longitu-
dinal relaxation was best characterized by a single component as spin diffusion is rapid in
the swollen coals. The dynamics of spin diffusion were revealed using a partially modified
Goldman-Shen pulse sequence and analyzed by a simple mathematical model of a two-
phase system. Given a model of the domain structure, the diffusive path length was con-
verted to the interdomain spacing, di. The resultant d i evaluated for the one- and two-di-
mension models is near that derived by SANS. These results suggest that the domain struc-
ture developed from osmotic dilation may reflect a roughly laminar structure with sufficient
irregularity, e.g., curvature, to yield proton diffusion behavior that would lie between our
simple one- and two-dimensional models. More significantly both the SANS and NMR
data show that a strictly affine treatment of macroscopic osmotic dilation experiments is not
supported at the nan 9 At least part of the volume expansion of coals immersed in ex-
cellent swelling solvents is related to the formation of phase-separated domains. This
makes the interpretation the statistical distribution of elastic chain lengths derived from
macroscopic swelling measurements and the application of even modified statistical me-
chanical theories of swelling dubious.
E. Vapor Sorption M e a s u r e m e n t
Coals are highly porous solids. Green and Selby (1994) reported that pyridine sorption
isotherms could be explained by a dual-mode sorption model that has been widely applied
to the sorption of glassy polymers. Shimizu et al. (1998) carried out research on organic
vapor sorption using various ranks of coals and found that sorption data for Illinois # 6 coal
could be treated by the Langmuir-Herry equation regardless of the organic vapors
(methanol, benzene, pyridine and cyclohexane) used. Takanohashi et al. (2000d) reported
116 2 Chemical and Macromolecular Structure of Coal
2.5 u I , I I , I I I I I I I I I I I ~ I -
go
~9
1.5
0.5 - - -
O- , , I , n i/ i i i I i i i-
0 0.2 0.4 0.6 0.8
P/Po (--)
Fig. 2.29 Sorptionisothermof various alcohols at 30 ~ [Reproducedwith permission
from Takanohashi, T. et al., Energy Fuels, 14, 917 (2000)]
that sorption in the residues from coals with high extraction yields greatly increased com-
pared to the raw coals, suggesting that more microporosity has developed due to the extrac-
tion. Thus, in this section, the investigation of macromolecular structure of coal using va-
por sorption measurement is introduced.
Isotherms for absorption of methanol, ethanol, n-propanol and n-butanol by Upper
Freeport coal are shown in Fig. 2.29 (Takanohashi et al., 2000a). At relative pressures <
0.05, methanol and ethanol gave similar values, after which the rate of increase was larger
for methanol. In contrast, the sorption for n-propanol and n-butanol, which have bulky
alkyl groups, was quite small over the entire range of relative pressures. The difference in
adsorption of the alcohols used may be the result of alkyl group steric effects. The size of
the alkyl groups (methyl, ethyl, n-propyl and n-butyl) was estimated using DMol calcula-
tion (a density functional theory) (Painter, 1987); the distance between the edge of the alkyl
groups and the center of the oxygen atom was 2.7, 3.0, 4.0 and 4.0, if the oxygen atom of
the alcohols is assumed to be the adsorption site with interacting sites in coals. It is reason-
able to assume that the difference in volume of the alkyl group is responsible for the differ-
ence in sorption among the alcohols, i.e., methyl > ethyl > propyl -- butyl. Upper Freeport
coal seems to have more micropores into which the relatively bulky n-propanol and n-bu-
tanol can diffuse only marginally.
Interactions between coals and various organic compounds were studied using an in-
verse liquid chromatography (ILC) technique in which coals were used as the stationary
phase. All samples, including pyridine, NMP and tri-aromatics, yielded relatively low ca-
pacity factors (the increment of the elution volume of the probe relative to the elution vol-
ume of the carrier solvent), and some produced negative values, showing that the interac-
tion between the samples and Upper Freeport coal is small compared to the lower-rank
coals. It has been reported that the swelling ratios of Upper Freeport raw coal in ordinary
solvents such as methanol and benzene were relatively low. This suggests that raw coal has
porosity and cross-link structures so fine that they are not easily penetrated (Takanohashi et
al., 1995; Aida et al., 1991).
Capacity factors for straight-chain alcohols with different numbers of alkyl groups with
toluene as the Gamier solvent are shown in Fig. 2.30 (Takanohashi et al., 2000b). For high-
2.3 Macromolecular Structure of Coal 117
2.5 I ' I ' I ' ' I ' I ' I ' I ' I '-
o 1.5
"~ 1
0.5
0 2 4 6 8 10 12 14 16 18 20
Number of carbons in n-alcohol
Fig. 2.30 Capacity factors of straight-chain alcohols against the number of alkyl groups on the alcohols.
[Reproduced with permission from Takanohashi, T. et al., Energy Fuels, 14, 725 (2000)]
rank Upper Freeport and Pocahontas # 3 coals, the capacity factors greatly decreased for
one-three carbons, after which there was essentially no change. The n-alcohols with more
than three-carbon alkyl groups hardly penetrate the coal bulk. These results suggest that the
diameters of the micropores of Upper Freeport and Pocahontas # 3 coals are in the range
4-7/~. Hayashi et al. (1995) carried out ILC using n-alkane probes for Pocahontas #3
coal, and concluded that there were few pores of 10-100/k diameter in the coal. Larsen et
al. (1995) proposed that pore s in Argonne Premium coals are isolated and can be reached
only by diffusion through the solid. For high-rank coals, the above data are in agreement
with the model of coal pores developed by Larsen et al. (1995).
The results of ILC and solvent swelling also show that the rate of diffusion of ordinary
organic compounds into the coal micropores is relatively small.
Fig. 2.31 Eight coal molecules placed randomly in a rectangular cell (top view). [Reproduced with permission
from Takanohashi, T. et al., Prepr. ACS, Div. Fuel Chem. 45, 242 (2000)]
1900
O
1800
O
1700
1600
1500 -u
1400
O Q
1300 qi
1200 ~ .'It.
lp-
1100 i i i
0.9 1 1.1 1.2 1.3
Density (g/cm 3)
Fig. 2.32 Plot of calculated density against total energy for the molecular model of Upper Freeport coal.
[Reproduced with permission from ed. Iino, M., Takanohashi, T. et al., Primary and Higher Order
Structures of Coal and Their Influence on Coal Reactivity-Final Report on "Research for the
Future" Coal Research Project-, 68 (2001)]
al., 2000c). The total energy remained unchanged at around 1.27 g/cm 3. The size of cell
was reduced manually, and MM-MD calculation was carried out; this procedure was re-
peated several times. Finally, the energy-minimum point was observed at a density of 1.28
g/cm 3 (Fig. 2.32), in good agreement with the experimentally obtained value of 1.30 g/cm 3.
Figure 2.33 (a) shows the conformation at minimum energy for model molecules in the
cell (Takanohashi et al., 2000c). The size of cell was 53.6 ,~ X 55.8 ,~ X 4.2/~. An
anisotropic associated structure was obtained. Cody et al. (1988) have reported anisotropic
swelling behavior for bituminous coals: the swelling ratio was greater perpendicular to the
bedding plane than parallel to it. Fig. 2.33 (b) shows two cells enclosing the model struc-
ture at minimum energy (Takanohashi et al., 2000c). Aromatic rings seem to interact with
one another perpendicular to the bedding plane. The distribution of the distances between
aromatic clusters in the model structure was 3.5-5.5/~, and the average distance was 4.1 A.
2.3 Macromolecular Structure of Coal 119
i:..-"
(a) (b)
Fig. 2.33 Molecular model for Upper Freeport coal in a basic cell; (a) shows the gross anisotropic structure, and
(b) shows two cells enclosing the model. [Reproduced with permission from ed. Iino, M.,
Takanohashi, T. et al., Primary and Higher Order Structures of Coal and Their Influence on Coal
Reactivity-Final Report on "Research for the Future" Coal Research Project-, 68 (2001)]
X-ray diffraction measurement by Wertz and Bissell (1994) showed the average distance
between the polycyclic aromatic planes of Upper Freeport coal to be 3.6/~. The value ob-
tained here are slightly higher than that measured by X-ray. Large distances ( > 4.5/k) ob-
served at sites of more strained structures in the model seem to increase the average value.
When a powerful solvent such as the CS2/NMP mixed solvent is used for extraction of
Upper Freeport coal, the solvent can relax the strained structure (as shown in Fig. 2.33) by
breaking of aromatic-aromatic interactions and solvating the molecules released from the
gross structure (Iino et al., 1989; Takanohashi and Iino, 1990). The extraction yield with
pyridine at room temperature is only 2.8 wt% (daD, while the amount of pyridine-soluble
material obtained from fractionation of whole CS2]NMP extract is 29 wt% (daD. This indi-
cates that a considerable amount of solvent-soluble component remains in the raw coal even
when pyridine is used for extraction, suggesting that the coal has an associated structure
(Iino et al., 1989).
scription of the degradation characteristics. Morwell brown coal was treated in 5N HC1
aqueous solution and oxidized using atmospheric oxygen gas at 358 K in an aqueous solu-
tion of Na2CO3. By a series of separations including extraction, washing and drying fol-
lowed by analysis, it was found that aromatic carbon in the residual solid was selectively
oxidized into carboxyls and water-soluble non-aromatic acids such as oxalic acid, acetic
acid, glycolic acid and formic acid as well as carbon dioxide, while the involvement of
aliphatic carbon in the reaction was negligible. It was also found that the mass yield of sol-
vent-extractables from the oxidized coal increased from 0.15 to 0.96 with progress of oxi-
dation. Fig. 2.34 shows the yield from SO (raw coal) and S12 (coal oxidized for 12 h) as a
function of Hildebrand's solubility parameter of single or mixed solvents. It is observed
that the yield depends on the solubility parameter rather than the donor number and reaches
a maximum at 12.0 cal ~ cm-1.5. The same trend was observed for the other oxidized sam-
pies. Although not shown, the yield for nonpolar solvents was 0.05 at most. This indicates
that electron-donor ability is essential for extensive extraction. Since pyridine gives a low-
er yield than methanol-THF mixtures despite its larger donor number, the yield for the
donor solvents with a DN of 20 or more depends not on the donor number but on the solu-
bility parameter.
Based on these results a reaction mechanism, which oxidation converts aromatic car-
bons bonded to bridges into peripheral carboxy groups on the neighboring clusters, and also
converts the other aromatic carbons into non-aromatic acids or carbon dioxide, is illustrated
in Fig. 2.35. In this mechanism, the elimination of a cluster (indicated by 3) accompanies
the formation of carboxy groups at the chain ends of the neighboring clusters (1, 2 and 4).
Further, the lattice statistical method was applied to quantify the degradation process and to
examine the validity of the oxidative degradation method in determining the characteristics
of macromolecular structure of coal (Hayashi et al., 1999b). Further, a particular type of
statistical lattice, Bethe lattices, was used to describe the change in the fraction of solvent-
extractable material and its molecular mass distribution as a function of the cluster (site)
i I i I i I~ ' I i I i
S12
_v 0.8
~ .6
0.4
>. 0.2
0 i I i I i I i I i I i
9 10 11 12 13 14 15
Hildebrand solubility parameter (cal ~ cm-1.5)
Fig. 2.34 Yield of extract for ML-A and 12-h oxidized coal as a function of Hildebrand solubility parameter of
single or mixed solvents. Solvents: (1) THF, (2) THF-methanol (9:1 vol/vol), (3) pyridine, (4) THF-
methanol (8:2), (5) THF-methanol (7:3), (6)DMF, (7) THF-methanol (6:4), (8) THF-methanol (5:5),
(9) THF-methanol (3:7), (10) methanol. [Reproduced with permission from Hayashi, J.-i. et al.,
Energy Fuels, 13, 76 (1999)]
2.3 Macromolecular Structure of Coal 121
O O
OH HO
Oxidation ..-"- O CO2
~..~ (COOH)2
HCOOH
CH3COOH
etc.
Fig. 2.35 Elimination of aromatic cluster and conveesion of inter-cluster bridge into carboxyl group by oxidation.
[Reproduced with permission from Hayashi, J.-i. et al., Energy Fuels, 13, 1231 (1999)]
elimination. It was found that the model could reasonably describe the observed increase in
the fraction of DMF-extractable material and its molecular mass distribution as a function
of the fractional loss of aromatic clusters that are recognized as sites of the Bethe lattice.
The coordination number estimated by the model, 2.2, may provide a simplified macromol-
ecular structure of the coal, in which 80% of the clusters are connected to two bridges while
the remaining 20% is connected to three bridges. If the latter clusters can be recognized as
the cross-linking (or branching) points in the macromolecular network, the model estimates
the average number of repeating units between the points to be about five. Although there
are great difficulties in this approach, the reader is reminded that the work described is only
the beginning.
c5
Pure water
O
9 o
YL
o.
tr
LY
.t=
O
X MW
SB
BZ
WY
IL
I BL I I
Fig. 2.36 DSC thermograms of the coal samples and pure water.
[Reproduced with permission from Norinaga. K. et al., Energy Fuels, 12, 576 (1998)]
and Naas-O'Rourke, 1979; Barton and Lynch, 1994). For LY, MW and a brown coal,
peaks are seen around 258 K and 226 K. The peaks around 258 K indicate the congelation
of water with properties nearly identical to those of bulk water, while those around 226 K
are attributed to the congelation of water condensed in pores of diameter less than several
micrometers. It is also observed that the other coals retain the latter type of water exclu-
sively. No exothermic peaks are detected at temperatures lower than 213 K. Fig. 2.37 il-
lustrates the relationship between the quantity of heat evolved, AH, and the water content
for selected coals. The partially dried samples were prepared from the original samples by
allowing them to lose water slowly at ambient temperature in a nitrogen atmosphere. For a
coal, A H decreases linearly with decreasing water content ranging from 1.3 to 0.6 g/g-daf
coal, where the exothermic peak around 258 K diminishes with the extent of drying. The
slope of 333 J/g-water is in good agreement with the congelation heat of bulk water, 334
J/g. Thus, water desorbed in this range is ascribed to water having no specific interaction
with the coal and hereafter referred to as "free water". For water contents ranging from 0.6
to 0.3 where the peak around 226 K diminishes, A H decreases with a slope of 188 J/g. The
peak is due to the congelation of water condensed in pores and referred to as "bound wa-
ter". Table 2.6 lists the content of free and bound water of four selected coals. It should be
noted that the sum of these two types of water accounts for only 35-78% of the total water
content. This indicates the presence of another type of water that does not undergo conge-
lation as a primary phase transition; hence, this type of water is referred to as "non-freez-
able water". The DSC analysis also revealed that the desorption of free, bound and non-
2.3 Macromolecular Structure of Coal 123
250
~)YL~B~ / ~
o 200
150
100
50
0 n I d1 l. i i [ I - i I I I , I , I ,
(b) A BZ
40 [] SB
"~ 9 IL
30
|
exl)
20
"~ 10
0 , ~ , i , _ , , I ,
Table 2.6 Contentsof Different Types of Water Determined by DSC and 1H-NMR
Method DSC NMR
Free Bound Non-freezable Free Bound Non-freezable
Coal (g-H20/g-mf coal) (g-H20/g-mf coal)
YL 0.70 0.35 0.30 0.74 0.34 0.27
SB 0 0.26 0.20 0 0.24 0.22
BZ 0 0.17 0.31 0 0.17 0.31
WY 0 0.15 0.24 0 0.15 0.24
[Reproduced with permission from Norinaga. K. et al., Energy Fuels, 13, 1058 (1999)]
ing Gaussian decay in the transverse relaxation. Moreover, if non-freezable water is still in
a mobile liquid-like state even when the other types of water are frozen, it can be observed
as a slowly decaying component and distinguished from the others. In the analysis of each
original coal sample, proton transverse relaxation was measured using a solid-echo pulse
sequence at temperature intervals of 2-5 K while cooling from 293 to 213 K. The relax-
ation signal at each temperature was recorded after temperature equilibration. It was found
that the amount of mobile proton of water decreases rapidly at 273-263 K for LY, MW and
YL coals but slowly at 263-213 K for all coals. The decrements in the former and latter
temperature ranges based on the mass of water agree well with the amounts of free and
bound water, respectively, as shown in Table 2.5. In addition, for each coal, the amount of
proton that is mobile even at 213 K is in good agreement with that of non-freezable water.
By means of NMR using a solid echo pulse sequence, the change in molecular mobility
of coal macromolecules was investigated (Norinaga et al. 1998b). Fig. 2.38 shows the con-
tent of mobile coal hydrogen, CM,, as a function of the water content w for YL, BZ and WY
coals. It is observed that CMH decreases linearly with a decrease in the content of non-freez-
able water at a rate of about 0.5 mol-H/mol-H20 while it is unchanged by the loss of free
and bound water. This suggests that a portion of coal hydrogen is mobilized in the NMR
sense due to solvation by non-freezable water molecules and their desorption renders the
hydrogen immobile. Further, it was also found that after sequential drying, deuteration and
swelling in D20 vapor of coal resulted in changes in the molecular mobility of coal. The
mobile hydrogen remaining after drying consists entirely of hydroxylic hydrogen. When
the deuterated YL coal is swelled in D20 vapor, CMH increases to 7 mol while no hydroxylic
hydrogen is involved, indicating the mobilization of other types of hydrogen. This method
can also be applied to the analysis of coals swollen in deuterated organic solvent and the
extent of mobilization of hydroxylic hydrogen can be measured (Norinaga et al., 2000).
In general, when partially or completely dried brown coal or lignite is exposed to wa-
ter, it swells but often does not regain its original volume. This irreversible change in vol-
ume in the cycle of water removal and swelling was examined. Based on freezing property
of pore condensed water, Norinaga et al. (1999) proposed a pore model to evaluate the ef-
fect of predrying on the porous structure of water-swollen coal. Using the Gibbs-
Thompson equation, the freezing point temperature, Tr, of water condensed in a microspace.
is a function of its size. Hence, Tf of bound water could be converted to the size of pores
=6
.o ~
~D
e o:
~9 4 o
O Q_o
O
~-, 2 ~2
0 , , | I , , , I , , , i , , , O |
0.0 0.4 0.8 1.2 1.6 0.0 0.1 0.2 0.3 0.4 0.5 0.6
w (kg-H20/kg-mfcoal) w (kg-H20/kg-mfcoal)
Fig. 2.38 Quantityof heat generated by congelation as a function fo water content:
(a) YL; (b) BZ, SB, and IL. [Reproducedwith permission from Hayashi, J.-i. et al.,
Energy Fuels 15, 908 (2001)]
2.3 Macromolecular Structure of Coal 125
where the water resides. For water confined in micro- and mesopores of solid materials, the
Tf is related to the pore dimension (Ishikiriyama et al., 1995a, 1995b) (diameter for cylin-
drical pore and width for cylindrical pore) as
Dp = a ] (273.15 -- Tf) + 213 (2.3)
where a is the constant depending on porous material and can be determined analytically
from the contents of non-freezable and bound water, fl is the thickness of the layer of water
molecules that acts as a shield between the pore surface and the core of ice atom-bound wa-
ter. Table 2.7 shows the average pore dimensions calculated by the model as a function of
w. Regardless of the pore shape, the model could explain the irreversible decrease in the
volume of pore water by that in the pore dimension, i.e., the shrinkage of pores.
Table 2.7 Effect of the Extent of Drying on Dimension of Pores
after Swellingin Water
In recent years, porous structures of solid materials sorbing water have been investigat-
ed by 'H-NMR employing Carr-Purcell-Meiboom-Gill (CPMG) pulse sequence. The
CPMG method can realize transverse relaxation of water protons with relatively long relax-
ation times by removing the effect of the magnetic field in homogeneity upon linewidth and
reducing the diffusion term, which is manifest in the spin echo sequence (Farrar and
Becker, 1971). The pore size distributions for porous ceramic materials have been estimat-
ed by analyzing the relaxation characteristics of water protons sorbed in the materials based
on a theoretical relationship among the longitudinal or transverse relaxation time constant
(T1 or T2) for bulk water and for pore water and the dimension of pores (Brownstein and
Tarr, 1979; D'Orazio et al., 1989; Halperin and Jehng, 1994). The NMR analysis provides
the size of pores in moistened coals and allows the elimination of the above-described pore
model.
Figure 2.38 exhibits the estimated changes in pore dimensions, dpc (diameter when pore
is cylindrical) and dps (thickness when pore is slit-like in shape), upon drying for BZ and a
coals (Hayashi et al., in press). Table 2.8 compares the dimensions for YL (w = 0.64) and
BZ (w -- 0.53) with those calculated by Eq. 2.3. It is seen for both coals that dps agrees well
with Dps while dpc is appreciably greater than Dpc. This is evidence that the pores are slit-like
rather than cylindrical in shape, dps decreases linearly with decreasing content of bound
water atom 2.8 to 1.4 nm for YL and atom 2.2 to 1.4 nm for BZ, and decreases further as
the non-freezable water is desorbed. A structural model of moist coal in which separation
of hydrophobic and water impervious phase and pore water phase are separated on the
nanoscale, is shown in Fig. 2.39.
126 2 Chemical and Macromolecular Structure of Coal
[Reproduced with permission from Hayashi, J.-i. et al., Energy Fuels 15, 908 (2001)]
Non-freezable water
dps
Bound water
Fig. 2.39 Model of slit-like pores of moistened coal. The mobile coal protons age identical with those of hydrox-
ylic groups for YL with w - 0.30 and BZ with w --- 0.31. All of the hydroxylic groups are solvated by
pore water at pore surface. Bound water is not distinguished from non-freezable water unless it freezed.
3
Pyrolysis
coal macromolecule induced by thermal energy that leads to the formation of free radicals.
Gases and tars will be produced when the thermally induced formation of free radicals is
accompanied by their prompt stabilization by an externally added, or by an internal, hydro-
gen source. The liquid phase hydrogenation of coal (or liquefaction) is also closely related
to pyrolysis. A significant difference between thermal decomposition in the gas phase (py-
rolysis or hydropyrolysis) and in the solvent lies in the size of the fragments removed from
the coal matrix. The size of the fragments produced in the gas phase is limited by their
volatility, while in the solvent the escaping fragments can be substantially larger.
Consequently, a significantly larger amount of char is usually produced in the pyrolyses.
Furthermore, in the absence of hydrogen atom donors, the radicals may instead undergo
"condensation" reactions which lead to char that is even more cross-linked and refractory
than the starting coal.
Pyrolysis is the most important reaction in coal technology since it is the basic process
for the manufacture of coke, tar and gases from coal and the initial and accompanying reac-
tion of coal hydrogenation, combustion and gasification. Since pyrolysis proceeds under
rather mild reaction conditions, low temperature and low pressure, attention has been paid
to the recovery of liquid products in high yields by utilizing pyrolysis: this has been called
"mild gasification of coal" or "skimming of coal" (Babu et al., 1990). Pyrolysis for this
purpose is performed at rather high heating rates of over 1000 K/s, and is called "flash py-
rolysis". Fig. 3.1 shows schematically how flash pyrolysis of coal proceeds (Miura, 2000).
/ o~,
o;~ t,.Y."I
/Stabilization
, /ofra ,ca, s
~o ~ |
/5 9, ,,,J.~ + ~ o ~ ' "~ o4"6~-c]" ~ Polymerization and
~ L~176 " < ~ ~ ~ ........~ ~ ~ ' \ condensation reactions
,'
/ ,~ , 2 H2
Coal particle ,,/ / ~ ~ ~ . ~ o CO
,'"
4t~H2
1.5 I I I I , [
1
J
,r t
/
/
/
1.0
0.5
0
800 1000 1200
Temperature (K)
Fig. 3.2 Ratio of TVM to PVM as a function of pyrolysis temperature (Miura, 2000). Circles and bars represent
data of Miura; (solid bar) Menster; (oblong with dots) Solomon; (big broken bars) Tyler; (oblong with
diagonal lines) Scott; (tiny broken bars) Teo.
[Reproduced with permission from Miura, K., Fuel Process. Technol., 62, 122, Elsevier (2000)]
Flash pyrolysis consists of two sets of reactions: primary devolatilization reactions and sub-
sequent secondary gas phase reactions. The former reactions are very rapid reactions which
consist of radical formation reactions, polymerization-condensation reactions, radical re-
combination reactions, hydrogen addition reactions, etc., and the latter reactions are decom-
position reactions of the volatile products produced through the primary reactions. Flash
pyrolysis has attracted much attention as a method for pr+ducing liquid products because it
is known to increase the total volatiles (TVM) over the volatile matter of the proximate
analysis (PVM). Fig. 3.2 shows the ratio of TVM to PVM against pyrolysis temperature
(Miura, 2000), The ratio increases from 1.05 to 1.45 above a pyrolysis temperature of 800
~ Although the mechanism by which TVM is increased is not clear, it is clear that the pri-
mary reactions must be controlled to further increase TVM. The secondary gas phase reac-
tions solely change the distributions of TVM which are important for increasing the yields
of some specialized products.
Many attempts have been made to increase the yields of tar and aromatic compounds
such as benzene, toluene and xylene (BTX) through controlling either the primary reactions
or the secondary gas phase reactions. The most commonly employed method is the so-
called flash hydropyrolysis, under a high hydrogen pressure of over 20 bars at high temper-
atures of over 900 ~ This method is actually effective for increasing both TVM and the
BTX yield significantly (Borrill and Noguchi, 1989), but it requires expensive hydrogen
and rather severe reaction conditions. Much milder reaction conditions are needed to meet
the demand for energy saving and economical coal conversion processes. The following
pyrolysis methods have been performed to increase TVM and/or the BTX yield under
rather mild experimental conditions.
3.1.1 P y r o l y s i s in R e a c t i v e G a s A t m o s p h e r e s
These methods are intended to supply CH3 and/or OH radicals to coal fragments in addition
to H radicals by performing pyrolysis in methane (Smith et al., 1989), toluene (Doolan and
130 3 Pyrolysis
Mackie, 1985) or methanol (Cing et al., 1987) atmosphere. However, they were not able to
produce a significant increase in TVM probably because the rate of the supply of radicals
from the gases did not match the formation rates of coal fragments (Miura, 2000).
3.1.2 P y r o l y s i s of P r e t r e a t e d Coal
Several methods of pyrolyzing coals pretreated with various gases such as H2, He, CO2,
H20 (Cypres and Baoqing, 1988) and hydrogen donor (Huettinger and Sperling, 1987) were
proposed, but the effectiveness of the pretreatment has not been elucidated. For example,
Graft et al. (1987, 1989) proposed that pyrolyzing the coal pretreated under 50 bar and 320
~ to 360 ~ of steam increases both TVM and the liquid yield significantly. The increases
are presumed to occur through the breakage of ether linkages in coal during the pretreat-
ment. However, such a significant effect has not always been detected in spite of reexami-
nations of the method by many investigators. The effect is believed to be significantly coal
rank dependent (Serio et al., 1992).
The trial of Ofosu-Asante et al. (1989) is unique. They alkylated coal by the so-called
O-alkylation method to replace OH groups by alkyl groups. By pyrolyzing the pretreated
coal they realized significant increase of TVM. It is also reported that TVM can be in-
creased slightly when ion-exchangeable Ca was removed before pyrolysis for low rank
coals (Franklin et al., 1983). The success of these methods in increasing TVM is realized
by suppressing the cross-linking reactions between OH groups during the pyrolysis.
3.1.3 Catalytic P y r o l y s i s o f C o a l
Several trials have been reported on catalytic flash pyrolysis of coal, but the catalysts are
generally effective only in secondary gas phase reactions. However, catalysts are of course
effective in slow pyrolysis performed at rather low heating rates in the presence of high
pressure of hydrogen (Snape et al., 1989).
3.1.4 P y r o l y s i s M e c h a n i s m o f Coal
Although pyrolysis is implicit in all coal conversion processes operating at temperatures
above 673 K, such as combustion, liquefaction, gasification, carbonization, etc., some re-
cent advances have been reviewed (Gavals, 1982; Fynes et al., 1980). Pyrolysis dispropor-
tionates coal in a technically simple manner (although this involves very complex physico-
chemical mechanisms) into hydrogen-rich tar and gas together with a hydrogen-deficient
char or coke with minimal use of energy. Moreover, the breakdown of the coal structure
release part of the organically bound sulfur, oxygen and nitrogen as volatiles which may be
collected in environmentally acceptable ways. Considering the complexity of even so-
called simple pyrolysis reactions (Albright et al., 1983) and the complexity of the coal
structure, it is not surprising that part of our lack of understanding of coal pyrolysis mecha-
nisms stems from a poor understanding of the structure of coal (Meyers, 1982). This is
compounded by difficulties in following rapid changes on the molecular scale within
porous solids and in characterizing the high molecular mass primary and intermediate py-
rolysis products. In recent years, mathematical modeling of the mechanisms and kinetics of
coal pyrolysis and hydropyrolysis has attracted academic interest because of its theoretical
basis, and industrial interest because of its application to systems designs. The advances in
kinetic modeling are described below.
Jungten and Heek (1979) have modeled nonisothermal pyrolysis reactions in terms of
volatile evolution and compared model predictions with reality; their studies were extended
to hydropyrolysis formation of light aromatics by postulating two separate BTX-forming
3.2 Pyrolysisof Coal Tar 131
reactions of first and second order (Bunthoff et al., 1983). Their results have been summa-
rized (Heek and Jungten, 1984).
Reidelbach and Summaerfield (1975) developed a general model of pyrolysis in vari-
ous reactors and expanded the model to cover a wide range of conditions. A model for en-
trained-flow reactors with solid heat carriers and radiative heat transfer, where particle size
distribution is of particular importance, was developed (Reidelbach and Algermisser, 1978).
Halchuk et al. (1983)coupled the physical and chemical mechanisms of pyrolysis with
the micropore structure controlling the tar transport. Unger and Suuberg (1983a, b) found
that mass transfer limitations existed under all the conditions they studied and only at their
limits of high temperature and high vacuum did the tar quality approach that of solvent ex-
tractable residue. This led to a model based on simultaneous tar formation, evaporation and
polymerization which was tested by correlating reaction conditions with changes in molec-
ular mass. Simons (1983) stressed the importance of fluid transport in mathematical mod-
els, suggesting that the transport was driven by internal pressures of up to 3000 bar through
pores arranged either in a tree or in a random mode. When coal melts, bubble transport is
obviously important and the effect of limited mass transfer of hydrogen into coal when it
softens under hydropyrolysis conditions (with rapid heating) was highlighted by Schaub
et al. (1981).
Goyal and Gidaspow (1982) developed a one-dimensional model which fitted the data
from Cities Service R&D pyrolysis of subbituminous coal. They assumed that coal consist-
ed of 11 solid species with nine gaseous species being formed, and obtained 53 nonlinear,
first-order differential equations. Sitnai (1976) developed a flash pyrolysis model for opti-
mizing tar yields from Australian coals at atmospheric pressure; a model for the combined
drying and devolatilization of lignite in a fluidized bed has been constructed (Agawal,
1984). Solomon et al. (1981) used FTIR analysis of the structural elements of coal to de-
velop a kinetic model based on distributed activation energies to predict temperature- and
time-dependence of volatile evolution.
Modeling coal pyrolysis and hydropyrolysis reactions is a useful exercise since it con-
centrates the mind on the process patterns and enables such patterns to be tested against real
results. However, our limited understanding of the complex mechanisms involved requires
that, for practical purposes, such as the scale-up of a new pyrolysis process, the normal pro-
cedure of going from conceptual idea to bench-scale apparatus to process demonstration
unit to pilot plant should be adhered to before an industrial-scale plan is built.
However, new technologies have been investigated to improve the quality of coke and to
save high-grade caking coal. One of the most effective methods is to utilize a binder mater-
ial such as coal tar or coal tar pitch. The binder material is mixed with the raw coal and
charged into the coke oven.
In the matter of raw coal, only a limited group of coals, as defined by rank, type, and
grade, is used for coking. Bituminous rank coals are used commercially to produce coke.
Rank refers to the maturity of coal; lower rank coals have excessive oxygen content and
produce unstable metaplast, which decomposes before the temperature of coal plasticity is
reached. These coals are sintered but are not coked. Coals of rank higher than bituminous
produce insufficient metaplast to produce well-fused coke.
For many years, there has been widespread interest in new coke-making processes
(Bujnowska, 1992; Bhatia, 1992; Serio et al., 1987a). Worldwide depletion of good quality
coking coal is the one single factor which has acted as the driving force for the develop-
ment of many advanced technologies in the field of coke making. Most of the proposed
processes use noncoking coals or coal chars with various binders, such as coal tar or pitch.
Some of the better known processes are: FMC, DKS, AUSCOKE and CTC (Bujnowska,
1992). These new technologies are attractive because they do not dependent on good cok-
ing and costly coals but use binder material. These processes are in various stages of devel-
opment, and it is only a matter of time, politics, and/or economic pressure before coke-mak-
ing by these new processes will be commercialized on a large scale. In these processes, the
yield and composition of coal tar produced from such new pyrolysis processes using coal
tar as binder material have not yet been made clear, and the development of a technique to
estimate the product yields and composition, such as the presence of naphthalene, in coal
tar is needed.
At present, in treating coal tar, the recovered crude tar is distilled in a distillation tower
to give the five standard fractions in order of ascending boiling ranges: light oil, naphtha-
lene oil, creosote and anthracene oil, and pitch (Eisenhut, 1981). The light oil fraction
(boiling range: ca. < 195 ~ contains benzene, tar acids and tar bases and resembles crude
benzole. The naphthalene oil fraction (boiling range: ca. 195-230 ~ contains naphthalene
and a range of tar acids and tar bases. The tar acids and bases are extracted by washing
successively with alkali and acid. The neutral fraction usually contains about 30 % naph-
thalene, which is removed either by crystallization and hot processing, continuous fraction-
3.2 Pyrolysisof Coal Tar 133
ation, or by crystallization and centrifugal washing. Pure naphthalene is used to make ph-
thalic anhydride, an intermediate for plasticizers, polyesters and resins. Although petrole-
um-based phthalic anhydride obtained via o-xylene has largely replaced this source, naph-
thalene is an invaluable component in coal tar. The creosote fraction (boiling range: c a .
230-300 ~ contains substituted naphthalene, the higher boiling tar acids and base oils.
The components are not usually separated, and creosotes are used in bulk for timber presen-
tation and for bending with pitch to make road tar. The anthracene oil fraction (boiling
range: ca. 300-350 ~ contains mainly polynuclear hydrocarbons, such as anthracene,
phenanthrene and pyrene, and high boiling tar acids. These oils are used much like cre-
osote, and they may be a good source of carbon black oil. The pitch residue is used gener-
ally as a binder, e.g., to hold together the grist in electrodes for producing aluminum from
bauxite and the arc steel furnace, as a briquetting binder. It is also used in starting material
for green pitch cokes. In the end, the amount of naphthalene in the naphthalene oil fraction
mainly defines the value of coal tar and becomes the center of interest in the coal tar indus-
try.
3.2.2 C o a l Tar P r o p e r t i e s
Coal pyrolysis in a coke oven consists of two reactions in series (Bhatia, 1992; Serio et al.,
1987a; Pitt and Millward, 1979). When coal is heated, primary tar is produced by the pri-
mary pyrolysis of coal. As temperature is increased, the primary tar must pass through a
high temperature zone, and secondary pyrolysis of tar in the gas phase occurs. If coal tar is
used as the binder material, the secondary pyrolysis of added coal tar also occurs. In this
process, the development of a technique to estimate the product yields and composition of
coal tar is needed to meet the demand for coal tar derivatives as a source of energy and
chemical feedstock. In particular, the composition of light fractions obtained by distillation
such as light oil, naphthalene oil, creosote oil and anthracene oil are important as chemical
feedstock.
Many researchers (for example, Tayler, 1980, Collin et al., 1980, Stiles and Kandiyoti,
1989, Katheklakis, 1990 and Gonenc, 1990) have investigated the effects of reaction time
and temperature of coal pyrolysis and secondary pyrolysis of coal tar on the yield and mole-
cular composition of tar. Hayashi et al. (1992, 1993) studied changes in the molecular
structure of tar in a fluidized-bed reactor divided into two regions: a dense bed for the pri-
mary reaction and a freeboard for the secondary reaction in the gas phase. Since, in flu-
idized-bed pyrolysis, secondary reactions are known to occur in the dense bed to some ex-
tent (Fletcher et al., 1990), the behavior of pyrolysis of tar cannot be analyzed precisely.
Solomon et al. (1990) studied the pyrolysis of a low-lank coal by minimizing the secondary
reaction. It was difficult to obtain the desirable tar yield and composition by controlling the
primary reaction of coal alone. Thus the secondary pyrolysis of coal tar should be con-
trolled as well. However, secondary pyrolysis of coal tar has not been investigated suffi-
ciently. Xu and Tomita (1989) and Serio et al. (1987) separated the primary and secondary
reaction zones by using two-stage reactors which include a fixed bed of coal and a tubular
reactor connected downstream. However, they could not exclude the possibility that the
yield and composition of tars evolving from the fixed bed reactor would reflect some influ-
ence of secondary pyrolysis within the coal particles prior to tar entrainment in the carrier.
Since coal tar is a complicated mixture, the chemistry involved in the pyrolysis of coal
tar is exceedingly complex. The dominant compounds in the coal tar are aromatic hydro-
carbons with one to eight tings (Solomon et al., 1990; Xu and Tomita, 1989). Naphthalene
is the most simple model compound of coal tar. Many insights into the mechanism of py-
134 3 Pyrolysis
rolysis and carbonization have been obtained in studies using naphthalene. Badger et al.
(1964) studied the pyrolysis of naphthalene at 700 ~ It was concluded that carbon-hydro-
gen fusion gives naphthyl radicals, which react with naphthalene to yield binaphthyls, and
that cyclodehydrogenation of the binaphthyls leads to the perylene and benzofluorathenes.
More recently, Jess (1996) reported the kinetics of the thermal conversion of aromatic hy-
drocarbons in the presence of hydrogen and steam, using a naphthalene, toluene and ben-
zene as model compounds at 700-1400 ~ The mechanisms of primary and consecutive
reactions are presented as reaction schemes that are supported by kinetic calculations. The
reactivity decreased in the order: toluene > naphthalene > benzene. Lewis (1980) de-
scribed general mechanisms for the pyrolysis and carbonization of naphthalene, dimethyl-
naphthalene, and dibenzothiophene. Fitzer et al. (1971) gave a detailed mechanism for the
pyrolysis of acenaphthylene. Studies using these aromatic compounds showed that poly-
merization through loss of side chain and hydrogen was the main chemical reaction. Stable
free-radicals are formed during the polymerization process. A continual increase in molec-
ular weight through polymerization and loss of low molecular weight volatiles results in the
transformation to mesophase, coke and ultimately carbon. These studies on pyrolysis and
carbonization of model compounds help in predicting the type of thermal phenomena in-
volved in coal tar and pitch pyrolysis. However, at present it is impossible to predict the re-
activity of compounds such as naphthalene in coal tar because there are so much complex
components.
3.2.3 Pyrolysis Mechanism of Coal Tar
Coal tar is a complex mixture consisting of a variety of compounds of different functionali-
ty and wide ranging molecular weight. The dominant compounds are polycyclic aromatic
hydrocarbons (PAH) with one to eight tings. In addition to PAH, heterocyclic compounds
containing oxygen, nitrogen and sulfur are also present in small amounts. Although the re-
action of a representative model compound provides useful information on the mechanism
of coal tar pyrolysis, it is still difficult to predict the actual reactivity of coal tar.
On the other hand, the pitch residue of coal tar is a versatile starting material for carbon
products of higher commercial value. It is well known that the properties of high perfor-
mance carbon materials are affected greatly by the molecular structure of raw coal tar pitch
(Patrick et al., 1983; Marsh et al., 1977; Mochida, 1981). The molecular structure of coal
tar is affected not only by the primary pyrolysis of coal but also by the secondary pyrolysis
(Kabe et al., 1998). In designing the properties of high performance carbon materials, it is
important to elucidate the molecular structure of such a heavy fraction and its changes in
the secondary pyrolysis of coal tar.
Takeuchi et al. (1994) published a method of functional group analysis based on 1H-
NMR and elemental analysis data to characterize the structure pitch and the concentrations
of the functional groups in the pitch. The details of the method are described in Section
3.3. lB. The premise of the method is that the structural pattern of pitch is divided to func-
tional groups, i.e., condensed aromatic with one-ring to eight-ring, alkyl chain, methylene
bridge, and hydroxy and hydroaromatic groups. On the other hand, secondary pyrolysis of
coal tar yield from commercial coke oven was investigated to provide pyrolysis data which
are useful for the estimation of the secondary pyrolysis occurring in the coke oven and to
improve the yield of desirable compounds (Godo et al., 1998a, b). The results are discussed
below.
3.2 Pyrolysisof Coal Tar 135
80 . 80 b: 800 ~ 80
c: 900 ~
60 ~- 60 60
40 ~ 40 40
20 ~ 20 >= 20
0 0 ' o
0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25
Residence time (sec) Residence time (sec) Residence time (sec)
Fig. 3.3 Effectof residence time on product yields of tar pyrolysis.
O: THFI A: CFI-THFS D: HI-CFS
O: HS A: Gas + H2P
[From Godo. M. et al., J. Jpn. Inst. Energy, 77, 122 (1998)]
2.0
N2:90 ml/min
1.5
1.o
0.5
0.0
700 800 900
Temperature (~
Fig. 3.4 Effectof temperature on gas yields in the pyrolysis of tar.
O: H2 A: CH4 D: CO2
O: C2H4 A: C2H6 m: C3H6
[From Godo. M. et al., J. Jpn. Inst. Energy, 77, 123 (1998)]
136 3 Pyrolysis
Figure 3.4 shows the effect of temperature on gas yields at the gas flow rate of carrier
nitrogen of 90 ml/min (residence time: 9-11 s). The gaseous products in the secondary py-
rolysis of coal tar were identified to be H2, CH4, C2H4, C2H6, C3H6 and CO2, with the main
products being H2 and CH4. CH4 by cracking and dealkylation of hydrocarbons with
methyl groups increased monotonically with rise in temperature. H2 by dehydrogenation of
heavier hydrocarbons, which are thermodynamically favored, was promoted markedly at el-
evated temperatures from 800 to 900 ~ This shows that the dehydrogenation reaction in
the secondary pyrolysis of coal tar becomes more important at temperatures over 900 ~
12 12 12
- b: 800 ~ I c: 900 ~
10 10 10
~8 g8
6 ~6 g6
~9 4 ~z 4 ~ 4
2 2 2
0 T I I I I 0 0 I
0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25
Residence time (sec) Residence time (sec) Residence time (sec)
Fig. 3.5 Changesin compositionsin coal tar with reaction conditions.
X: 1-Ring with substituent A: 3-Ring with substituent
O: Naphthalene [-]: 4-Ring without substituent
O: 2-Ring with substituent I1: 4-Ring with substituent
A: 3-Ring without substituent +: 5-Ring without substituent
[From Godo. M. et al., J. Jpn. Inst. Energy, 77, 127 (1998)]
3.2 Pyrolysis of Coal Tar 137
Naphthalene
Perylene B enzefluoranthene
Fig. 3.6 Possible mechanism for the pyrolysis of naphthalene.
[From Godo. M. et al., J. Jpn. Inst. Energy, 77, 128 (1998)]
. n ~ ~ CH3
H3C
~ CH3
H3C
CH3
9Coke
H3C :"~
Fig. 3.7 Mechanism of pyrolysis of the hexane-soluble fraction.
[From Godo. M. et al., J. Jpn. Inst. Energy, 77, 128 (1998)]
41 0 700 ~
t A 800 oc
3~ - I-1 900 ~
~2
~
i I , I i I i I i
0 0 5 10 15 20 25
Residence time (sec)
Fig. 3.8 Changes in the yields of dibenzofuran.
[From Godo. M. et al., J. Jpn. Inst. Energy, 77, 129 (1998)]
0.3
0 700 ~
A 800 oc
I-! 900 ~
0.2
__ 0---
..~
0.1
| I / ,
0.0 ,
0 5 10 15 20 25
Residence time (sec)
Fig. 3.9 Changes in the yields of quinoline.
[From Godo. M. et al., J. Jpn. Inst. Energy, 77, 129 (1998)]
This shows that dibenzofuran is unstable at 900 ~ and that the amount of dibenzofuran is
controllable by the secondary pyrolysis temperature.
Quinoline and carbazole are major compounds among heterocyclic compounds con-
taining nitrogen. Figs. 3.9 and 3.10 show the change in the yields of quinoline and car-
bazole, respectively, in feed with the secondary pyrolysis of coal tar. The amount of quino-
line did not change in comparison with feed tar at 700 and 800 ~ However, quinoline al-
most disappeared in 13 s at 900 ~ In contrast to quinoline, the amount of carbazole de-
creased gradually with lapse of residence time and rise in reaction temperature, and 25%
carbazole remained in feed tar in the pyrolyzed products. This suggests that the nonbasic
3.2 Pyrolysis of Coal Tar 139
1.0
t O 700 oc
/~ 800 oc
0.8 l-1 900 ~
~" 0.6
0.4
0.2
0.0 t i i , i ,
0 5 10 15 20 25
Residence time (wt%)
Fig. 3.10 Changes in the yields of carbazole. [From Godo. M. et al., J. Jpn. Inst. Energy,
77, 129 (1998)]
compounds containing nitrogen of the pyrrole type may be thermally more stable than basic
nitrogen compounds like the 6-membered heterocycles such as pyridine and quinoline.
s 0.4
0.2
0.0 [
15 20 25 30
Carbon number of straight-chain alkanes
perature, the shapes of distribution are very similar to one another with a maximum at C21
or C22. This result suggests that the reactivities of normal alkanes from C~5 to C30 in sec-
ondary pyrolysis of coal tar do not change significantly.
C H 3 ~ . CH3
n ~ CH3 CH3
~ CH3 ~ ~ HCH3 CH
CH3
H
- S
CH
H H
C H ~ CH3 C H ~ CH3
H H
creased. This indicates that the aliphatic substituents, the methylene bridge and the hy-
droxy group decomposed at the initial stage of secondary pyrolysis. However, in long reac-
tion time range at 800 and 900 ~ the amount of Har decreased and those of Ha, H e, and H~
increased in comparison with the short reaction time range at the same reaction tempera-
tures. These reaction conditions corresponded to the significant increase in THFI and the
142 3 Pyrolysis
decrease in the HS fraction. It was suggested that some portion of the HS fraction contain-
ing alkyl-substituted polycyclic aromatic hydrocarbons (PAHs) converted to the HI-CFS
fraction by condensation and remained as aliphatic substituents in the HI-CFS fraction, as
shown in Fig. 3.12.
The structural parameters calculated from hydrogen distribution and elemental analysis
are given in Table 3.5. After the secondary pyrolysis of coal tar in the short reaction time
range, fa increased, and cr and p decreased, indicating that dealkylation and condensation of
the aromatic rings occurred. However, in the long reaction time range at 800 and 900 ~ fa
decreased, and cr and/9 increased in comparison with the short reaction time range at the
same reaction temperatures. This result corresponded to the hydrogen distributions deter-
mined using 1H-NMR spectra data.
,CH3
It was determined the components in the HS fraction of raw and pyrolyzed coal tar us-
ing FID gas chromatography. The majority of the identified compounds are unsubstituted,
alkyl-substituted and hydroaromatic derivatives of one- to five-ring PAHs. This shows that
PAHs of one- to five-rings can be dissolved in hexane and PAHs of more than five rings are
contained in the HI-CFS fraction. The molecules in HI-CFS have a higher molecular
weight than the five-ring PAHs with aliphatic chain. The separation of products in sec-
ondary pyrolysis was achieved by solvent solubility. The HI-CFS fraction is assumed to
contain molecules having almost the same solubility. However, it was clarified that there
was a difference in the concentrations of the functional groups of HI-CFS fraction obtained
by the secondary pyrolysis of coal tar. These results are basically consistent with those ob-
tained from hydrogen distributions and average structural parameters, and provided more
effective information in comparison with the average structural parameters.
Table 3.6 Functional Group Concentrationin HI-CFS
Feed 11.4 14.8 31.8 26.2 10.4 3.29 0.68 0.16 1.00 0.22
700 10.5 5.89 13.8 29.1 31.9 14.5 2.91 0.66 0.20 0.78 0.34
700 16.2 13.5 16.3 22.5 35.0 9.59 2.35 0.25 0.1 1 0.44 0.17
700 22.1 6.63 13.8 20.9 44.3 11.1 2.32 0.25 0.12 0.40 0.14
800 5.6 18.2 20.1 52.8 3.84 1.84 2.14 0.36 0.17 0.38 0.17
800 9.6 8.82 19.8 54.7 11.9 2.10 1.63 0.29 0.13 0.48 0.20
800 14.7 17.0 19.6 49.4 10.3 0.5 1 2.01 0.38 0.18 0.50 0.17
900 4.8 16.9 22.3 48.9 7.00 2.06 1.72 0.37 0.15 0.45 0.09
900 8.8 24.1 29.8 37.9 4.54 0.31 1.95 0.54 0.17 0.50 0.14
900 13.3 8.90 22.3 39.3 17.2 8.38 2.32 0.75 0.26 0.50 0.13
[From Godo. M. et al., J. Jpn. Inst. Energy, 77,239 (1998)l
144 3 Pyrolysis
3.2.4 I s o t o p i c S t u d i e s o f N a p h t h a l e n e R e a c t i v i t y in P y r o l y s i s o f C o a l Tar
As mentioned above, coal pyrolysis consists of two successive reactions. The primary py-
rolysis of coal yields tar containing appreciable quantities of naphthenes and paraffins, of-
ten with alkyl and hydroxy substituents. This primary tar must pass through a high-temper-
ature zone; consequently, secondary pyrolysis of tar in the gas phase occurs and results in
the conversion of the naphthenic structures to aromatics and removal of side chains.
Therefore, coal tar yield and its composition are determined not only by the condition of
primary pyrolysis of coal, but also by that of secondary pyrolysis. If coal tar is mixed with
a raw coal as a binder material, further secondary pyrolysis of added coal tar also occurs.
Many researchers (for example, Stiles and Kandiyoti, 1989; Katheklakis, 1990;
Hayashi et al., 1993; Fletcher et al., 1990; Solomon et al., 1990; Xu and Tomita; 1989;
Serio et al., 1987) have investigated the effects of reaction time and temperature of the coal
pyrolysis on the yield and molecular composition of coal tar. On the other hand, in this sec-
tion, the secondary pyrolysis mechanism of coal tar, in particular, the hydrogen transfer ac-
companying dehydrogeneration during the pyrolysis of coal tar which is the dominant reac-
tion in the pyrolysis, is discussed in detail. The coal tar containing 14C-labeled naphthalene
was used to trace the behavior of naphthalene in the pyrolysis of coal tar (Kabe et al., 1997,
1998). The 14C tracer method is convenient for elucidating the reaction mechanism with re-
spect to carbon in compounds. Furthermore, the hydrogen behavior of naphthalene in the
pyrolysis of coal tar was elucidated using a tritium tracer technique by pyrolyzing coal tar
containing tritiated naphthalene (Kabe et al., 1997, 1998).
80
[] Gas + loss
D HS
60
[ ~ HI-THFS
THFI
~, 40
20
/
Feed Blank 800-50 900-50 950-50
Run No.
Fig. 3.14 Product yield in pyrolysis of tar containing 3H-naphthalene. [Reproduced with permission
from Kabe, T. et al., Fuel, 77, 817, Elsevier (1998)]
3.2 Pyrolysis of Coal Tar 145
soluble fractions, respectively. The amounts of the HS, HI-THFS and THFI fractions in
feed coal tar were 66.9, 31.1, and 2.0 wt%, respectively. Compared with "Feed", it is ob-
served that the preheating treatment in "Blank" scarcely causes the pyrolysis of coal tar.
For the reaction at 800 ~ the amount of THFI significantly increased, while the total
amount of HI-THFS and THFI hardly changed in comparison with the feed tar. At 900 ~
the amount of HI-THFS decreased markedly together with increase in the amount of THFI.
These results suggest the precursor of THFI to be mainly HI-THFS at 900 ~ Although a
considerable portion of the HS was converted into THFI at 950 ~ for 50 s, the HS fraction
was more stable than the HI-THFS fraction.
Figure 3.15 shows the changes in the amount of naphthalene with reaction temperature.
The amount of naphthalene in feed tar was 12.4 wt%. After the pyrolysis, the amounts of
naphthalene decreased with a rise in temperature, and were 9.9, 9.1, and 5.3 wt% at 800,
900 and 950 ~ for 50 s, respectively. The amount of naphthalene at 950 ~ decreased
markedly. This result corresponds to the decrease in the yields of the HS fraction. This
also shows that naphthalene has significant reactivity in the pyrolysis of coal tar under this
condition.
15
10
0
~0 5
<
E
100
80 I
[7 Naphthalene
[~ 3,4-Ring
I I Perylene
= 60 HI-THFS
O THFI
~ 40
r,.)
i i i i i i
20
l i l i l l
~11111
I I I l i i
l i l l i i
Fig. 3.16 14C distribution in the products of tar pyrolysis (Godo, 1998).
9Coke
~ after 50 s, the radioactivity levels of the 3- and 4-ring compounds and perylene were
21.0 and 0.6%, respectively. When the temperature was increased to 900 ~ the level of
radioactivity for the 3- and 4-ring compounds had hardly changed at all. The radioactivity
of the perylene, however, increased and reached 8.7 % at 900 ~ after 50 s. This indicates
that naphthalene can be converted into a 5-ring compound, such as perylene at 900 ~ In
summary, naphthalene in coal tar can be converted into 3- and 4-ring compounds at 800 ~
into a 5-ring compound at 900 ~ and into THFI at 950 ~
From these results, a possible mechanism for the pyrolysis of naphthalene can be pos-
tulated, as shown in Scheme 3.1. The carbon-hydrogen bond scission gives a naphthyl radi-
cal, which reacts with another naphthyl radical to yield binaphthyl. The dehydrocyclization
of binaphthyl leads to perylene, which further condenses to give coke. This indicates that
naphthalene can be intransformed to coke in the pyrolysis of coal tar, although naphthalene
is regarded to be a very stable hydrocarbon compared with substituted aromatic hydrocar-
bons.
C. 3H Distribution in Products
After the pyrolysis, the radioactivity was detected in each fraction. Naphthalene in the HS
fraction was separated and identified by HPLC, and the radioactivity of naphthalene was
measured by the radioanalyzer synchronizing the retention time. The peaks detected after
naphthalene were assigned to HS compounds in Fig. 3.17. Fig. 3.17a shows the 3H distrib-
ution in various fractions of pyrolysis products. The results obtained in the reaction under
3.2 Pyrolysis of Coal Tar 147
/////,
/////, (a)
. . . . . . f l l l l ~
, / / / / . , f lJ. J.J-J
80 .
r
. . . .
~'11111
#.1111~
+ ' / / / / 1 p- #- i i J ~ ~ 1 1 1 1 1
f l l l l # ~ l l J J J
O f / / i l l r
. . . . .
~9 60 ~ ' 1 1 1 l I
8 . 1 1 1 1 J
r i
r
J i z.,
r
~
;5555;
"~ 40 r
:5555;
r
, 1 1 1 1 ~
, 1 1 1 1 ~
.., . . . . I::1:::1::1
~ 2o r l l l l #
::::::::::: ii~i!iii!i!
, 1 1 1 1 ~
i .....................
..j
80 Y'///"~~
}'///z
/ / / / # ~
.o= 60 9 z z z z, / / / / / i
;"///z . . . . . "~
/ / / / / ~
r / / / / /
"~ 40
~ / ./ 4 ./ 4 ~
,..........i
P
20 :...:.1:::: : : -" 9: :." : : ,.'i
!iiiiiii!i ::::ii:::.i!::ii::
i!i!i!ii~! ' i
Feed Blank 800-50 900-50 950-50
Run No.
[]: Naphthalene; D" HS; []" HI-THFS; i " THFI.
the same conditions using 14C-labeled naphthalene described in the previous section B are
shown in Fig. 3.17b.
The total radioactivity yield of tritium recovered from liquid and solid fractions de-
creased with rise in temperature, and was about 52.5% for the reaction at 950 ~ while the
yield of 14C for the same reaction was about 80%. This shows that about one-half the hy-
drogen in the initial naphthalene was released into gas phase through the dehydrogenation
in the polymerization of compounds which contained the tritium atom. Almost all tritium
radioactivity in the products at 800 ~ was distributed in naphthalene and the HS fraction.
The distribution of tritium radioactivity in naphthalene remarkably decreased and that in
HS, HI-THFS and THFI increased with a rise in temperature. However, in comparison
with the distribution of 14C radioactivity, the amount of tritium radioactivity in HS in-
creased and that in THFI decreased at 950 ~ Generally, it was assumed that the tritium
distribution in each fraction reflected the extent of condensation and dehydrogenation be-
tween naphthalene and the heavier component. This assumption means that the amount of
tritium distribution in some fractions derived from naphthalene is always smaller than that
of 14C, because some portion of the hydrogen is released into gas phase by dehydrogena-
tion. However, the distribution of radioactivity of tritium in HS at 950 ~ was much higher
than that of 14C, indicating that hydrogen in naphthalene can exchange with the hydrogen in
148 3 Pyrolysis
compounds of HS. The fact that the distribution of tritium radioactivity in naphthalene is
smaller than that of ~4C radioactivity also demonstrates the hydrogen exchange of naphtha-
lene. It is assumed that the hydrogen in naphthalene exchanges not only with HS but also
with heavier compounds such as HI-THFS and THFI. However, the hydrogen exchange of
the heavier compound from the tritium distribution could not be estimated because the de-
hydrogenation reaction predominates in such highly polymerized fractions.
~9 a 80-
O
60
E
.~ 40-
"~= 20-
>, o
(b)
~-
.,.~ 80-
~ 60-
"~N 40-
,,~
= 20-
1
The coal tar used in the study contained unsubstituted, alkyl-substituted and hydro-
genated derivatives of PAHs. If dealkylation of alkyl-substituted PAHs and dehydrogena-
tion of hydrogenated derivatives of PAHs with two rings occur, the amount of naphthalene
increases. It is assumed that the amount of naphthalene in pyrolyzed coal tar is defined by
the sum of the amount of unconverted naphthalene in the feed tar and the amount of naph-
thalene generated.
T T T
R. + 9 + RT (q.l]
T T T ~ T
T T T T
(3.2)
T T T H
T T
T T T
T
(3.3)
T ~ T
T T
T T T T
+ HT (3.4)
Scheme. 3.2 [Reproduced with permission from Kabe, T. et al., Fuel, 77, 819, Elsevier (1998)]
150 3 Pyrolysis
tios of the ~/mount of naphthalene and its radioactivity of tritium in the recovered tar were
identical at 800 ~ showing that the generation of naphthalene and the hydrogen exchange
of naphthalene scarcely occur in the pyrolysis of coal tar at 800 ~ for 50 s. In contrast, at
900 and 950 ~ the ratios of tritium radioactivity of naphthalene were smaller than those of
~4C radioactivity of naphthalene. At 900 ~ the ratio of tritium radioactivity of naphthalene
was about one-half that of the recovered amount of naphthalene, although the ratio of 14C
radioactivity of naphthalene was about two-thirds the amount of naphthalene recovered. At
950 ~ the ratio of tritium radioactivity of naphthalene was about one-third that of the re-
covered amount of naphthalene, although the ratio of ~4C radioactivity of naphthalene was
about two-thirds the recovered amount of naphthalene. The difference between the ratios
of tritium and 14C radioactivity can be attributed to the presence of hydrogen exchange in
naphthalene and other compounds in the coal tar. The results indicate that hydrogen mobil-
ity was promoted with a rise in temperature over 900 ~
From the above results, a possible mechanism for the pyrolysis of naphthalene and the
hydrogen exchange reaction between naphthalene and PAHs are shown in Scheme 3.2, Eqs.
(3.1)-(3.4). In this sequence, anthracene is selected as the model compound of PAH be-
cause the coal tar used in the study contains a large amount of anthracene. The coal tar
used also contains compounds with functional groups such as a hydroxy group which gen-
erates radicals Ro easily. When an R~ is formed under the operative conditions, a 3H-naph-
thyl radical will be formed by the reaction of 3H-naphthalene with an R~ as shown in Eq.
(3.1). Further, if the 3H-naphthyl radical is quenched by a PAH such as anthracene to form
naphthalene, the hydrogen exchange of naphthalene is complete, as shown in Eq. (3.2).
The hydrogen exchange between 3H-naphthalene and PAHs can proceed via radical hydro-
gen transfer reaction depending on the concentration of the naphthyl radicals. A 3H-naph-
thyl radical reacts with another radical such as an anthryl radical to yield 9-1-naphthylan-
thracene in Eq. (3.3). The dehydrocyclization of 9-1-naphthylanthracene leads to the for-
mation of a 1,2-benzoperylene and HT molecule in Eq. (3.4). Therefore, the formation of
the naphthyl radical in this system may be the rate-determining step for both the polymer-
ization of naphthalene into more condensed PAH and the hydrogen exchange.
3.2.5 I s o t o p i c S t u d i e s o f H y d r o g e n M o b i l i t y in C o a l Tar
Coal tar is a complex mixture which consists of a variety of compounds with different func-
tional groups and has wide ranging of molecular weight. Functional groups with het-
eroatom in coal tar such as hydroxy groups, imino groups, etc. play a critical role in the py-
rolysis of coal tar because they constitute the more polar fraction of the coal tar and stabi-
lize free radicals (Attar and Hendrickson, 1982; Shinn, 1984). Consequently, it is very im-
portant to know the forms in which they appear in coal tar and their precise concentration in
coal tar for clarifying the very complex structure of the coal tar and for developing the con-
version techniques of coal tar.
Oxygen in coal tar is present mainly in the hydroxy group and such oxygen functional
groups have been indicated to play an important role in the pyrolysis of coal tar (Surygla
and Sliwka, 1994). On the other hand, nitrogen in coal tar is present mainly in the form of
pyrrolic and pyridinic nitrogen. Fourier transform infrared (FTIR) spectroscopy has been
recognized as a useful technique for providing information about these functional groups
with heteroatoms in heavy hydrocarbons (Surygla and Sliwka, 1994; Granda et al., 1990;
Kaihara et al., 1991; Diez et al., 1994; Guillen et al., 1995ab). If a substance can be dis-
solved in a solvent such as CS2, good results can be obtained by estimating the absorbance
for the specific band and by comparing it with that of a standard compound. However, coal
3.2 Pyrolysisof Coal Tar 151
tar has poor solubility in solvent because it contains high molecular weight compounds.
Alternatively, a tritium tracer technique can be used to quantify accurately the amount
of hydrogen in the functional groups of substances that have a macromolecular structure
such as coal and coal tar (Ishihara et al., 1993, 2000, 2001, Kabe et al., 1990b, 199 lb, Qian
et al., 1997). This technique was also reported to be effective in determining the mobility
of h y d r o g e n in coal and coal-related c o m p o u n d s u n d e r coal l i q u e f a c t i o n conditions
(Ishihara et al., 1993a, b, c, 1994, 1995, 1996, 1999, 2002a, b; Kabe et al., 1991, 1993,
1995, 1997).
In this section, the amount of hydrogen in functional groups of coal tar and the hydro-
gen mobility in its macromolecular structure are discussed. The examination was made by
hydrogen exchange reaction of coal tar or its representative model compounds with tritium-
labeled water (Kabe et al., 1998). Coal tar used in the study was typical high temperature
coal tar produced from blast furnace coke, and its properties are shown in Table 3.7.
12t
1.o lOO ~
C) --
0.8
~ 0.6
0.4
0.2
0.0
0 5 10 15 20 25 30
Reaction time (h)
Fig. 3.19 Effectof reaction time on the ratio of hydrogen exchange.
[From Kabe. T. et al., J. Jpn. Petrol. Inst. 41, 165 (1998)]
152 3 Pyrolysis
tio of exchangeable hydrogen in the coal tar to the total amount of hydrogen in the original
coal tar as presented in the analytical data of Table 3.7. The data of Fig. 3.19 show that the
HERs of the coal tar gradually increased and approached a constant value of about 0.9% at
100 ~ after 6 h. It is assumed that hydrogen in the functional groups of the coal tar such as
hydroxy and imino groups was rapidly exchanged through the proton exchange between
water and the coal tar at 100 ~ Based on the amount of hydroxy and imino groups ob-
tained from the FTIR analysis (shown in Table 3.7), the calculated ratio of hydrogen in
these groups to total hydrogen in the coal tar is 0.65%. The constant value of the HER of
coal tar at 100 ~ is slightly higher than the predicted value above. This discrepancy is con-
sidered to be due to the presence of a small amount of other functional groups such as thiol,
amino group and carboxylic acid. The hydrogen in these functional groups would also be
exchanged through the proton exchange. Therefore, the HER of coal tar at lower tempera-
tures may represent the amount of hydrogen in all functional groups present in the coal tar.
The result also shows that much longer time is needed for it to react up to the equilibrium
state as the exchange reaction is run at a much lower temperature (50 ~
Figure 3.20 shows the effect of temperature on the HER of coal tar for 3 and 6 h.
HERs of coal tar increased remarkably over 100 ~ This indicates that hydrogen in coal
tar other than that in the functional groups become exchangeable with hydrogen in water
30
20 6h
&.
10 3h
0
0 100 200 300 400
Temperature (~
Table 3.8 Hydrogen Exchange Ratio of Model Compounds with Tritiated Water (%)
Experimental data b
Compounds Predicted HER a
100 ~ 300 ~
Naphthalene 0.0 0.04 0.39
Naphthol 12.5 13.1 38.6
Indole 14.3 15.0 49.7
a Ratio of hydrogen in functional group to total hydrogen in each compound.
b HER obtained from exchange reaction with tritiated water for 6 h.
[From Kabe. T. et al., J. Jpn. Petrol. Inst., 41, 166 (1998)]
3.3 Pyrolysis of Coal Tar Pitch 153
over 100 ~
well known that the properties of carbon and graphite are controlled to the greatest extent
by the nature of the precursor pitch and the mechanism of its transformation into coke
(Mochida, 1991). The coking of pitch occurs in the liquid state and is the most amenable
stage of carbonization for study. The subsequent conversion of coke to carbon is basically
a solid-state process and is much more difficult to consider (Lewis, 1980). The chemistry
involved in the transformation of pitch to coke and carbon is exceedingly complex. Much
insight into the mechanisms of carbonization has come through studies on pitches derived
from single aromatic compounds. Lewis (1980) described general mechanisms for the car-
bonization of naphthalene, dimethylnaphthalene and dibenzothiophene. Fitzer et al. (1973)
gave a detailed mechanistic scheme for the pyrolysis of acenaphthylene. In addition, it is
worthy of note that the carbonization pathway of a compound in the presence of a catalyst
is completely different from that in the absence of a catalyst (Mochida, 1990; Mochida et
al., 1975). These studies on carbonization of model compounds help in predicting the type
of thermal phenomena involved in pitch pyrolysis. However, because pitches are very
complex mixtures of aromatic compounds with different functionalities and molecular
structures and a broad molecular weight distribution, a better understanding of the car-
bonization of pitch has not yet been achieved.
In this section, the hydrogen mobility or reactivities of various pitches are first estimat-
ed to clarify the relationship among hydrogen mobility of pitch, molecular structural fea-
tures, and optical texture developed at the early stage of carbonization. Then the hydrogen
behavior during carbonization of pitch is addressed.
3.3.1 B e h a v i o r o f H y d r o g e n d u r i n g H y d r o g e n a t i o n o f C o a l Tar P i t c h
A. Structural Features and Hydrogen Mobility of Coal Tar Pitch
In recent years, a number of attempts have been made to produce pitch-based carbon fibers.
These works are often fundamental studies concerning the manufacture of high perfor-
mance carbon fiber from coal tar pitch, ethylene tar pitch, FCC decant oil, etc., since they
are quite cheap and can be obtained in quantity. Among them, carbon fibers derived from
coal tar pitch are usually classified as a high performance carbon fiber (HPCF) or a general
performance carbon fiber (GPCF), depending on the properties of the starting pitch and
preparative method of the precursor pitch. In general, the coal tar pitch derived from coke
oven is initially heat-treated at 350-450 ~ under reduced pressure, then divided into two
portions. The portion containing mesophase pitch is used as the raw material for HPCF,
and the remaining portion as the raw material for GPCF.
It was known that hydrogen in pitch molecules plays an important role in many
processes for the manufacture of carbon fibers (Mochida and Marsh, 1980; Obara et al.,
1980; Yamada et al., 1984; Yokono et al., 1981), such as heat treatment (Kabe et al., 1990),
oxidative stabilization (Wang et al., 1991) and carbonization (Shono et al., 1991). So far,
hydrogenation by hydrogen molecules and Benkeser reduction have been investigated in or-
der to introduce hydrogen into pitch (Mochida et al., 1981, 1982, 1987). However, the re-
activity of hydrogen in pitch in hydrogenation has been investigated.
Three kinds of pitches were hydrogenated. The differences in chemical and physical
properties of the pitches were estimated by tracing the reactivities of tritium-labeled hydro-
gen with pitches. Attention was focused mainly on the interdependence of hydrogen mobil-
ity on the structural features.
The hydrogenation of a coal tar pitch for HPCF (HP-A) and two coal tar pitches for
GPCF by tritiated gaseous hydrogen (GP-A, GP-B) was carried out. The experimental tri-
tium distributions for three pitches after the hydrogenation are shown in Fig. 3.21. The
156 3 Pyrolysis
amounts of tritium in gas phase linearly decreased with the passage of time at each temper-
ature for all three pitches. The tritium was transferred from gas phase to pitch via hydrogen
100
g
=9 80
N 60
O 300 ~
[] 350 ~
A 400 ~
, I , I , I , I , I
40 ,
0 60 120 180 240 300
Nominal reaction time (min)
Fig. 3.21 Change in tritium concentration in gas phase with reaction time (HP-A Pitch). [Reproduced with
permission from Wang, X. et al., Fuel Process. Technol., 38, 73, Elsevier (1994)]
0.4
O Addition (a)
9 Exchange 300 ~
0.3
0.2
0.1
0.0~
.~ 0.4
d=
[] Addition (b)
9 Exchange 350 ~
~ o.3
0
r~
~ 0.2
~ O.1
0.0 ~_____~---r---7-~- - - - - - - - - - ~
, I , I ~ I
.~ 0.4
A Addition (c)
=
9 Exchange 400 ~
0.3
0.2
0.1
0.0 , , i i |
Fig. 3.22 Amount of hydrogen transferred from gas phase to coal tar pitch (HP-A pitch). (Wang, X. 1993)
3.3 Pyrolysis of Coal Tar Pitch 157
addition and exchange. The amounts of hydrogen addition and hydrogen exchange were
calculated from the experimental tritium distributions, and the results for the three pitches
are shown in Fig. 3.22. It was found that the amounts of hydrogen addition and exchange
also increased linearly with reaction time at each temperature, although, at 400 ~ plots de-
viated slightly from the straight line. This may show that side reactions such as dehydro-
genation and polycondensation occurred significantly at 400 ~ It was assumed that tri-
tium was transferred from gas phase to pitch by a first order reaction depending on the con-
centration of tritium in gas phase. Thus, the rate constants of hydrogen addition and ex-
change reactions at each temperature were obtained as shown in Table 3.9. Further, the ap-
parent activation energies of h y d r o g e n addition and e x c h a n g e were calculated f r o m
Arrhenius plots of the rate constants, and the results are summarized in Table 3.10. As giv-
en in Table 3.9, the rate constant of hydrogen addition for HP-A pitch is larger than those
for GP-A or GP-B pitch, and the rate constant of hydrogen exchange for HP-A pitch is
smaller than these for GP-A or GP-B pitch at each temperature. The rate constants of hy-
drogen addition and hydrogen exchange for GP-A and GP-B are very close to each other at
every temperature. The apparent activation energies for the hydrogen addition of three
pitches are approximately the same (ca. 9 - 1 0 kcal/mol), but those for the hydrogen ex-
change of three pitches are different from each other (HP-A, 15.4; GP-A, 2.9; GP-B, 3.7
kcal/mol). In particular, the apparent activation energy of hydrogen exchange of the raw
pitch for HPCF was significantly larger than those of the raw pitches for GPCF. The kinet-
ic parameters for the hydrogen transfer may reflect the differences in the structures and re-
activities of component molecules in pitches. Since hydrogen addition from gas phase to
pitch in the absence of both vehicle solvent and catalyst usually proceeds by a radical
158 3 Pyrolysis
mechanism, hydrogen molecules are added into pitch to stabilize the radicals produced by
thermal cracking (Sasaki and Sanada, 1991). The rate of hydrogen addition could depend
mainly on the thermal reactivities of pitches but was hardly related to the hydrogen mobili-
ty or hydrogen distribution. This proposition was supported by the fact that the apparent
activation energies of hydrogen addition from gas phase to three pitches were almost the
same. However, the hydrogen exchange between gas phase and pitch was associated main-
ly with the bond energy of C-H, O-H, N-H etc. in pitch molecules. Consequently, informa-
tion concerning the distribution and the mobility of the hydrogen in pitch could be obtained
by estimating the kinetic parameters of hydrogen exchange.
The mechanism of the hydrogen exchange between gaseous hydrogen and coal tar
pitch is not well known because of the complicated composition of coal tar pitch. As evi-
dence of structural difference, it has been shown that the GP-A pitch polycondensate more
easily than the HP-A pitch to form a semi-coke. The more the thermal decomposition of
molecules takes place easily, the more the exchange reaction becomes rapid. In addition, it
was reported that the peri-condensated aromatics, such as pyrene, were inert toward ex-
change (Derbyshire and Whitehurst, 1981). Such peri-condensated aromatics would be pre-
sent in the pitch for HPCF in quantity since they are major species constituting mesophase
(Mochida and Korai, 1982). To discover the differences in the kind and the amount of ex-
changeable components in the pitches, detailed composition analyses are required.
However, the kinetic parameters obtained from the hydrogenation may be used to compre-
hensively evaluate the structural feature of pitch. There are some differences in the rate
constants and the apparent activation energies between GP-A and GP-B pitches. It is con-
sidered that the basic aromatic structure does not change at the initial heat treatment.
In a recent work on carbonization of tritiated coal tar pitch (Wang et al., 1991), the be-
havior of hydrogen in coal tar pitch during carbonization showed that the hydrogen in pitch
which is easily exchanged with H20 or H2 will be released at the initial stage of carboniza-
tion, and that such exchangeable hydrogen is strongly related to the polycondensation of
pitch molecules. The hydrogen in raw pitch for HPCF can be reserved in pitch molecules
100
O'-"
80
4 e,l
E
~9 60
3 .~
tD 2 =
r~
O
40 0 Methane
[] Ethane 1 .~
A Ethylene
9 Propane 0 ~
~ 20 [] Partial pressure
Fig. 3.23 Changes in composition and partial pressure of noncondensable hydrocarbon gas with reaction time
(GP-A pitch, 350 ~ 5.9 MPa initial H2 pressure). [Reproduced with permission from Wang, X. et
al., Fuel Process. Technol., 38, 77, Elsevier (1994)]
3.3 Pyrolysisof Coal Tar Pitch 159
till higher temperature, so that the fluidity of the pitch required for development of optical
features can be maintained to higher carbonization temperatures. The fact that the activa-
tion energy of hydrogen exchange of the pitch for H P C F was significantly higher than that
of raw pitch for G P C F reflects the differences in the aromatic structures of c o m p o n e n t mol-
ecules between the two kinds of coal tar pitches. The hydrogen in the pitch for H P C F is
difficult to exchange with gaseous hydrogen in comparison with that in the pitch for GPCF.
The changes in pitch composition after h y d r o g e n a t i o n were investigated. Fig. 3.23
shows the gas phase composition and partial pressure of non-condensable hydrocarbon gas
after the reaction of GP-A pitch at 350 ~ The relative content of methane and the partial
pressure of hydrocarbons increased remarkably with prolonged time, and saturated at 120
min. Methane was a major product in gas phase, and hydrocarbons other than methane
were found in extremely low contents and scarcely changed with prolonged time. These re-
sults suggest that, in the hydrogenation of coal tar pitch in the absence of catalyst and vehi-
cle solvent, the hydrocracking of substitutional chains in aromatic tings would proceed at
the initial stage of the reaction, and that the alkyl chains longer than methyl were hydroc-
racked only at the initial stage. Table 3.11 shows the results of solvent extraction of the
raw pitches and the pitches hydrogenated under various conditions. THFIS fraction of HP-
A pitch significantly increased with temperature and slightly increased with reaction time
except for 400 ~ 300 min. THFIS fraction of GP-A pitch increased significantly with
both temperature and reaction time. This indicates that polycondensation of pitches occurs
160 3 Pyrolysis
tional groups to the experimental data through a set of balance equations. The set of all bal-
ance equations can be expressed in the matrix form
where the Yj (j = 1..... n) represent the unknown functional group concentrations, bi (i - 1 .....
m) are quantities representing the elemental and NMR data, and the Aij represent stoichio-
metric coefficients. If the functional groups of Fig. 3.24 are combined with the data of ele-
mental analysis and hydrogen distribution, the stoichiometric coefficients and data vector b
result. In addition to the constraints imposed by the data, the concentration must satisfy the
following equation:
Yj_>0 (3.6)
For the mathematical problem presented by Eqs. (3.5) and (3.6), two general classes of so-
lution are possible. In the first class, the number of balance Eq. (3.5) is greater than the
number of unknown functional group concentrations. In this case, the concentrations are
determined using the weighted least square method. In the second class of solutions, the
number of equations is smaller than the number of unknowns. In this case, Eq. (3.5) can
yield either no solution or an infinite number of solutions. If the equations have no solu-
tion, then the proposed set of functional groups is insufficient to describe the observed data
1 0
9 (3------ CH3
10 ~ CH3
4
11 D CH3
12
13 O------ CH2---C)
6 14 C) OH
Fig. 3.24 Probable functional groups present in coal tar pitch. [Reproduced with permission
[From Takeuchi, M. et al., J. Jpn. Petrol. Inst., 37, 139 (1994)]
162 3 Pyrolysis
and must be revised. If a space of solution exists, then the mixture can be characterized by
selecting a single solution from the feasible space. Selecting a single solution provides a
valid structural characterization because the range of feasible group concentrations is limit-
ed, even through the number of solutions is mathematically defined as infinite. The equa-
tions dealt with in the present work belong to this case.
To select a single solution from the feasible region, we can use the following computa-
tional procedure. The concentrations I11. . . . . Yn are chosen such that a function P (Y1. . . . . Yn)
is minimized, subject to the constraints of Eqs. (3.5) and (3.6). The form of the function P
can vary depending on which, if any, data are available in addition to elemental analysis
and NMR. Data other than elemental analysis and NMR can be introduced as additional
balance equations or can be incorporated into the function P. Here, we adopt the minimiza-
tion objective function Eq. (3.7):
P = [(fa-faexp) 2 + (O'-O'exp)2 + (p-pexp)2] 1/2 (3.7)
where fa, cr and/9 are the structural parameters determined by Brown-Ladner's method, and
faexp, Crexpand p~xp are the structural parameters determined from the initial solutions of Eqs.
(3.5) and (3.6). The precise concentrations of functional groups can be implemented by
minimizing Eq. (3.7) subject to Eqs. (3.5) and (3.6).
A coal tar pitch was hydrogenated at 410 ~ to 470 ~ for 60 min by tetralin or tetrahy-
droquinoline (THQ). The hydrogenation procedure has been described in detail elsewhere
Table 3.12 Hydrogenation Conditions of Pitches Used to Structure Analysis (Wang et al. 1993)
1 R-BI m __
60 g of pitch and 120 g of solvent were charged into a 500-mL stainless autoclave with an
inner tube equipped with a stirrer assembly, then replaced by nitrogen before it was subjected
to heat.
Table 3.13 Functional Group Concentrationsin Each Fraction of HP-B Pitch Hydrogenated by Tetralin (Wang et al. 1993)
R-BI 0.0 0.0 0.0 0.0 0.0 0.0 17.8 19.9 34.8 20.7 Q.0 0.0 6.8
H-T41O-BI 0.0 0.0 0.0 0.0 0.0 0.0 50.1 0.1 3.3 17.3 22.9 0.0 6.3
H-T430-BI 0.0 0.0 0.0 0.0 0.0 0.0 61.1 0.1 7.1 18.2 13.5 0.0 0.0
H-T450-BI 0.0 0.0 0.0 0.0 0.0 0.1 59.9 12.8 0.0 7.5 19.0 0.0 0.0
H-T470-BI 0.0 0.0 0.0 0.3 0.8 0.0 52.9 35.1 0.0 4.5 6.7 0.0 0.0
R-BS 0.0 0.0 9.0 7.1 12.3 19.5 0.0 24.2 12.1 7.5 2.4 5.9 0.0
H-T4 10-BS 0.0 0.0 0.0 0.0 9.6 11.3 0.0 24.5 32.7 16.1 0.0 5.8 0.0
H-T430-BS 0.0 0.0 0.0 1.5 13.2 13.9 0.0 28.2 21.4 7.1 0.0 14.7 0.0
H-T450-BS 0.0 0.0 0.0 0.0 14.0 14.4 0.0 30.1 22.7 5.3 0.0 13.5 0.0
H-T470-BS 0.0 0.0 11.4 17.9 2.2 20.1 0.0 16.3 7.6 5.5 12.6 6.4 0.0
R-HS 0.0 18.9 19.4 9.8 5.3 0.0 0.0 24.0 11.4 0.0 0.0 9.6 1.6
H-T410-HS 0.3 17.0 22.0 13.9 0.0 0.0 0.0 26.2 0.0 0.0 7.1 11.3 2.2
H-T430-HS 0.0 11.4 13.6 9.8 6.4 0.0 0.0 33.8 2.6 1.0 6.1 11.3 4.0
H-T450-HS 0.0 20.7 4.7 2.8 12.3 0.0 0.0 30.7 14.8 2.2 0.0 9.8 2.0
H-T470-HS 1.8 14.4 15.7 13.6 0.0 0.0 0.0 30.1 10.5 1.9 2.0 9.0 1 .o
R-BI 0.0 0.0 0.0 0.0 0.0 0.0 17.8 19.9 34.8 20.7 0.0 0.0 6.8
H-Q4 10-B I 0.0 0.0 0.0 0.0 0.0 0.0 31.8 29.8 12.8 2.7 0.0 14.7 8.2
H-Q430-BI 0.0 0.0 0.0 0.0 0.0 0.0 29.4 31.1 17.3 4.7 0.0 10.5 7.0
H-Q450-BI 0.0 0.0 0.0 0.0 0.0 0.0 37.0 26.8 10.7 0.0 0.0 16.6 8.9
R-BS 0.0 0.0 9.0 7.1 12.3 19.5 0.0 24.2 12.1 7.5 2.4 5.9 0.0
H-Q410-BS 0.0 0.0 0.0 0.0 0.0 0.0 8.4 22.7 41.7 21.8 0.0 3.2 2.2
H-Q430-BS 0.0 0.0 0.0 0.0 0.0 0.0 10.7 24.2 40.8 20.2 0.0 3.1 1.0
H-Q450-BS 0.0 0.0 0.0 0.0 0.0 0.0 13.2 23.6 39.9 19.8 0.0 2.3 1.2
R-HS 0.0 18.9 19.4 9.8 5.3 0.0 0.0 24.0 11.4 0.0 0.0 9.6 1.6
H-Q410-HS 0.0 3.8 18.4 9.2 5.8 0.0 0.0 31.7 16.9 1.8 0.0 12.4 0.0
H-Q430-HS 0.3 18.7 6.5 21.0 0.0 0.0 0.0 27.5 0.0 2.2 11.9 11.4 0.5
H-Q450-HS 2.4 15.0 22.8 17.7 0.0 0.0 0.0 15.4 0.0 3.8 15.9 6.3 0.7
0- Bond directly to an aromatic ring.
+Bond to a carbon alpha to an aromatic ring.
Bond a carbon beta or further from an aromatic ring.
3.3 Pyrolysisof Coal Tar Pitch 165
(Shono et al., 1990). The raw pitches and hydrogenated pitches were separated into three
fractions (BI, BS and HS) by solvent extraction. The fractions obtained were used as sam-
ples for 1H-NMR and elemental analysis. Functional group concentrations were estimated
by using the method described above for all samples listed in Table 3.12, and are given in
Tables 3.13, 3.14. The 13 concentrations, which have been given in Tables 3.13, 3.14, pro-
vide detailed structural profiles of the coal tar pitches unhydrogenated and hydrogenated by
donor solvents and can provide a reasonable starting point for modeling the physico-chemi-
cal characteristics of these samples. The results reported in this work show that the struc-
tural analysis based on functional group concentrations is effective for characterizing pitch.
The results of functional group analyses are in good agreement with those of the average
structural parameters calculated according to the Brown-Ladner method, indicating that
functional group distributions provide at least as much as information as average parame-
ters.
Examination of the functional group concentration reveals that the effect of hydrogena-
tion by tetralin was different from that by THQ. The pyrolysis or hydropyrolysis of aliphat-
ic substitution chains are easy to induce with the B IS fraction of pitch under hydrogenation
conditions, regardless of solvent, but stabilization of radicals produced by pyrolysis is
strongly related to hydrogen donor capacity of solvent. It seems that tetralin has higher
donor capacity than THQ, since the hydrogenation by tetralin significantly increased the
concentration of eight aromatic tings and decreased the concentration of substituted aliphat-
ic chains. In the case of hydrogenation by THQ, the concentration of eight aromatic tings
and aliphatic substitution chains also increased and decreased, respectively, but the extent
of the increase or decrease was much smaller than that in the case of hydrogenation by
tetralin. In addition, a considerable number of eight aromatic tings in BS were formed in
the case of hydrogenation by THQ. This is considered to be from the recombination of rad-
icals produced by pyrolysis due to the poorer donor capacity of THQ. For the BS or HS
fraction of pitch, however, significant change in concentration distribution of functional
groups in the case of hydrogenation by THQ was observed. The concentration of seven
aromatic rings in the BS fraction of the pitch decreased from 19.5% to 0% and the concen-
trations of aliphatic substitution chains, especially longer aliphatic chains, increased
markedly. The hydrogenation by THQ significantly changed the concentration distribution
of functional groups, but this does not indicate that THQ is a good donor solvent because
the concentration of naphthenic rings did not increase. Summarizing these results, it is sug-
gested that tetralin is a more effective solvent for hydrotreatment of coal tar pitch, especial-
ly for heavier fractions of pitch, than THQ.
Cronauner et al. (1978) studied the hydrogen transfer cracking of dibenzyl in tetralin
and related donor solvents; and it was concluded that the reaction rate of hydrogen transfer
cracking is independent of the donor solvent, but the product distribution is dependent upon
the type of solvent. Our results for the hydrogenation of coal tar pitch by donor hydrogen
solvent agree well with this conclusion.
In the hydrogenation of coal tar pitch, the reaction can be considered to proceed in ac-
cordance with the same or similar mechanism as that in coal liquefaction. Only the extent
of various reactions in hydrogenation may differ depending upon the structural features of
the feedstock. In the hydrogenation of coal tar pitch, the cleavage of substituted aliphatic
chains may be the dominant reaction by which the aromaticity of constituent molecules in-
creases. In comparison with coal, coal tar pitch consists of more condensed aromatic tings
with some aliphatic substitution chains. In addition, the bridge structure in coal tar pitch is
much less than that in coal since the former underwent severe thermal hysteresis. So the
166 3 Pyrolysis
basic structures will not change so much as in the conversion of coal. The hydrogenation
by donor hydrogen solvent mainly improves the hydrogen distribution of coal tar pitch. In
addition, partial saturation of condensed aromatic rings also can occur if a solvent having
high capacity for hydrogen donors is used. This reaction is the most expected since the
condensed aromatic rings with naphthenic groups are favorable components for the devel-
opment of anisotropic texture during heat treatment.
Bond scission of strongly bonded pitch structure occurs when the hydrogenation tem-
perature is sufficiently high. At above 470 ~ the concentration of naphthenic rings which
formed at lower temperatures and eight aromatic rings in BIS fraction of the pitch de-
creased as shown in Table 3.13, indicating that cleavage of strongly bonded aliphatic or
aromatic rings occurred at 470 ~ This temperature for cleavage of rings is higher than
that usually observed in coal liquefaction, suggesting that pyrolysis of coal tar pitch is more
difficult than pyrolysis of coal. The scission of strongly bonded pitch structures is consid-
ered to be related less to the action of the donor solvent. This has been illustrated by
Vlieger et al. (1984). In the temperature, pressure and contact time ranges commonly en-
countered in thermal coal liquefaction processes using donor solvents a good donor solvent
will not directly break carbon-carbon bonds of the type encountered in dibenzyl. This bond
must first be thermally broken to form free radicals. Once free radicals are formed, they
will readily react with any donor solvent. The nature of the product distribution will, of
course, depend t~pon the nature of the donor solvent.
3.3.2 H y d r o g e n Behavior during Carbonization of Pitch and M e c h a n i s m of
Carbonization
A. Effect of Hydrogenation on Pyrolysis Reactivity of Coal Tar Pitch
It is known that solubility and fluidity of pitch can be modified by hydrotreatment so as to
facilitate production of the mesophase required for HPCF in the successive carbonization
process. Several authors have reported on the influence of hydrogenation under various
conditions in the development of optical anisotropy of coal tar pitch (Mochida et al., 198 lb;
Yamada et al., 1987) or petroleum pitch (Mochida et al., 1981b)during successive car-
bonization. It is clear that, when the hydrogenated pitch is carbonized, the optical texture in
the resultant coke is modified. In these two mutually contrasting processes of hydrotreat-
ment and heat treatment, however, what roles the hydrogen in the pitch would play have not
yet been clarified. Here, a coal tar pitch was hydrotreated using tritium-labeled tetralin as
donor solvent without catalyst prior to heat treatment, after which the tritiated pitch was
carbonized at 30-1000~ In addition, the features of heat treatment of each component of
both pitch and hydrotreated pitch were compared by non-isothermal TGA.
The coal tar pitch was hydrotreated by tritium-labeled tetralin to obtain the tritiated
pitch. The mass and hydrogen balances, after hydrotreatment of the pitch, are shown in
Fig. 3.25. The products and tritium distributions are summarized in Fig. 3.25a. The HIS-
BS fraction was 63.7 wt%, higher than that of the raw pitch. The HS and THFIS fractions
of the hydrotreated pitch were 11.2 and 12.1 wt%, respectively, which were lower than that
of raw pitch. Tritium was introduced from tetralin into every fraction of the products. As
shown in Fig. 3.25b, hydrogenation of pitch proceeded along with hydrogen addition and
exchange reactions. The amount of hydrogen added from solvent to pitch was 0.31 g. The
amount of hydrogen exchanged between pitch and solvent was 0.13 g, which was less than
the amount of hydrogen added. The hydrotreatment of pitch by donor solvent is considered
to proceed by a radical mechanism, as in the hydrotreatment of petroleum heavy oil (Kabe
et al., 1983b) or liquefaction of coal (Kabe et al., 1989). The lighter fractions such as gas,
3.3 Pyrolysis of Coal Tar Pitch 167
Gas 0.7%
-633500 dpm/g
G
Naphtha 3.1%
186500 dpm/g HS 11.2%
0.07 0.13\
52000 dpm/g N.
Product - - _Light oil 7.5%
65900 dpm/g HIS-BS 63.7%
40800 dpm/g
Hydrotreaed LO
-pitch 88.6% m
BIS-THFS 12.9%
41600 dpm/g 37800 dpm/g
G: Gas; N: Naphtha; LO: Light oil;
S: Solvent; HP: Hydrotreated pitch.
THFIS 12.1% (Hydrogen (g)/Pitch (30 g))
31600 dpm/g Hydrogen addition:
Hydrogen exchange:
(a) (b)
Fig. 3.25 Mass and hydrogen balances after hydrotreatment of pitch.
[From Wang. X. et al., J. Jpn. Petrol. Inst., 34, 316 (1991)]
naphtha and light oil were formed by addition of hydrogen into radicals, produced through
pyrolysis of branched chain in aromatics and oligomeric polyarylene-type compounds, with
methylene groups as linkages. Scission of C-C bonds, in a short alkyl substituent on aro-
matic rings, was the first stage of the reaction (Kabe et al., 1989) and the polymerization of
resultant aromatic radicals caused increase in the HIS-BS fraction. A considerable amount
of tritium was contained in the heavier fractions such as BIS-THFS and THFIS of pitch, in-
dicating that substantial hydrogen exchange reactions occurred in these fractions.
The pyrolysis of the tritiated pitch was conducted at 30-1000 ~ and the results are
shown in Fig. 3.26. The amounts of tritium and hydrogen in pitch decreased with a rise
from 400 to 1000 ~ The rate of decrease of hydrogen content in pitch was greater than
that of tritium in the range of 400 to 600 ~ The rates of decrease of tritium and hydrogen
in pitch coincided at over 600 ~ when the weight loss of pitch could scarcely be seen.
The difference in the concentrations of tritium and hydrogen in pitch occurred in the range
of 400 to 600 ~ indicating that the hydrogen, not having been tritiated during preparation
of tritiated pitch, was released in this temperature range. This suggests that considerable
amounts of saturated hydrocarbon and branched chains in aromatics are present in the raw
pitch, and these are released by pyrolysis at 400-600 ~ Further, it is assumed that only
the tritium and hydrogen in aromatic rings were released and that the carbonization of pitch
was proceeding in this temperature range, since the weight loss could not be seen over
600 ~ and the rates of decrease of hydrogen and tritium were nearly the same.
The features of the heat treatment of raw and hydrotreated pitches were compared by
non-isothermal TGA, as shown in Fig. 3.27. TG curves of both pitches are similar, but
weight loss of the hydrotreated pitch in final stage was greater than that of the unhydrotreat-
ed pitch. The TGA analyses of each component of the raw and hydrotreated pitch are
shown in Figs. 3.28a-d. TG curves of HS and THFIS fractions of both pitches were practi-
cally the same, but TG curves of HIS-BS and BIS-THFS fractions of both pitches displayed
curves different to each other. Differences in the TG curves are most probably due to the
structures of the component molecules that are present between raw and hydrotreated pitch.
The results suggest that, in the hydrotreatment of pitch, a considerable change in the molec-
168 3 Pyrolysis
100
80
100 60 .,~ ~-
,.t=
~9 60
. ,...~
4O 03 m
20
I I I I I I I I I
Fig. 3.26 Weight loss and change in tritium and total hydrogen concentrations with temperature in pyrolysis of
tritiated pitch. (The initial total hydrogen content in pitch was 5.15 wt%, and the initial tritium content
was 45214 dpm/g. The experimental points were plotted on the basis of these initial values.)
[From Wang. X. et al., J. Jpn. Petrol. Inst., 34, 316 (1991)]
I00
80
.~
exo 60
"~ 40
20
i i i i
0 !
0 200 400 600 800 1000
Temperature (~
Fig. 3.27 Non-isothermal TG curves for raw and hydrotreated pitches at a heating rate of 12 ~
[From Wang. X. et al., J. Jpn. Petrol. Inst., 34, 317 (1991)]
ular structure of pitch takes place mainly in HIS-BS and BIS-THFS fractions of pitch, si-
multaneously. Further, the small weight loss of the THFIS fraction occuring in the range of
500 to 850 ~ (Fig. 3.28d) suggests that the THFIS fraction is extremely heat stable, and
3.3 Pyrolysisof Coal Tar Pitch 169
~, 80
.~ 60 .~ 60
~ 4o ~ 4O
20 20 . . . . . .
0 0
0 200 400 600 ' 800 ' 1000 0 ' 2;0 ' 4()0 ' 6;0 ' 8;0 1000
Temperature (~ Temperature (~
(a) (b)
~, 80 80
~9 60 60
"~ 40 40
nr
20 20
00 ' 2()0 ' 4()0 ' 6()0 ' 8()0 ' 1000 0 ' 2()0 ' 400 ' 660 ' 8()0 ' 1000
Temperature (~ Temperature (~
(c) (d)
0 Raw pitch 9 Hydrotreated pitch
Fig. 3.28 Non-isothermal TG curves for each component of the raw and hydrotreated pitch
at a heating rate of 12 ~
[From Wang. X. et al., J. Jpn. Petrol. Inst. 34, 317 (1991)]
is considered that the increase of pyrolysis activation energy of HS fraction is due to the in-
crease in thermally stable components that come from the products of hydrocracking of the
heavier fraction in pitch, and the increase in pyrolysis activation energy of THFIS fraction
is due to the decrease in decomposable components and that the hydrotreatment did not sat-
urate its aromatic rings.
100 I ~ ~
~- 80
~40
.~ 9 Weight
20 I-l Hydrogen
A Tritium
I i
00 200 400 600 800 1000
Temperature (~
Fig. 3.29 Changes in residual weight, hydrogen and tritium concentrations in hydrogenated HP-B pitch
with temperature. [Reproduced with permission by Wang. X. et al., Fuel Process. Technol.,
38, 51, Elesvier (1994)]
tritiated by isotope exchange. The rate of dehydrogenation was faster than that of NP triti-
ated by isotope exchange, but slower than that of detritiation itself, although the difference
between the two rates was smaller than that in the case of hydrogenated HP-B. This result
indicates that hydrogenation is also effective for improving thermal reactivity of naphtha-
lene pitch, although the structural features of naphthalene pitch may be significantly differ-
ent from those of coal tar pitch.
The start and end of the carbonization process are indefinable. To facilitate compre-
hension of the subject, the following postulates for the start and end of the carbonization
process are used: Carbonization begins when chemical reactions such as thermal degrada-
tion or molecular condensation occur. The temperature range for the commencement of
carbonization depends on the reactivity of the feedstock (Tillmans, 1979). Carbonization
finishes when the starting material is completely solidified and low molecular weight prod-
ucts are removed completely by thermal degradation.
100 [7~,
o ~ 80
.O.O
~ .,.
o ".~ 60
O
~ 40
= ~ 9 Weight
20 D Hydrogen
A Tritium
0 , I , I , I , I ,
0 200 400 600 800 1000
Temperature (~
Fig. 3.30 Changes in residual weight, hydrogen and tritium concentrations in hydrogenated NP
with temperature. [Reproducedwith permission by Wang. X. et al., Fuel Process.
Technol., 38, 52, Elesvier (1994)]
3.3 Pyrolysisof Coal Tar Pitch 173
It is well known that the crystalline and textural structures of carbon materials depend
on the structural characteristics of their feedstocks (Tillmans, 1979). The present experi-
mental results indicate that the release of hydrogen in parent pitch during carbonization is
related to the optical features of pitch developed at an early stage of carbonization. It was
observed that (Wang et al., 1994b), in hydrogenation of coal tar pitches for different perfor-
mance carbon fiber using gaseous hydrogen in the absence of catalyst or solvent, the pitch
having the higher activation energy of hydrogen exchange will show better flow texture at
the early stage of carbonization. Although a correlation between the abundance of ex-
changeable hydrogen and the optical texture of pitch developed at the early stage of car-
bonization could not be observed with a series of pitches with different properties, it was
found that the rate of dehydrogenation or detritiation during carbonization of pitch at
400-700 ~ corresponded well to the development of flow texture. That is, the better the
development of flow texture, the slower the rate of dehydrogenation or detritiation at
400-700 ~ A similar phenomenon has been observed (Wang et al., 1994b). When pitch-
es were heat-treated under hydrogen atmosphere at 300-400 ~ the amount of THFIS frac-
tion formed from the pitch for general performance carbon fiber (GPCF) was greater than
that formed from the pitch for high performance carbon fiber (HPCF), indicating that the
thermal reactivity of pitch for GPCF was higher than that of pitch for HPCF. The THFIS
fraction has a much lower H/C ratio than the fractions lighter than THFIS, indicating dehy-
drogenation significantly occurred accompanying the polycondensation of pitch. These
facts allow us to believe that too rapid dehydrogenation and polycondensation especially at
lower temperatures may result in isotropic carbon because dehydrogenation will increase
viscosity of the system and carbonization under high viscosity state does not benefit the de-
velopment of anisotropic texture. Mesophase is generally formed at a temperature range of
350-450 ~ where pitch is in the liquid state. The orientation of condensated aromatic mol-
ecules appearing at lower temperatures can remain up to higher carbonization temperatures,
so that a parent pitch for HPCF should be required to maintain low thermal reactivity at
lower temperature except under those conditions suggested by Mochida and Korai (1985).
The hydrogen in parent pitch can be divided into two types in terms of the exchange-
ability of hydrogen. The first type is the hydrogen which is easy to exchange with isotope
and the second type is the hydrogen which is difficult to exchange with isotope. The ratio
of the first type to the second type seems to depend upon many factors such as the structural
features of the component molecules, size of molecules, etc. Tritium represents the ex-
changeable hydrogen in pitch, and is released rapidly during the carbonization relative to
hydrogen, which is difficult to exchange isotopically. This implies that the exchangeable
hydrogen in pitch actively participates in the thermal reaction. However, the thermal be-
havior of hydrogenated pitches was quite different from that of pitches tritiated by isotope
exchange. The original hydrogen in hydrotreated pitch was released easily. This phenome-
non was also observed with Eureka pitch hydrogenated with Ni-Mo/A1203 (Mochida et al.,
1987). It is generally assumed that hydrogenation improves thermal decomposition charac-
teristics of pitch, hence the carbonization of hydrogenated pitches begin at lower tempera-
tures. Moreover, it was found that the tritium added to pitch was difficult to release relative
to the original hydrogen in pitch. In processes for deriving a high performance carbon fiber
from coal tar pitch, the hydrotreatment of raw pitch by gaseous hydrogen in the presence of
a catalyst or by hydrogen donor solvent is usually used to improve the properties of starting
pitches, i.e. the texture of optically anisotropic mesophase during the carbonization. In this
process, hydrogen is generally introduced into a polyaromatic ring to form a structure with
naphthene groups to enhance the fluidity of pitch required for forming mesophase. The re-
174 3 Pyrolysis
Fig. 3.31 Scheme of isotope exchange and addition reaction for coal tar pitch.
Note: The ratio of tritium to hydrogen in aromatic rings after exchange can be inferred to be about 50%
according to the reported results (Garnett and Kenyon, 1971; Keith and Barnett, 1975; Asante and
Stock, 1986), if it is assumed that two hydrogen atoms in the methyl group were isotopically ex-
changed. However, the tritium orientation in the aromatic rings is unclear.
[Reproduced with permission from Wang. X. et al., Fuel Process. Technol., 38, 54, Elesvier (1994)]
H(T) H ~.
H(T) H
HT2
9 or etc. 9Further
Fig. 3.32 Carbonization scheme of coal tar pitch tritiated by isotope exchange.
[Reproduced with permission from Wang. X. et al., Fuel Process. Technol., 38, 54, Elesvier
(1994)]
~ Hc T H
H~
-- Further
(c)
Fig. 3.33 Carbonization scheme of coal tar pitch tritiated by isotope addition.
[Reproduced with permission from Wang. X. et al., Fuel Process. Technol., 38, 54, Elesvier
(1994)]
suit obtained from our experimental confirmed the viewpoint mentioned above. The hydro-
gen introduced into pitch by hydrogen addition could be retained at higher temperatures
during carbonization, and thus the fluidity of pitch can be maintained to higher tempera-
tures.
Based on the above results, the carbonization mechanism of coal tar pitch was estimat-
ed using the model system shown in Figs. 3.31-33. Methylnaphthalene (A) which is a typi-
cal structure in coal tar pitch (Zander, 1987) was used to represent a general pitch molecule.
Fig. 3.31 presents the isotope exchange and addition reactions for the preparation of tritium-
labeled pitch. The orientation of tritium in the product after exchange reaction has been de-
termined to be benzylic and aromatic positions by Garnett et al., who used similar model
compounds (Garnett and Hodges, 1967; Garnett and Kenyon, 1971; Keith and Garnett,
1975; Benjamin et al., 1982; Asante and Stock, 1986). The hydrogenation of coal tar pitch
on Ni-Mo/A1203 led to partial saturation of polycyclic aromatic hydrocarbon as presented in
Fig. 3.31 (Wang et al., 1994). The pyrolysis pathways of tritiated model molecules were
proposed as in Figs. 3.32 and 3.33. The pyrolysis pathways proposed for two model mole-
3.3 Pyrolysisof Coal Tar Pitch 175
cules of coal tar pitch were based on the following two hypotheses suggested by Hurd et al
(1962): (1) The aromatic ring structure without side chains is more stable than one with side
chains: and (2) methyls attached to an aromatic ring are reactive and the polymerization of
such aromatics is always initiated by the formation of benzyl radicals or by the elimination
of methyl groups. In Fig. 3.32, initially oligomers of the model compound of pitch are giv-
en via the formation of radicals from starting compounds or elimination of methyl side
chains, which are accompanied by the evolution of a great amount of tritium. This pathway
explains the phenomenon observed from the carbonization of coal tar pitch tritiated by iso-
tope exchange, where tritium was released more rapidly than hydrogen. In Fig. 3.33, how-
ever, thermal dehydrogenation or cleavage of an unsaturated ring may also occur as side re-
actions. This scheme illustrates that, in hydrogenated pitch, the release rate of hydrogen
was more rapid than that of tritium and that the release rate of hydrogen in hydrogenated
pitch was more rapid than that of hydrogen in unhydrogenated pitch. The cleavage of un-
saturated rings would be a reason for the excessive weight loss of hydrogenated coal tar
pitch.
XPS study provided some information on the behavior of carbon during the plastic
phase pyrolysis of pitch. The carbonization process can be presented as the model shown
in Fig. 3.34. On heating pitch to a certain temperature, the oil fraction will be volatilized
and the softened pitch will tend to sphericize due to surface tension. Nakagawa et al.
(1985) reported the temperature at which spherical pitch forms to be above 250 ~ and that
it would depend on the properties of the parent pitch. The concept of spherical pitch used
here is different from that of the mesophase sphere because spherical pitch can consist of
either anisotropic or isotopic texture. With rising temperatures above 300 ~ carbonization
begins from the surface of spherical pitch, then advances toward the center of the sphere
with annealing time, as described in Fig. 3.34.
Table 3.18 Compositionsof Raw Pitch and Hydrogenated Pitches by Solvent Extraction
Run Composition (wt%)
HS HIS-BS BIS-THFS THFIS
Raw pitch 14.7 52.6 5.5 27.2
Run 1 16.3 33.6 3.7 46.4
Run 2 6.8 49.5 17.9 25.8
Run 3 5.9 54.5 12.2 27.4
Run 4 8.1 56.2 11.2 24.5
Run 5 3.7 62.2 15.7 18.4
Run 6 5.2 63.9 5.6 25.3
Run 7 9.5 62.5 10.2 17.8
Run 8 2.0 70.2 8.3 19.5
(Wang, 1993)
The raw pitch and hydrogenated pitches were fractionated with solvents and the com-
positions of those are shown in Table 3.18. The changes in pitch composition after hydro-
genation with catalysts (Runs 2-8) were quite different from that after hydrogenation with-
out catalysts (Run 1). In the absence of catalyst, the lightest HS fraction and the heaviest
THFIS fraction increased, and the middle HIS-BS and BIS-THFS fractions decreased after
hydrogenation. This result can be interpreted as follows: the radicals produced by pyrolysis
of pitch molecules can be stabilized only partially by the inter- and intramolecular transfer
of hydrogen to form low molecular weight species (Yasuda and Miyanaga, 1989).
However, most radicals will not be stabilized and recombine with other species to form
larger molecules, since the gaseous hydrogen is difficult to be activated in the absence of
catalyst. Further, in the absence of catalyst, hydrogenation of aromatic rings hardly occurs.
In the presence of catalyst, on the contrary, both HS and THFIS fractions decreased, and
middle fractions increased in comparison with raw pitch. These results suggest that in the
catalytic systems, the radicals were effectively stabilized by activated molecular hydrogen.
The amount of HS fraction decreases because it is more easily hydrocracked into lighter
distillates such as naphtha and light oil than the heavier fractions. In addition, it seems that
the organometallic complexes are more effective in converting the THFIS fraction into
middle fractions than supported catalysts.
Compared with molybdenum-based supported catalysts, the organometallic complexes
were more effective in catalyzing the hydrogen transfer from gaseous hydrogen to coal tar
pitch. The catalysts derived from ruthenium complexes give the highest activities, but they
were ineffective in catalyzing the hydrocracking of pitch molecules into light fractions.
Further, comparison of FT-IR spectra for the raw pitch and hydrogenated pitches shows
that the ruthenium complexes can effectively catalyze the hydrogenation of not only the
low molecular weight aromatics but also the high molecular weight aromatics in pitch.
Results obtained from the XPS or XRD analyses for the THFIS fraction of catalytically hy-
drogenated pitch showed that the organometallic complexes were well dispersed into pitch
and converted into fine metal particles of reduced states. These metal particles are consid-
ered to be active for hydrogenation of polynuclear aromatics in pitch.
It is well known that the carbonization behavior of pitches is influenced by the pres-
ence of particulate matter in the pitches, such as carbon blacks, natural graphite and mica,
carbon felt and silica (Bradford et al., 1970; Forrest and Marsh, 1983; Tanaka et al., 1986).
For instance, Obara et al. (1985) investigated the carbonization of silica-containing pitch by
means of in situ ESR, and found that spin concentration of pitch increases with increasing
silica content. Kuo et al. (1987) reported that particulate matter (diameter < 1/tm) in pitch
178 3 Pyrolysis
9o1A
IF v
Naphthalene pitch
~h,
W
A
W
~9 70
o
. ,...~ Coal tar pitch
O
o
.~ 5o
O
30 I , I
significantly influences carbonization behavior and reduces the size of the optical texture in
resultant coke. On the other hand, it has been clarified that most of the elements in the peri-
odic table are effective as catalysts on graphitization of carbon (Oya, 1980). However, few
works have been done to investigate the effects of such metal particles or metal compounds
on dehydrogenation and polycondensation during low temperature carbonization of heavy
hydrocarbons.
As mentioned above, organometallic complexes such as Mo(CO)6, Ru3(CO)12 and
Ru(acac)3 have higher activity for hydrogen transfer from gas phase to pitch molecules and
can inhibit the formation of the light fraction by hydrocracking, as compared with conven-
tional supported catalysts. As the hydrogenated coal tar pitches containing fine metal parti-
cles derived from organometallic compounds are subjected to carbonization, however, it
has not yet been clarified how the fine metal particles produced during the hydrogenation
behave and affect the carbonization of coal tar pitch.
Figure 3.35 shows the effect of the content of molybdenum on carbonization yield of
the pitches. The pitches containing molybdenum particles were prepared by two methods.
Method 1: Pitch and Mo(CO)6 were mixed mechamically in an inert atmosphere at room
temperature. Method 2: Pitch and Mo(CO)6 were mixed in an autoclave under 5.9 MPa of
nitrogen and at 250 ~ for 2 h. In the case of coal tar pitch, the addition of Mo, especially
according to Method 2, increased remarkably the yield of carbonization. It was found that
the fine metal particles derived from molybdenum carbonyl increased the carbonization
yield of coal tar pitch and selectively catalyzed the dehydrogenation and polycondensation
of saturated hydrocarbons in the pitch below 700 ~ Further, when hydrogen in pitches
was labeled by tritium with the reaction of pitch and tritiated water catalyzed by Pt/A1203,
the release of tritium during carbonization of pitch was independent of molybdenum parti-
cles and was more rapid than that of hydrogen. Since tritium is incorporated mostly into
benzyl positions in pitch molecules, the result implies that hydrogen at benzyl positions was
initially released independent of the presence of metal particles. On the contrary, increase
in carbonization yield by adding molybdenum was scarcely observed. This is considered to
be due to the differences in structure and composition between the two pitches.
The influence of molybdenum particles on the structure of resultant coke was also in-
3.3 Pyrolysisof Coal Tar Pitch 179
H T~.,T H T- T H HT. ~, H H T H T H
3 I H(T) H(T) H(T) H(T)
D
~ H H T H T H
1
H
(T)~. Further c o n d e n s a t i o n
III CH - Aromatics and C,-C5 hydrocarbons
I 3
H (T)~,,r~CT2_CH2_CH_CT2.
H T
H T H H T H
H(T) ~ C H 3
- Further condensation
H(T)
H(T) ~ f H H f ~ H(T)
H T H T T H
vestigated by means of XRD and XPS. The results indicated that growth of graphite crystal
structures in the presence of molybdenum particles was slightly slower than that in the ab-
sence of molybdenum particles.
Based on the results above, the carbonization mechanisms of coal tar pitch with and
without molybdenum particles could be illustrated comprehensively in Fig. 3.36. On the
basis of the structural features of coal tar pitch (Qian and Ling, 1990), 1,2,3,4,-tetrahydro-2-
methylnaphthalene was chosen as the model compound of coal tar pitch. The orientation of
tritium in the product of an isotope exchange reaction catalyzed by alumina-supported plat-
inum using similar model compounds has been determined to be benzyl positions as well as
aromatic hydrogen (Garnett and Hodges, 1967; Garnett, J.L. and Kenyon, 1971; Keith and
Garnett, 1975; Benjamin et al., 1982b; Asante and Stock, 1986), Routes I, II and III in Fig.
3.36 represent the dehydrocondensation, cracking and dehydrogenation, respectively, which
are the possible reactions during the carbonization of pitches (Alonso et al., 1992). It is
thought that molybdenum particles promote not detritiation but selective dehydrogenation
in Route I. As a result of selective dehydrogenation promoted by molybdenum, Route II
was inhibited thus the carbonization yield of coal tar pitch increased. On the other hand,
the detritiation of coal tar pitch by Route ii or III proceeds independent of molybdenum,
since these reactions are easier to occur than those in Route I. In the case of naphthalene
pitch, since most hydrogens in pitch molecules are aromatic ones which could be consid-
ered to be nearly equivalent in the ease of isotope exchange, the selective dehydrogenation
by Route I and selective detritiation by Routes II and III observed for coal tar pitch would
scarcely be observed. Thus, the effect of molybdenum as well as the differences between
the rates of dehydrogenation and detritiation was much smaller. These propositions can
provide a satisfactory explanation for the phenomena observed from the carbonization of
pitches.
The Summary, pyrolysis is the first step in coal conversion by several methods such as
gasification, liquefaction and combustion. Since pyrolysis proceeds under rather mild reac-
tion conditions of low temperature and low pressure, attention has been paid to the recovery
of liquid products in high yield by utilizing pyrolysis. Pyrolysis is also one of many routes
180 3 Pyrolysis
from coal to chemicals. Most of the materials for the petrochemicals industry can be made
from coal when the need arises, i.e., when crude oil supplies become difficult. However,
the formidable problems of process development must be solved before shortages become
acute.
Pyrolysis methods that were performed to increase total volatile matter and/or the ben-
zene, toluene and xylene yields can be grouped into three categories: (1) pyrolysis in reac-
tive gas atmosphere, (2) pyrolysis of pretreated coal, and (3) catalytic pyrolysis of coal.
During the coming decades, these categories will be improved in many major and minor
ways. The utilization of coal tar and coal tar pitch also plays an important role in the cost
cutting of the pyrolysis process of coal to make these processes commercially feasible.
4
Liquefaction of Coal
4.1 Introduction
Today, the issue of energy security is essential for survival, i.e., for the sustainable develop-
ment of the global community. The world population is expected to reach 7.0 billion by
2010, and over 55% of this number will be taken up by the population of Asia. Because the
demand for petroleum and natural gas in the industrialized nations continues to rise, world
petroleum supplies are anticipated to become either unreliable or inadequate in the near fu-
ture. The local distribution of petroleum in the world has resulted in several crises and sup-
ply interruptions, especially since 1973 when the Organization of Petroleum Exporting
Countries (OPEC) first achieved sufficient power to have a major impact on world oil mar-
kets (Wilson, 1980). Nowadays, there is no doubt that coal is a valuable resource and may
be the major fuel of the 21st century. Coal can be considered to be one of the most attrac-
tive alternative sources of petroleum oil for the following two reasons. 1) The amount of
coal deposits estimated worldwide is ten times larger than that for the other carbonaceous
resources. 2) Coal resources are located more widely throughout the world than are oil re-
serves. Therefore, coal will be more widely available than crude oil in the future. It is nec-
essary to develop new processes for transformation of solid coal into clean liquid from the
viewpoint of energy security in the future.
There are many technologies for coal liquefaction: indirect liquefaction, refinement of
coal tar obtained in carbonization at 630-770 K, flash hydropyrolysis in the high pressure
of hydrogen, extraction with solvent and/or critical gas, and direct liquefaction by hy-
drogenolysis in the present of catalyst, solvent and high pressure of hydrogen. Regarding
the gasification of coal involving flash hydropyrolysis and carbonization is described in
chapter 3. The chapter emphasizes on the direct liquefaction.
In indirect liquefaction, coal is gasified at 1300 K or over in the presence of steam and
oxygen to produce a synthesis gas containing mostly carbon monoxide and hydrogen. This
synthesis gas (syngas), after being cleaned of impurities and adjusted to the desired H2/CO
ratio (if required), is converted to liquid fuels in the presence of catalysts. A unique feature
of the indirect liquefaction is the ability to produce a broad array of sulfur and nitrogen free
products including motor fuels, methanol, oxygenates (octane enhancers), and chemicals
with the use of different combinations of catalysts and process conditions. The conversion
of syngas to motor fuels is known as Fischer-Tropsch (F-T) synthesis. Commercial indirect
liquefaction plants in operation since 1955 have included coal based plants in South Africa
and the U.S., and natural gas based plants in South Africa, New Zealand and Malaysia. In
all the plants, the syngas is converted in gas phase reactors. Because of the high exotherm
associated with the reactions, it has long been known that a liquid phase reactor could offer
cost and operability advantages over gas phase reactors due to its superior heat transfer ca-
pabilities. Earlier efforts in developing a liquid phase F-T reactor after World War II were
suspended in the late 1950s because of the availability of cheap petroleum crude (Poutsma,
1980). Interest in this area was revived in late 1970s with the rise in petroleum crude price.
Scoping economics studies supported by the US Department of Energy (DOE) and the
Electric Power Research Institute (EPRI) indicated that the capability of a liquid phase re-
actor to process a low H2/CO ratio syngas from advanced coal gasifier could offer signifi-
cant cost advantages over gas phase reactors (Gray et al., 1980, Brown et al., 1982). In co-
operation with industrial organizations, DOE in 1981 began to support a research and de-
velopment program to advance the liquid phase reactor technology for coal-based syngas
conversion beyond that of the late 1950s. The initial focus of this program has been on the
liquid phase reactor technology development for methanol and F-T synthesis. The details
of the program have been reviewed (Shen et al., 1996). Liquid phase methanol develop-
ment was successfully completed at the proof-of-concept (POC) scale in 1989, and ad-
vanced to commercial demonstration in 1993 under the support of DOE Clean Coal
Technology program. Development of liquid phase reactor technologies for F-T synthesis
and for syngas conversion to oxygenates and chemicals have been under way at the POC
unit.
Direct coal liquefaction involves the conversion of solid coal to liquids without the pro-
duction of synthesis gas, a mixture of carbon monoxide and hydrogen, as an intermediate
step. Direct coal liquefaction should be the most energetically efficient method of produc-
ing liquids and this method makes it possible to obtain the highest oil yield. Although the
development of coal liquefaction will depend on both economics and the reliability of pe-
troleum and natural gas supplies from the Middle East and other main exporting areas, it is
important to develop coal liquefaction technology to insure new sources of energy in the fu-
ture.
The direct liquefaction of coal has been studied extensively. Most of the current
processes for the liquefaction of coal have been developed from early works (Bergius and
Billiviller, 1918) and have some common features. In most of the conversion processes, as
shown in Fig. 4.1 which is a flow diagram of the coal liquefaction process, coal is transport-
4.1 Introduction 183
Hydrogen
~
v Distillation M._._ Distillation} Heavy oil
ed to a coal slurry preparation process, in which the coal is dried and ground, then mixed
with a hydrogen-donor solvent and/or a coal-derived recycle oil to form coal slurry. Then
the slurry is pumped into the liquefaction process, in which the coal is liquefied by hydro-
gen at a high temperature and pressure in the absence/presence of a catalyst in one and/or
several high-pressure reactors. The liquefaction product is separated in the distillation
process. The heavy oil (residue) and a portion of light gas oil is hydrogenized in the sol-
vent hydrogenation process and transferred to the hydrogen-donor solvent.
4.1.2 Mechanism of Coal Liquefaction
It has become almost axiomatic to formulate coal liquefaction as a free radical process.
The concept has its origins in the contribution of Curran et al. (1967) who pointed out the
significance of the relationship between the extent of the conversion of the intractable coal
molecules to soluble products and the amount of hydrogen transferred to the liquid coal
products. They proposed a five-step reaction [Eqs. (4.1)-(4.5)] sequence focused on the
homolyses of carbon-carbon bonds in the coal molecules. In this reaction sequence, the
radicals produced in the initial reaction, Ri, react with other coal molecules or with hydro-
gen-atom donor-solvent molecules, DH2, to form other radicals. A variety of recombina-
tion reactions terminate the chain reactions.
Coal --->2Ri~ (4.1)
structures are formed and react readily at coal decomposition temperatures (-350 ~
Methyl and hydroxy substituents serve to increase the overall free radical reactivity of the
molecules to which they are attached. The present analysis accepts the importance of free
radicals in coal chemistry and attempts to develop a more detailed and unified view of their
structures and properties.
In many respects, liquefaction is closely related to pyrolysis. They share an identical
initial step - the thermal generation of radicals from the coal by way of homolytic bond
scission. In pyrolysis, these radicals are either capped by an internally transferred hydrogen
or they combine with carbon to form material of heavier molecular weight (char). These
two events also occur in liquefaction, along with transfer of hydrogen to the radicals from a
hydrogen source. The net effect is that liquefaction produces greater amounts of liquid and
gaseous products than conventional pyrolysis, but at the expense of additional hydrogen
consumption. Liquids from hydroliquefaction are substantially depleted of heteroatoms as
compared with either the parent coal or pyrolysis liquids. A wide range of different tech-
niques is used to make liquids from coal, even though they all share the thermal conversion
step. These methods differ in whether the hydrogen is provided from an organic donor or
from molecular hydrogen, either catalytically or non-catalytically. They also differ in
whether a solvent, and what kind of solvent, is used. Thus, a study of the physical proper-
ties of solutions of coal macromolecules in various solvents as well as the colloidal nature
of the solutions would be helpful. An understanding of the phase behavior at high tempera-
tures and under high hydrogen pressures would help to elucidate the liquefaction process.
Catalysis in liquefaction has received much attention, although thus far the use of such cata-
lysts as cobalt-molybdenum has not altered process temperature or pressure requirements
(Johnson, 1978). Research should be carded out to develop catalysts that will positively af-
fect the initial coal conversion. It is relatively easy to affect the course of reactions after the
primary products are out of the coal particle. However, by this time the product distribution
may already have been determined. If a catalyst that could influence the product distribu-
tion of the primary products as they are formed could be found, entirely different types and
quantities of products might result. It would be less important, but still valuable, to deter-
mine whether the use of catalysts in coal liquefaction improves the quality of the liquid
product. Comparisons of the heteroatom content, aliphatic/aromatic ratios, viscosity and
compatibility with petroleum liquids of catalyzed and noncatalyzed coal liquids would be
valuable in determining the best disposition of various coal liquid fractions. It would also
be valuable to: (1) understand the relationship between coal structure and its reactivity; (2)
establish the ultimate dispositions of elements; (3) better understand the kinetics and mech-
anism of the reaction involved in the catalytic effect, hydrogen transfer, etc.; and (4) devel-
op data which link coal characteristics to process conditions and product type, quality and
ultimately, utilization. In this chapter, the advances in the areas mentioned above are de-
scribed.
4.1.3 Hydrogen Transfer Reaction in Coal Liquefaction
The hydrogen-transfer reactions that occur during coal liquefaction reactions are essential
for the conversion of intractable coal molecules into liquids and soluble products. Virtually
all the practical processes for coal liquefaction, such as the solvent-refined Coal II process
(Schmid and Jackson, 1981), the Exxon donor-solvent process (Furlong et al., 1976), the in-
tegrated two-stage liquefaction process (Whitehurst et al., 1980, Neuworth and Moroni,
1981), and the Chevron coal liquefaction process (Rosenthal et al., 1982), use a portion of
the liquid coal products as a solvent for the dissolution reaction. In the recently developed
4.2 Coal Structure and Reactivity 185
Chevron coal liquefaction process, the liquefaction reaction is carded out in two separate,
but closely coupled, reactors. A slurry of the coal in a portion of the coal liquid (recycle
oil) is introduced into the first-stage reactor and the product of this phase of the reaction is
then fed into the second-stage reactor. The large coal molecules are decomposed and, in
part, dissolved in the first reactor and the initial product is refined catalytically in the sec-
ond reactor to yield the coal liquefaction products which include a fraction suitable for use
as the solvent for the reaction. The conversion reactions require not only the addition of
hydrogen but also the redistribution of the hydrogen atom already present in the coal mole-
cules. Thus, hydrogen transfer reactions occur between the coal molecules and the compo-
nents of the reaction solvent and between the coal molecules and the added hydrogen.
Hydrogen transfer reactions also take place between the liquid coal products, the solvent
molecules, and gaseous hydrogen. Many of these reactions, particularly those that occur in
the initial stages of the coal liquefaction process, are quite rapid even in the absence of cata-
lysts. Indeed, some coals are such good hydrogen atom donors that they only need to be
heated in a fluid medium to cause extensive degradation of the carbon skeleton with an at-
tendant redistribution of the hydrogen atoms (Neavel, 1982). More often, however solvents
that are good hydrogen-atom donors are used in coal liquefaction reactions to provide a flu-
id medium for the products as well as to provide a convenient, mobile source of hydrogen
for the decomposing coal molecules. In addition, these solvent molecules enable the trans-
fer of hydrogen atoms between the array of hydrogen donors in the solid, liquid, and gas
phases and the reactive coal molecules. The reaction pathways important for the transfer of
hydrogen atoms during coal liquefaction have been studied intensively in the past few years
to establish a more secure basis for the development of efficient methods of coal liquefac-
tion using the available hydrogen atoms in the coal molecules as well as the hydrogen
atoms in donor-solvent molecules and added hydrogen. The recent work on this matter is
also discussed in this chapter.
hydrocracking take place. These reactions occur rather slowly because fewer reactive
bonds are involved in this stage, which produces distillate range small molecules. In the
earlier Bergius process, the reaction at this stage was performed under very high pressure at
high temperature with disposable catalysts of low activity and was completed in a single
step. Current liquefaction processes utilize two or three stages under more moderate condi-
tions. Hydrogen donor solvents also assist in moderating the conditions required. Thus,
the primary products in the first stage, together with the used solvent, are further hydroc-
racked and/or hydrorefined products as well as rehydrogenated solvent. Various types of
feeds, distillates, nondistillable liquids free of minerals, the catalyst, preasphaltenes and un-
reacted coal of the first stage or whole products, including the catalyst and minerals, are
charged to the second stage, depending on the liquid/solid separation procedure utilized and
the durability of the catalyst in this stage. Staged heating is sometimes utilized in the first
stage where the reaction temperature of each reactor is controlled separately to obtain the
best oil yield with minimum formation of hydrocarbon gases and avoidance of coking
(Burgess and Schobert, 1991, Davis et al., 1989). The oil is further refined in the following
stages. Such a process scheme practiced at present is called multistage liquefaction. A se-
ries of reaction temperatures is expected to improve selectively specific reactions at differ-
ent temperatures in a series of consecutive reactions. Higher degrees of desulfurization and
denitrogenation, longer catalyst life, less sludge formation, and higher yields of distillate
are reportedly obtained by the multistage processing and refining of petroleum products
(Mochida et al., 1988a; 1990a). The function of the catalysts in the various liquefaction
stages are described in the following sections.
4.2.2 C o a l D i s s o l u t i o n , D e p o l y m e r i z a t i o n , and R e t r o g r e s s i v e R e a c t i o n s
The liquefaction of coal is the conversion of an ensemble of macromolecules as described
above into smaller hydrocarbon molecules that are distillable. Shinn (1984) has described
the changes in representative molecular structures of intermediates in the three steps of liq-
uefaction as shown in Fig. 4.2a-c. The first step in the liquefaction of solid coal is the for-
mation of liquid phase. Small molecules of the coal fuse above 350 ~ to form a liquid
phase together with solvent (if present); some macromolecules may be dissolved in this liq-
uid phase (fusion and dissolution mechanisms). Other molecules undergo thermal fission at
their weakest bonds, such as methylene and benzylether bonds, producing fragmented radi-
cals (Whitehurstet et al., 1980). When the radicals are capped with hydrogen from the sol-
vent or the catalyst, they form smaller molecules that are soluble in the solvent or even
fusible by themselves (first mechanism) increasing the quantity of liquid phase (Poutsma,
1990). This pyrolysis continues while the reactive bonds and stabilizing hydrogen are
available. Atomic or molecular hydrogen available in the reactor system can hydrogenate
reactive sites on the aromatic rings. When the ipso-position of the strong aryl-aryl bond in
the aromatic ring is hydrogenated, the bond becomes weakened and bond cleavage becomes
possible via the first mechanism of depolymerization and facile stabilization (second mech-
anism) (McMillen et al., 1987; Kamiya et al., 1988; Mochida et al., 1988b). Very reactive
hydrogen may attack the aryl-aryl bond directly, leading to its breakage (third mechanism)
(Mochida et al., 1990b). Aromatic rings are very stable unless they are hydrogenated to
naphthenic tings, which may be thermally or catalytically cracked to open the ring (fourth
mechanism) (Malhotra and McMillen, 1990). Unless the fragmented radicals are stabilized,
they recombine or react with other molecules, forming thermally stable bonds. Repetition
of recombination reactions produces large molecules that have resistance toward depoly-
merization. Coking takes place when such large molecules remain at elevated temperatures
4.2 Coal Structure and Reactivity 187
OH HO O O OH
H20
(b)
for long time periods, for example, in the locations of low flow rate, such as near reactor
walls, in bends of transfer lines, or on catalyst surfaces (retrogressive mechanism) (Bate
and Harrison, 1992). When radicals are trapped in the cage of coal macromolecules, such
188 4 Liquefaction of Coal
Table 4.2 Half-life Estimated for Carbon-Carbon Bond Homolysis in Some Representative
Compounds at 400 ~ in Tetralina
Compound b Half-life (min)
1,2-Diphenylethane, C6HsCH2-CH2C6H5 1680
2,3-Diphenylbutane, C6HsC(CH3)H-C(CH3)HC6H5 20
2,3-Diphenyl-2, 3-dimethylbutane, C6HsC(CH3)2-C(CH3)2C6H5 0.2
9-(1-Phenylethyl) anthracene, 9-CI4H9CH2-CH2C6H5 5
9-Benzyl-9,10-dihydrophenanthrene, (9,10-C14HI1)-9-CH2C6H5 14
Bitetrayl, (1-CIoHIl)-(1-Cl0Hll) 1
Benzyl phenyl ether, C6HsCH2-OC6H5 0.7
Benzyl phenyl thioether, C6HsCH2-SC6H5 1
aThe values were selected from the compilation provided by Stein (1981).
bThe carbon-carbon, carbon-oxygen, or carbon-sulfur bond cleaved in the hemolytic reaction is
indicated in the structural reprasentation.
[Reproduced with permission from Stein, S. E., ACS Symposium Ser, No. 169. New Approaches
in Coal Chemistry, 7, 104 (1981)]
retrogressive reactions come to act and are accelerated as the radicals frequently encounter
each other. The cage hinders liberation of radicals and the participation of donors. Hence,
the dissolution of depolymerized coal molecules to break the cage is very important and ef-
fective in the liquefaction process (Sakata et al., 1990). Strong dissociative properties of
the solvent are important in minimizing macromolecular interactions of coal components or
coal-derived products.
4.2.3 F r e e R a d i c a l s in C o a l L i q u e f a c t i o n
It is generally accepted that free radicals are key reactive intermediates in thermal coal
chemistry. However, because of the chemical complexity of coal, structural and kinetic
properties of these radicals cannot be directly obtained from experiments on coal itself.
Therefore, the work that is based on results of well controlled "model" experiments and
predictive theory has been carded out (stein 1981). This work extends a previous analysis
of coal chemistry (Stein, 1981) in which basic predictive theory was first applied to coal-re-
lated molecules and reactions and then used to analyze selected results of model compound
experiments. The present analysis uses these predictive methods and results along with rel-
evant experimental data to divide coal-derived free radicals into classes according to their
reactivity and to examine the probable behavior of each of these radical classes in coal con-
version chemistry. The primary focus of this work is on reactions below 500 ~ To make
this analysis tractable, the following working assumptions are made:
1) Coal is composed of randomly oriented, substituted, hydroaromatic clusters tied together
by short covalent linkages (biphenyl, methylene, ether (Whitehurst et al., 1980)
2) Free radical reactions account for all covalent bond breaking and forming processes and
most types of hydrogen transfer
3) All varieties of elementary free radical reactions of importance in coal conversion are al-
ready known (O'Neill and Benson, 1973)
4) Steric and diffusional effects do not influence reaction kinetics.
In the past decade, many contributions have been presented to support the general reac-
tion scheme as shown in Eqs. (4.1)-(4.5). The first reaction in the sequence involving the
homolytic cleavage of carbon-carbon bonds in the coal molecules is the critical step.
Studies of the kinetics of the decomposition of a variety of compounds with relatively weak
carbon-carbon bonds have shown that the rates of decomposition of these substances by
well-known free radical pathways are comparable with the rates of decomposition of coal
4.2 CoalStructure and Reactivity 189
molecules. This feature is well illustrated by the data presented in Table 4.2 (Stein, 1981),
One of the best lines of evidence for the involvement of free radicals in these processes
stems from the study of the rapid pyrolysis of coal in the cavity of EPR spectrometers
(Petrakis and Grandy, 1981; Sprecher and Retcofsky, 1983). In one quite pertinent study.
Sprecher and Retcofsky investigated the thermal decomposition or a bituminous coal sus-
pended in silica (Sprecher and Retcofsky, 1983). The concentration of radicals increased
by a factor of 4 in about 5 min at 470 ~ The addition of an equal weight of 9,10-dihy-
drophenanthrene to the reaction mixture inhibited the formation of radicals, whereas the ad-
dition of an equivalent amount of phenanthrene had no effect on the radical concentration.
In addition, it was found that the radical concentration increased by an additional factor of 2
when the volatile products formed in the pyrolysis of this coal were allowed to escape from
the reaction vessel. These results strongly support the idea that coal molecules decompose
by homolytic reactions to yield transient reactive radicals. In addition, the results are com-
patible with the view that hydrogen-atom-abstraction reactions occur rapidly with effective
donor molecules such as 9,10-dihydrophenanthrene. As illustrated in reactions (4.2) and
(4.6), a new coal molecule is produced together with a mobile radical. These react with
other coal molecules as illustrated in Eqs. (4.3) and (4.7) to yield new coal molecules and a
different series of coal radicals resulting from the abstraction of hydrogen atoms rather than
from the homolyses of carbon-carbon bonds.
Ri 9 + volatile products ---->Rill nt-R3" (4.7)
The fourth reaction in the sequence describes the behavior of donor-solvent molecules,
such as tetralin and 9,10-dihydrophenanthrene, that can readily be oxidized to aromatic
compounds. The abstraction reactions of the dihydronaphthalenes [Eq. (4.9)] are more
rapid than the abstraction reactions of tetralin. The removal of hydrogen atoms from the in-
termediate radicals by reactions with other coal radicals are also presumably very rapid
processes. Some radicals derived from the coal and from the solvent engage in dehydro-
genation reactions. The fifth reaction in the basic sequence illustrates the notion that hydro-
gen atoms are transferred among coal molecules to yield both hydrogen-rich and hydrogen-
poor compounds during the coal liquefaction process (Neavel, 1982). These processes are
@o + Ri 9 ~ ~ -+Rill (4.10)
complemented by another series of reactions which are initiated by the abstraction of hy-
drogen atoms from the coal molecules to provide another unstable series of radicals [Eq.
(4.3)]. These reactions are certain to be important in coals with abundant hydroaromatic
190 4 Liquefactionof Coal
structures because the benzylic carbon-hydrogen bonds are relatively weak and the activa-
tion energies for the transfer of hydrogen atoms from these positions to other coal radicals
are modest. While some of the radicals formed in benzylic hydrogen-atom abstraction
eventually undergo aromatization, other radicals formed as outlined in Eq. (4.3) decompose
by the very well-known zr-scission process (Sweeting and Wilshire, 1962; Collins et al.,
1979; Poutsma and Dyer, 1982; King and Stock, 1984) to yield a new radical and a highly
reactive alkene as illustrated for 1,3-diphenylpropane in Eq. (4.11). The fragmentation re-
actions decrease the molecular weight of the coal molecules and the reactive products prop-
agate the liquefaction reaction.
C6HsCHCH2CH2C6H5 -'-) C6HsCH2 9+ C6HsCH - CH2 (4.11)
While most discussions have focused attention on the dominant free radical processes
that take place during the thermal decomposition of coal molecules during coal liquefac-
tion, it is apparent that pericyclic reactions can make a major contribution to the degrada-
tion reactions of the complex molecules under appropriate conditions. There are a variety
of plausible pericyclic reactions that must be considered in the development of an adequate
theory for the noncatalytic thermal reactions of coal molecules. Moreover, processes of this
kind are certain to be more important in the more severe reactions of coal molecules at tem-
peratures in excess of 500 ~ The reactions include the unimolecular transfer of hydrogen
from one hydrocarbon to another hydrocarbon. Doering and Rosenthal (1967) provided the
classic example of the reaction when they showed that the Z isomer, rather than the E iso-
mer, of 1,2-dimethylcyclohexane was obtained preferentially (6% yield) during the decom-
position of the dihydronaphthalene. However, few other authentic examples of the reaction
have been reported presumably because free radical chain reactions occur competitively ob-
scuring the nonchain pericyclic reactions (Fleming and Wildsmith, 1970). For example,
studies of the hydrogen-transfer chemistry of 1,2- and 1,4-dihydronaphthalene and other di-
hydroaromatic compounds unequivocally indicate that pericyclic processes are not involved
in the product-forming stages of the reactions with E-stilbene, phenanthrene, and tetracene
(King and Stock, 1981) and Heesing and Mullers (1980) demonstrated that the hydrogen-
transfer reactions which take place during the disproportionation of 1,2-dihydronaphthalene
at 300 ~ do not occur in a pairwise fashion or stereo specifically.
H H3C H3C H
(4.12)
H H3C U3C H
~ @ + H2 (4.13)
The notion that pericyclic reactions are important for carbon-carbon bond-cleavage re-
actions remains a matter of controversy. Virk and his co-workers have advanced the view
that retroene reactions and related processes are important in the decomposition reactions of
4.3 Catalysisin Coal Liquefaction 191
retroene process requires the formation of toluene-2-d [Eqs. (4.17)-(4.18)]. No more than
2% toluene-2-d is obtained during the decomposition reaction in tetralin at 400 ~ Thus,
reactions Eqs. (4.14)-(4.16) appear to be the principal toluene-forming reactions. While it
seems clear that retroene processes are not responsible for the product-forming reactions, it
is quite possible that such pericyclic processes may be responsibleindirectly for the
cvI:
C6HsCH2CD2CD2CH2C6H5 + C6HsCHzCD = CD2 (4.17)
H
D
[~ CH2 several
H steps > Toluene-2-d (4.18)
D
initiation of free radical chain reactions. This idea requires consideration because the
retroene reactions often produce very unstable intermediates. To examine the concept, we
recently studied the decomposition of 9-[3-(perdeuteriophenyl)propyl-3,3-d2]-phenanthrene
(Stock, 1985). Group-additivityconsiderations suggest that the retroene reaction is
CH2CH2CD2C6D5 CH2
endothermic by no more than 38 kcal mol-1. However, the products formed in this reac-
tion are quite unstable under the conditions of coal liquefaction and may initiate free radical
chain reactions with quite low effective activation energies. Thermal chemical analyses
(Benson, 1976) suggest that the molecule-induced homolysis shown in Eq. (4.20) is en-
dothermic by only 21 kcal mol-1. More important, the abstraction of the benzylic and
CH2 CH2 ~
allylic hydrogen atom in the 9-methylphenanthrene isomer is a very facile process [Eq.
(4.21)] which leads to a new radical capable of initiating the decomposition of the original
alkylphenanthrene by a conventional reaction sequence. Considerations of this kind sup-
port the view that certain low-energy pericyclic reactions may lead to the initiation of free
radical chain reactions. In this sense, such reactions may influence the rates of the thermal
decomposition reactions of coal molecules.
CH2 CH2 ~
Phenol --+ ~ O
-+ CO + Cyciopentadiene (4.22)
postulated as the active form. Its precursors are red mud, residue of bauxite after the sepa-
ration of alumina, iron ores of various sources, synthetic and natural pyrite, fine iron parti-
cles, iron dust from converters, iron sulfate, iron hydroxide, etc. (Montan et al., 1981;
Suzuki et al., 1989). It is also important to elucidate the catalytic roles played by iron-
based catalysts and sulfur, which are added during coal liquefaction.
The next most widely used materials are Co-Mo and Ni-Mo sulfides, which have been
widely used in petroleum refineries. They are usually supported on alumina of designed
pore structures in which the pore diameter is usually larger than that for conventional petro-
leum residue (Stephens et al., 1985; Derbyshire, 1989).
A third type of material is the chloride of transition metals, such as ZnC12 and SnC14
(Mobley and Bell, 1980; Mizumoto et al., 1985). This group of catalysts works in molten
state in contrast to the solid state of the previous two groups. The corrosive nature and in-
stability may exclude their practical application. No details are reviewed here. Ru has been
used as an additive to Co-Mo and Ni-Mo (Hirschon et al., 1987) to improve their hydro-
genation and denitrogenation activities. Hydrogen sulfide in the reaction atmosphere has
been reported to accelerate liquefaction directly, in addition to controlling the extent of sul-
tiding of iron, Ni-Mo, and Co-Mo catalysts (Ogawa et al., 1984; Hirschon and Laine,
1984). Recently, carbon black was reported to catalyze coal liquefaction (Farcasiu and
Smith, 1990, 1991); this may initiate radical reactions of bond breakage.
4.3.1 Preparation of Catalysts
Solid liquefaction catalysts have been prepared by three procedures.
1) Fine Powder Catalysts. Most iron catalysts are used in powdered form. Since their par-
ticle size strongly influences their activity, fine powders are preferred. Natural products are
ground extensively. Magnetite for the magnetic tape is needle-like crystal of which the di-
ameter is less than 1 /2m. Recently, ultra-fine powders (nanometer to tens-of-nanometer
size scale) of iron oxides and sulfides have been prepared by means of vapor-phase hydrol-
ysis of volatile compounds in a hydrogen-oxygen flarhe to produce nanometer-sized iron
oxides (aerosol) (Bacaud et al., 1993; Lacroix et al., 1989); rapid thermal decomposition of
solutes (RTDS), such as Fe(NO3)3 solutions (Matson et al., 1993); laser pyrolysis of
Fe(CO)5 and C2H4 to produce iron carbides followed by in situ sulfidation (Hager et al.,
1993); precipitation/crystallization sequence from the sulfated and oxyhydroxides of iron
(Pradhan et al., 1993); and a chemical reduction or an exchange/replacement reaction of
iron salts solubilized in inverse micelles of reaction media (Martino et al., 1993). Finer
powders of the iron sulfide are expected to be expensive as well as active. The cost/perfor-
mance should be carefully evaluated.
2) Supported Catalysts. Co-Mo and Ni-Mo sulfides are usually supported on alumina.
Selection of the specific alumina is conventionally studied on the basis of the pore size dis-
tribution and acidic characteristics. The supporting procedure and the amount of supported
sulfides are very influential in catalyst activity. Alternative supports to alumina are the fo-
cus of current research. Titania and carbon have recently been examined as supports for
iron and Ni-Mo sulfides (Mochida et al., 1984; Derbyshire et al., 1986a). Bifunctional and
strong interactive roles of the support should be emphasized in addition to physical proper-
ties (Ehrburger et al., 1976; Showa, 1982). The search for additives such as phosphate and
sulfate, which have been utilized for commercial Co-Mo and Ni-Mo base catalysts, has also
been receiving much recent attention (Lewis and Kydd, 1991; Morales et al., 1984;
DeCanio et al., 1991).
3) Highly Dispersed Catalysts on Coal. Sulfide catalysts have been dispersed directly on
194 4 Liquefaction of Coal
the coal surface. Very high dispersion on the catalyst may allow direct interactions be-
tween the catalyst and solid coal. The first application of this approach utilized molten
chloride as the starting material. Later, oil- and water-soluble iron precursors were impreg-
nated or ion exchanged onto the coal surface through the interaction with oxygen functional
groups (Cugini et al., 1991; Hirschon and Wilson, 1991a; Curtis and Pellegrino, 1989;
Snape et al., 1989; Pradhan et al., 1991). Recently, highly dispersed, highly active, or high-
ly functional catalysts have been extensively investigated to reduce the amount of catalyst
required for recovery and regeneration (Suzuki et al., 1984, 1987, 1991; Holloway and
Nelson, 1977; Takemura et al., 1989; Curtis and Cahela, 1989; Ryan and Stacey, 1984;
Mendez-Vivar et al., 1990). Very fine particles of iron sulfide are very promising catalysts
because of lower cost and moderate activity. Presulfiding treatments for activation, ion ex-
change, and dispersed impregnation of catalysts or catalyst precursors are combined to en-
hance the catalytic activity and reduce the amount of catalyst required (Naumann et al.,
1982; Yokoyama et al., 1983). The use of highly dispersed catalysts from soluble salts of
molybdenum is another approach to the reduction of catalyst amount because of their excel-
lent activity despite their higher price. Recently, metal carbonyl com-pounds, such as
Fe(CO)5, Ru3(CO)12, and Mo(CO)6 have been investigated as metal cluster catalysts.
Preparation involved their deposition and decomposition on catalyst support surfaces
(Paradhan et al., 1991; Burgess and Schobert, 1991; Cowans et al., 1987). It has been re-
ported recently that highly dispersed catalyst on coal grains can accelerate the liquefaction
of the coal grains without supporting catalysts (Cugini et al., 1991; Pradhan et al., 1991).
The fine powders of the catalyst are indicated to be mobile during the liquefaction, suggest-
ing no importance of direct interaction. Recoverable catalysts also offer a promising way to
economize the cost of liquefaction catalysts (Joseph, 1991; Pelofsky, 1979). Dow designed
a process that utilized fine powders of MoS2 that were reported to be recoverable by hydro-
clone; however, specific details have not been published (Whitehurst, 1980).
4.3.2 Fe-based Catalysts
It was reported that all Fe-based catalysts (FeOOH, Fe203, pyrite, FeSO4, etc.) show similar
yields of liquid when 0.5% Fe was added. On the other hand, FeOOH catalyst that was dis-
tributed on the surface of coal as fine particles via impregnation showed higher yield of liq-
uid than other catalysts, as shown in Fig. 4.3 (Cugini et al., 1994). The sintering of the
FeOOH nanometer particle catalyst still occurred in the sulfidation and resulted in a de-
crease of surface area even though its initial value was ca. 138 m2/g, resulting in lower cat-
alytic activity. This indicates that the higher the distribution of the sulfides is, the higher
the catalytic activity.
It was proposed that natural pyrite was activated via the grinding in NEDOL process
because of the lower activity of natural pyrite. The relationship between particle size of
pyrite and the yield of liquid in the liquefaction of Wandoan coal is shown in Fig. 4.4
(Hirano et al., 1997). The yield of oil significantly increased when the average particle size
was under 50 ~tm, and the yield increased from 40% for particle size of 100/~m, i.e., un-
ground pyrite, to 58% for particle size of 0.86/.tm, which is over the objective of the
NEDOL method. By XPS analysis, it was found that there is ca. 10% of SO~- on the sur-
face of natural pyrite. In contrast, most of the surface of ground pyrite in air is present in
the form of SO 2-. Thus, the grinding in the oil phase or the addition of sulfur after the oxi-
dation will restore the catalytic activity of the pyrite (Hirano et al., 1997).
The preparation of nanometer particle ~,-geitite (FeOOH) was developed by Kaneko et
al. (1995). In the liquefaction of Yallourn coal, the yield of oil was only 40-43% in the
4.3 Catalysis in Coal Liquefaction 195
100
75-
50
~.
~D
25
I I I I I I I
+ 2 0= 0 +~ z
<
=
o
o 0 0
2: 0 0
J M J
Pow~der Y
Impregnation
[] CH2C12 soluble, 9 Heptane soluble
Fe203 5000 ppm, FeOOH 2500 ppm, Ammonium heptamolydate (AHM) 1500 ppm
case using pyrite and a-FeOOH ground to a particle size of 0.5 pm. On the other hand, the
yield of oil was 50% using ~,-FeOOH of 1.2 pm, and increased to 55% when ~,-FeOOH of
0.4 pm was used. The higher activity of 7'-FeOOH was attributed to the difference in crys-
tallization, whereby ~,-FeOOH can be readily transformed to the active phase, i.e.,
pyrrhotite at temperatures under 250 ~ and growth of crystal at high temperatures was
suppressed. In the analysis of XRD of variable temperature units in the sulfidation atmos-
phere for FeS2, ?'-FeOOH and limonite, it was found that ?'-FeOOH and limonite were com-
pletely transformed into pyrrhotite at 300 ~ whereas the peak of FeS2 was still present at
300 ~ and a temperature of 400 ~ was necessary for its complete disappearance. Further,
the crystal size of sulfided ~,-FeOOH and limonite was c a . 10 nm at 250 ~ and merely
grew to 20 nm even at 400 ~ In contrast, the crystal size originating from FeS2 was c a .
40 nm and second particles further formed due to cohesion of first particles. Consequently,
the crystal size was 2-4 times greater than when ?'-FeOOH and limonite were used. Thus,
it is important to form active pyrrhotite at lower temperatures and to avoid the growth of
greater particles as much as possible (Suzuld and Ikenaga, 1998).
In addition, Fe(CO)5 and the Fe(CO)5-sulfur system are also often used in the liquefac-
tion of coal. It is important to elucidate the catalytic roles played by highly dispersed cata-
lysts and sulfur which are added during coal liquefaction although these are not well de-
fined. On the other hand, since the reactions involved in coal liquefaction include hydroc-
racking and hydrogenation by molecular hydrogen and donor solvents, a number of at-
196 4 Liquefactionof Coal
70
60-
9 50-
-.,,.
40
0.1 1.10 ll0 100
Average radius (~m)
Fig. 4.4 Relationshipbetween average radius of pyrite particle and oil yield in liquefaction of Wandoan coal at
450 ~ and under 75 kg/cm2of H2 for 1 h in a 5 L autoclave. [Reproducedwith permissionfrom Hirano,
K. et al., J. Jpn. Inst. Energy, 75, 911 (1997)]
tempts have been made to elucidate the mechanism of hydrogen transfer occurring during
coal liquefaction in the presence of solvents (Billmers et al., 1986; Camaioni et al., 1993;
Autrey et al., 1995; McMillen et al., 1985, 1991; Malhotra and McMillen, 1993; Murakata
et al., 1993). A useful method for clarifying the mechanism of hydrogen transfer in coal
liquefaction is to utilize isotopes such as deuterium and tritium tracers (Fu and B laustein,
1963; Franz and Camaioni, 1980; Brower, 1982; Schweighardt et al., 1976; Cronauer et al.,
1982; Skowronski et al., 1984; King and Stock, 1982; Collin et al., 1980). Recently, it has
been reported that tritium and carbon-14 tracer techniques were effective in quantitatively
monitoring the hydrogen during coal liquefaction (Kabe et al., 1986a, 1987a, 1987b, 1990d,
1991 a, 1991 b; Montano et al., 1981; Ishihara et al., 1993, 1995). These reports showed that
the hydrogen mobility of coal and coal-related compounds could be quantitatively analyzed
using the hydrogen exchange reactions occurring between coal, the gas phase and the sol-
vent, as well as by considering the hydrogen addition reactions. Recently, Godo et al.
(1997a) reported a liquefaction of Taiheiyo coal in the presence of conventional and dis-
persed iron catalysts (pyrrhotite and Fe(CO)5) and in the presence of sulfur, in which a tri-
tium and radioactive 35S dual-tracer method was used.
As we know, tetralin is one of the most simple and convenient model compounds be-
cause it contains both an aromatic ring and a naphthene ring, and because it can serve as a
hydrogen donor solvent. Before investigating the complex reaction with coal, the reaction
of tetralin with tritiated hydrogen was performed in the absence of coal. Figs. 4.5 and 4.6
respectively show the products and tritium distributions in the exchange reaction between
tetralin and tritiated hydrogen. Table 4.3 shows the amount of hydrogen exchanged be-
tween the gas phase and the solvent. Although naphthalene (NP), 1-methylindan (MI) and
n-butylbenzene (BB) were produced in the absence of a catalyst and sulfur (Run 1), the
yield of each product was very low, with the virtual absence of hydrogen exchange. When
sulfur was added (Run 2), the yield of NP increased significantly (from 0.7 to 3.0%), with
most of the added sulfur converted into hydrogen sulfide (Table 4.3). If all the added sulfur
reacted with tetralin to produce hydrogen sulfide, it would correspond to the production of
2.7% of NP. These results indicate that sulfur promotes the dehydrogenation of tetralin to
4.3 Catalysis in Coal Liquefaction 197
Feed
~:~:~:~1~
-- lg ~iii!zi~ D Butylbenzene
[-1 Methylindan
3 Pyrrhotite
4 Pyrrhotite lg
5 Fe(CO)5 lg
6 Fe(CO)5 2g ~:iii~iii
0 ' 92' 94' 96' 9'8 ' 100%
Fig. 4.5 Products in the exchange reaction between tetralin and tritiated hydrogen.
[From Godo, M. et al., Energy Fuels, 11, 726 (1997)]
2 - - lg 9 Solvent
[] Gas
3 Pyrrhotite
4 Pyrrhotite 1g
5 Fe(CO)5 1g
6 Fe(CO)5 2g
Tritium equilibrium
I I I ' I '
0 20 40 60 80 100%
Fig. 4.6 Tritium distribution in the exchange reaction between tetralin and tritiated hydrogen.
[From Godo, M. et al., Energy Fuels, 11,726 (1997)]
produce naphthalene. Further, the yield of MI also increased from 0.3 to 0.6%, and the
yield of BB from 0.1 to 0.3%; the tritium distribution in the solvent increased from 0.6 to
7.7%.
When pyrrhotite was added (Run 3), the yields of NP, MI and BB were respectively
2.0, 0.7 and 0.4 %, and the tritium distribution in the solvent was 59%. These values there-
fore increased significantly, but the amount of HzS generated was very small. When
pyrrhotite and sulfur were added simultaneously (Run 4), the tritium distribution in the sol-
vent amounted to 64%. This was very close to the equilibrium value of 82%, which was
calculated on the assumption that the hydrogen atoms were completely and randomly dis-
persed between the gas phase and the solvent. It was assumed that the independent effects
of the sulfur and pyrrhotite would considerably increase the amount of hydrogen exchange.
When Fe(CO)5 and sulfur were used (Runs 5 and 6), the yields of MI and BB were not
as high as that obtained with pyrrhotite and/or sulfur. The tritium distributions in the sol-
vent in Runs 5 and 6 increased to 16% and 27%, respectively, significantly greater than in
the absence of catalysts (Runs 1 and 2). However, these values were smaller than those ob-
198 4 Liquefactionof Coal
tained with pyrrhotite. In the case of Fe(CO)5, part of the added sulfur produced iron sul-
fide, with the remainder being converted into H2S (Table 4.3). After the reaction in the
presence of pyrrhotite and Fe(CO)5, iron sulfide was recovered from the inside of the auto-
clave. Compared with the pyrrhotite reaction, the particle size of the iron sulfide recovered
from the reaction with Fe(CO)5 and sulfur was larger and more uniform. Further, many
large fragments had plain surfaces. It was assumed that these fragments had been deposited
on the inside wall of the autoclave. Although Fe(CO)5 is one of the most highly dispersed
iron catalysts used in direct coal liquefaction, the particle size of the iron sulfide recovered
in the absence of coal was about 1.2 pm, which was larger than the particles of pyrrhotite
(about 0.6/.tm). It appears that compared with pyrrhotite an increase in particle size results
in a decrease in the amount of hydrogen exchanged.
It was assumed that the formation of products and the hydrogen transfer from the
tetralin proceeded via a tetralyl radical, which acted as an intermediate in the conversion
and the hydrogen exchange of tetralin (Kabe et al., 1990d-e; Ishihara et al., 1995). When
radicals generated in coal react with tetralin in the system, a tetralyl radical may be formed
easily. However, if coal is not included in the system, a tetralyl radical is difficult to gener-
ate. In the system of tetralin and gaseous hydrogen, tetralin may collide with itself to give a
tetralyl radical (Eq. (4.24)). The hydrogen exchange between tetralyl radicals and tritiated
hydrogen or tritiated hydrogen sulfide can be assumed to proceed via Eq. (4.25) depending
on the concentration of tetralyl radicals, which also controls the formation of methylindan
in Eq. (4.26). The products were methylindan by isomerization, butylbenzene by hydroc-
racking and the main product, naphthalene by dehydrogenation. Decalin by disproportiona-
tion was not formed. This is consistent with a previously reported result (Hooper et al.,
1979; Jan et al., 1984). Butylbenzene may be formed by the dissociation of tetralin with
hydrogen at on as shown in Eq. (4.27). Concerning the hydrogen exchange reaction of
tetralin, studies using deuterium have been reported extensively. In hydrogen exchange be-
tween tetralin-dl2 and hydrogen in coal at 400 ~ 1 h in a shaken autoclave system, pro-
tium was incorporated into H~ (66%), H~ (23%) and Hat (11%) positions in tetralin and that
H~ absorption of the recovered naphthalene in ~H nuclear magnetic resonance (NMR) was
approximately seven times as intense as H~ absorption (Skauronski et al., 1984). Collin
and Wilson (1983) showed in their NMR measurement that, in the reaction of tetralin with
deuterium and coal, the intensity of the signal at the H~ position was larger than that at the
H~ position. These reports indicate that the 1-tetralyl radical appears to be a more impor-
tant intermediate than the 2-tetralyl radical in the conversion and the hydrogen exchange of
tetralin, although it has been proposed that 2-tetralyl radical may be an intermediate in the
formation of methylindan (Franz and Camaioni, 1980). Therefore, the formation of the 1-
tetralyl radical in this system may be the rate-determining step for both the conversion of
4.3 Catalysisin Coal Liquefaction 199
~ + H. (4.24)
T
HT or TSH
+ H. or SH. (4.25)
CH2 9 CH3
H9
(4.26)
H9
+ H. . (4.27)
+ SH 9 (4.28)
+ H2S (4.29)
+ SH 9 (4.30)
tetralin and the hydrogen exchange. It was previously reported that the rate of conversion
of tetralin in the presence of H2S was nearly equal to that in the absence of H2S (Godo et
al., 1997a). In contrast to the result with H2S, it was considered that sulfur promoted the
formation of tetralyl radicals. The processes of radical formation could be represented as
Eqs. (4.28)-(4.31), which is consistent with the fact that sulfur promotes naphthalene for-
mation. Because sulfur promotes the formation of tetralyl radicals, the concentration of
tetralyl radicals will increase. As a result, the formation of products was promoted by sul-
fur. When pyrrhotite was added, the products yields increased and the tritium distribution
in the solvent increased significantly. Autrey et al. (1996) suggested that FeS catalysts are
reduced by donor solvent. It was considered that pyrrhotite promoted not only the forma-
tion of the tetralyl radicals in Eq. (4.32) but also the dissociation of the tritiated hydrogen
molecules in Eq. (4.33).
The liquefaction of Taiheiyo coal with tritiated gaseous hydrogen was performed in the
presence of catalysts and sulfur. Fig. 4.7 shows the distribution of the products from coal.
Without the catalyst and sulfur (Run 7), the product yield ( = 100-residue) was 74% and
the SRC was 71%. In the presence of sulfur (Run 8), although there was an increase in
light fractions (such as light oil), the product yields decreased. This suggests that sulfur si-
multaneously promoted both the thermal decomposition and polycondensation of coal. In
Cat.
-- + H-Cat. (4.32)
Cat.
T2 2T-Cat. (4.33)
200 4 Liquefaction of Coal
the presence of pyrrhotite or Fe(CO)5, the product yields in liquefaction increased to more
than 80%; the use of sulfur also increased the product yields. When Fe(CO)5 and 2 g of
sulfur were added (Run 12), the product yields reached 89%, which was higher than those
obtained with pyrrhotite. In the presence of Fe(CO)5 and sulfur, the yield of preasphaltene
decreased and the yields of oil and asphaltene increased in comparison with the use of
pyrrhotite. This indicates that Fe(CO)5 and sulfur could be used to obtain products with a
lighter composition. The result also indicated that the liquefaction activity of Fe(CO)5 was
higher than that of pyrrhotite.
Figure 4.8 shows the tritium distributions in Taiheiyo coal liquefaction. Compared
with the absence of sulfur (Run 7), the addition of sulfur (Run 8) increased the tritium dis-
tribution in the solvent and decreased that in coal. The use of pyrrhotite only (Run 9) in-
creased the distributions in both coal and the solvent. When pyrrhotite and sulfur were both
added (Run 10), the tritium distribution in the solvent increased and that in coal decreased.
It was believed that the sulfur promoted the hydrogen transfer from the gas phase to the
solvent, and that the reaction mechanism was the same as that occurring in the absence of
coal (described above). In contrast, when Fe(CO)5 and sulfur were used, the tritium distrib-
ution in the coal increased in spite of the presence of sulfur, suggesting that the catalyst de-
rived from Fe(CO)5 and sulfur acted directly on the coal and increased both the rate of coal
12 Fe(CO)5 2g
' I ' I ' I ' I '
0 20 40 60 80 100 daf%
Fig. 4.7 Products from coal in Taiheiyo coal liquefaction using gaseous tritium.
[From Godo, M. et al., Energy Fuels, 11,728 (1997)]
i Coal
8 -- lg I~ Solvent
Gas
9 Pyrrhotite --
10 Pyrrhotite 1g
iiiiiiii iiiiiiiiiiiii iiiii!iii i iiiiiiiill
11 Fe(CO)5 1g /iiiiiiii!ii!i!iiiiil
12 Fe(CO)5 2g
o 'o4o6'o8o oo
Fig. 4.8 Tritium distribution in Taiheiyo coal liquefaction using gaseous tritium.
[From Godo, M. et al., Energy Fuels, 11,728 (1997)]
4.3 Catalysisin Coal Liquefaction 201
-- Pyrrhotite 1 g
II HzS
10 + Pyrrhotite 1 g [-1 Solvent
I Coal (THFS)
-- Fe(CO)5 1g VA Coal (THFI) + Cat.
-- Fe(CO)5 2g
+ Fe(CO)5 1g
12 + Fe(CO)5 2g
0 20 40 60 80 100 %
Fig. 4.9 35S distribution in Taiheiyo coal liquefaction.
[From Godo, M. et al., Energy Fuels, 11,729 (1997)]
fer into tetralyl radicals and coal radicals. This would increase the amount of hydrogen ex-
change and addition via the radical reaction mechanism.
[35S]H25 [35S]HzS
[35S]SH~ H,
1 [35515H.
[ '~ ([3'S]Fe'-xS., ~
~..--...._.._.__..._-
Fig. 4.10 Possible mechanism of sulfue exchange.
[From Godo, M. et al., Energy Fuels, 11,729 (1997)]
60
50 O"
40
r~
"O
- - 30
9
.,..~
20
10
, IFI , I , I , I , I , I ,
00 _..40 60 80 100 120 140
Reaction time (min)
A Gas, 9 Oil, 9 AS, [] PA, 9 Residue
Fig. 4.11 Effectof reaction time on liquefaction of Wyoming coal in the presence of Ni-Mo/KB catalyst at
440 ~ and under 13 MPa of H2 (Tetralin/Coal = 1.5, Catalyst/Coal -- 3 wt%).
[Reproduced with permission from Sakanishi, K. et al., Energy Fuels, 10, 218 (1996)]
cyclohexane, with a commercial Co-Mo/A1203 catalyst in order to investigate the carrier ef-
fect. There is almost no difference among the catalysts, and the effect modified in the sur-
face by TiO2 cannot be recognized. This was attributed to that catalytic activity in the coal
liquefaction depended on surface area, and the micropore structure such as pore volume.
Thus although the surface nature of the prepared catalysts was modified, the micropore
structure remained still no change due to without the modification for the support.
Tian et al. (1996) investigated the effect when Mg or Mo is added to the iron sulfide as
the second element as well as the addition method. Catalytic activity was improved by
adding Mo while there was no influence on the activity and the selectivity of the iron sul-
fide catalyst when adding Mg. Especially, the conversion and oil yield increased respec-
tively 8% by impregnating Mo into the iron sulfide.
The oil soluble or water soluble compounds as a high distributed catalyst precursor,
which is considered to be effective in the coal liquefaction, such as metal carbonyl (Ikenaga
et al., 1997; Warzinski and Bockrath, 1996; Zhang et al., 1997), metal naphtenate (Yoon et
al., 1997) and ammonium tetrathiomolybdate (ATTM) (Burgess and Schobert, 1996; Song
et al., 1997a, 1997b; Schobert et al., 1997; Schroeder et al., 1997) has been discussed.
Ikenaga et al. (1997) found that in the presence of Ru3(CO)12 or Mo(CO)6 catalysts of
high catalytic activity the molecular hydrogen activated on the catalyst was supplied to the
radical directly and promptly even in the existence of the hydrogen donor solvent of the re-
dundancy as well. On the other hand, when large quantities of resolution radicals and aro-
matic compound existed in the system of reaction, it was found that the hydrogen from
tetralin mainly participates in the reaction than hydrogen activated on the catalyst.
Moreover, it was reported that the re-hydrogenation of the naphthalene formed by a dehy-
drogenation of the tetralin hardly happened.
A Mo(CO)6-H2S system catalyst was applicable in the solvent-free liquefaction of the
bituminous coal. Though Mo(CO)6 discomposed and became Mo carbide, which is not of
204 4 Liquefactionof Coal
high activity in the liquefaction of coal without a source of sulfur, the decomposition of
Mo(CO)6 was promoted, and MoS2 was formed, and coal liquefaction was promoted in the
hydrogen sulfide atmosphere. Thus, a solvent-free liquefaction of coal in an autoclave is
possible enough by using Mo(CO)6 and it is pointed out that hydrogen consumption speed
at the early stages of liquefaction is an important factor in liquefaction of coal (Warzinski
and Bockrath, 1996). It was also reported by Yoon et al. (1997) that a metal sulfide catalyst
of the high dispersion was formed, and that the amounts of hydrogen consumption in-
creased in order of Mo > Co > Fe, corresponding to the increase in the conversion and oil
yield when adding three-fold sulfur of the amount of stoichiometry to transition metal naph-
thenate (Mo, Co and Fe etc.). Burgess et al. (1996) and Song et al. (1997a) impregnated
coal to the water or the water/THF mixed solution of ATTM in order to form an active
phase-MoS2 in situ on the surface of the coal, and examined about the catalytic activity.
Moreover, the addition of water to the ATTM/coal system improved remarkably the con-
version and oil yield when liquefaction reacted by 2-step (350 ~ minutes and
400 ~ minutes) (Song et al., 1997b). This is because the surface area of MoS2 treated
with ATTM and water in 350 ~ became ca. 6 times in comparison with the surface area
of MoS2 treated with only ATTM.
On the other hand, Oshima et al. (1997) examined the influence of the shape of Mo cat-
alysts on coal liquefaction. The influence of a catalyst form (powder or grain-shaped) on
the internal diffuse in the pore depended on the kind of the solvent when a Co-Mo/A1203
catalyst was used. Further, no superior liquefaction yield was observed in the case impreg-
nating Mo onto coal than that obtained in the case dissolving an oil soluble Mo catalyst into
the solvent, suggesting that the Mo species supported on the coal in the former case was
eluted into the solvent, and was then sulfided and catalyzed the reaction just as a super-mi-
cro-particle catalyst of MoS2.
In order to develop practical processes of direct liquefaction of coal, it is important to
estimate the rates of hydrogen transfer during the liquefaction. A demonstration scale
process is now being developed in Japan. As the reactions in this process consist of both
hydrocracking by gaseous hydrogen and hydrogen transfer from donor solvents, the under-
standing of these reactions is regarded as significant for the process design.
Generally, coal liquefaction consists of processes in which coal is thermally hydroc-
racked and the produced radicals are hydrogenated by hydrogen atoms supplied from the
surroundings. Fig. 4.12 shows the hydrogen transfer model in which coal is liquefied in a
representative hydrogen donor solvent, tetralin, under pressurized hydrogen (Kabe, 1986).
In this scheme, hydrogen atoms are supplied to coal directly by tetralin (donation) and/or
gaseous hydrogen by the aid of the catalyst (spillover), or hydrogen atoms in coal are also
used (shuttling). Liquefied products are hydrocracked by gaseous hydrogen to give lighter
products on the catalyst. At the same time, naphthalene may be hydrogenated to tetralin on
the catalyst.
Many traditional approaches studying the product distribution under various experi-
mental conditions have been made (Moritomi, et al, 1983; Guin et al., 1979; Tsai and
Weller, 1979; Rottendorf and Wilson, 1979; Chow, 1981; Ueda et al., 1981) and a number
of attempts have been made to elucidate the mechanisms of hydrogen transfer by the use of
deuterium tracer, NMR and M.S. methods (Skowronski et al., 1984; Maekawa et al.,
1980). However, a few investigations using radioisotope tracer method have been made.
Poutsma et al. (1982b) performed the liquefaction of Illinois coal in a bench scale flow sys-
tem and investigated the behavior of tetralin during the reaction by a laC tracer technique.
They found that the grafting of tetralin to molecules derived from coal occurred to some ex-
4.3 Catalysis in Coal Liquefaction 205
3H-Donation / z - - ~ ~
H-Shuttling
3H
H-Spillover
ng
Rehydrogenation
Fig. 4.12 Hydrogentransfer model in coal liquefaction. [FromKabe. T., J. Jpn. Petrol. Inst., 39 345 (1986)]
tent. The 3H and 14C tracer methods allow a study of tritium incorporation by the addition
and exchange into the coal products from the gas phase as well as from the solvent (Kabe et
al., 1983a, 1983c, 1985). There are various interpretations for the effects of solvent and
catalyst on coal liquefaction. One of them is that the catalysts are effective in regenerating
donor solvents which are dehydrogenated during the liquefaction. From this point of view,
Moritomi et al. (1983) studied the hydrogen transfer at the initial stage of liquefaction.
Another interpretation that solvents are effective only to disperse coal and catalyst particles
and to dissolve gaseous hydrogen and then catalysts act to hydrogenate coal and coal prod-
ucts by the use of gas phase or dissolved hydrogen (Ueda et al., 1981). In order to elucidate
the behavior of hydrogen in coal liquefaction, the liquefaction of Taiheiyo coal by use of 3H
and 14C double-labeled toluene solvent (p-3H-toluene and methyl-lac-toluene), and 3H and
14C double-labeled tetralin solvent (3H labeled tetralin and a small amount of 1-14C-naph -
thalene or ~4C labeled tetralin) has been studied (Kabe, 1986).
Coal liquefaction was performed in a 350 ml autoclave containing 75 g of solvent, 25 g
of coal and 0 or 5 g of Ni-Mo-A1203 catalyst and filled with hydrogen at an initial pressure
of 0 or 5.9 MPa. It was heated at a heating rate of 10 ~ and held at 400 ~ during the
reaction. The reaction mixtures in the autoclave were separated by filtration, distillation
and extraction, as shown in Fig. 4.13. In this diagram, naphtha, light oil and SRC were the
distilled fractions under 200, 204-350 and over 350 ~ respectively. As for the solvent ex-
tractions, HS, BS and PS represent hexane, benzene and pyridine soluble products, respec-
tively; HIS, BIS and PIS represent the insoluble ones. SRC was separated into HS (oil).
BS-HIS (asphaltenes) and PS-BIS (preasphaltenes). The separated gas, coal products and
solvents were weighed and analyzed by GC. Naphthalene, decalin, methylindan and butyl-
benzene produced during coal liquefaction are thought to be converted from tetralin sol-
vent, and the recovered yields of the solvent containing them ranged from 97 to 98 wt%.
The specific activities of 3H and 14C in the reaction products were measured with a liquid
scintillation counter. Colorless or light colored products were directly dissolved into the
scintillator, while the colored liquid, solid and gas were oxidized to H20 and CO2 to avoid
206 4 Liquefaction of Coal
color quenching.
Figure 4.13 shows the separation process of products; material balances of products, 3H
and 14C were obtained in the flow of separation process as shown in Fig. 4.13. Specific ra-
dioactivities of 3H and ~4C of solvents and coal products in Fig. 4.13 suggest that 3H was
transferred from tetralin to the reaction products, but, except tetralin, there was little trans-
fer of 14C during the liquefaction. The 14C transferred to the liquefaction products was only
1.6% of the total ~4C, even in another experiment at 440 ~ thus it is assumed that incorpo-
ration of the solvent molecule to coal products scarcely occurs. The yield of hydrogenation
of naphthalene can be evaluated using the amount of 14C transferred from naphthalene to
tetralin. The amount of 14C which was contained in tetralin after the liquefaction with the
catalyst at 400 ~ for 30 min was about 1.2% of total 14C, and the remaining ~4C was left in
naphthalene. The re-hydrogenation of naphthalene to tetralin, therefore, was slight during
the primary liquefaction step within 30 min, and the hydrogenation-dehydrogenation reac-
tions between tetralin and naphthalene was also slight. When the liquefaction was conduct-
ed at 440 ~ and for 120 min, the amount of 14C transfer was much larger (18%) (Kabe et
al., 1983a), suggesting faster re-hydrogenation of naphthalene to tetralin at 440 ~ than at
400 ~
The hydrogen addition and exchange reactions between tetralin and gaseous hydrogen
without coal were also examined using 3H-labeled tetralin solvent and unlabeled gaseous
hydrogen under the same liquefaction conditions. In the absence of coal, 12% of the
tetralin was hydrogenated to decalin by gaseous hydrogen and 3% tetralin was converted
into naphthalene at 400 ~ and 30 min. The observed 3H distribution, 84.4% in solvent and
15.6% in gas, agreed with the calculated value, 16.0% assuming the complete scrambling of
hydrogen atoms between solvents and gaseous hydrogen. When the same experiment was
conducted without a catalyst, neither hydrogen addition to a solvent from gas phase n o r 3H
exchange was observed. Under the experimental conditions at 400 ~ and 30 min, there-
fore, the disproportionation of tetralin to naphthalene and decalin is less than 3% of tetralin
in the presence of the catalyst and 0% in the absence of catalyst.
Figure 4.14 shows the effects of reaction time and the catalyst on the product distribu-
tions for the experiments using 3H labeled tetralin and unlabeled gaseous hydrogen. The
product distributions indicate that the catalyst enhances the hydrocracking of preas-
phaltenes to oil and asphaltenes. Fig. 4.15 shows the amounts of 3H transferred to the same
products as in Fig. 4.14. The amount of tritium in gaseous hydrogen is significantly smaller
than the equilibrium value, 13%, of tritium exchange among gaseous hydrogen, coal and
solvent. Moreover, even in the presence of the catalyst, this value does not increase in the
reaction, thus coal or coal liquids may inhibit the hydrogen exchange reaction between
gaseous hydrogen and tetralin on the catalyst. Fig. 4.16 shows the concentrations of naph-
thalene produced during coal liquefactions in the presence and absence of the catalyst. In
the presence of the catalyst, the concentrations of naphthalene formed in only tetralin or in
the presence of coal products (HS and BS-HIS) are also plotted in Fig. 4.16. In the presence
of coal without the catalyst, the concentration of naphthalene increased dramatically. When
the catalyst was introduced, however, the concentration of naphthalene was always lower
than that without catalyst. On the other hand, when the experiment was conducted without
both gaseous hydrogen and catalyst, the amount of residue was the same as that in the pres-
ence of gaseous hydrogen without catalyst (Kabe et al., 1983c). The results suggest that the
hydrogen atoms used in liquefaction are largely supplied by tetralin solvent in the primary
liquefaction of coal in the presence or absence of the catalyst, but the supply of hydrogen
Without catalyst
With catalyst
~., 1 0 0 ~
lOOi x • x x x
~ ~ ~ Naphtha
~, Oil
O
~ 6o ~, 60
4o) 40
-.~
~ 2o
I
0
i
30
i
60 90
Nominal reaction time (min)
120 300 0
I
30
I
60
I
90
Reaction time (min)
120
i
Fig. 4.14 Effect of reaction time on the product distribution for Taiheiyo coal liquefaction at 400 ~ and under
initial pressure of H2 of 5.9 MPa. [From Kabe. T., J. Jpn. Petrol. Inst., 39 347 (1986)]
208
Without catalyst
1O0 With catalyst
10C
90 o ....
80 9C I o ~
-,~
5
...
a
0 30 60 90 120 300 0 30 60 90 120
Reaction time (min) Reaction time (min)
151 !
~=,~ l O ~
-S [ []
Z
5,-
I I I I
30 60 90 120
Reaction time (min)
atoms from tetralin is depressed in the hydrocracking steps of HS and BS-HIS as shown in
Fig. 4.16. This means that the hydrogen molecules dissociated on the catalyst are used in
the secondary hydrocracking of coal liquids.
In addition, the concentrations of decalin remained constant throughout the reaction
time, but the concentrations of methylindan and butylbenzene increased gradually, regard-
less of the presence or absence of the catalyst. This indicates that decalin is an impurity in
tetralin; methylindan and butylbenzene are produced from tetralin during the reaction.
Figure 4.17 shows the product distributions of the secondary hydrocracking of the pri-
mary coal liquids: HS, BS-HIS, PS-BIS and unseparated SRC, after the reaction for 120
min at 400 ~ in tetralin under pressurized H2 and in the presence of the catalyst. HS and
BS-HIS mainly produce light oil and HS, respectively. It is shown that in the presence of
BS-HIS, the hydrocracking of HS to light oil is somewhat depressed. PS-BIS is hydroc-
racked to produce BS. From another experiment in which the reaction time was varied, it
was shown that the amounts of HS and light oil remained almost constant during the reac-
tion of PS-BIS and, at the same time, PS-BIS decreased and BS-HIS was produced at a
constant rate. From these results, it is assumed that the presence of PS-BIS depresses fur-
ther hydrocracking of BS-HIS or HS (Kabe et al., 1984a).
The components of the source SRC are shown in Fig. 4.17. Calculation can be made to
elucidate what composition should be given when each component of SRC is hydrocracked
independently and put together in proportion to the source composition. The calculated re-
sult is shown by SRCcalc and compared with the result of the hydrocracking of the unsepa-
rated SRC (source SRC). A comparison shows the following results: (1) The presence of
PS-BIS in the system largely depressed the hydrocracking of BS-HIS to HS, and (2) the hy-
drocracking of PS-BIS takes precedence over that of BS-HIS. These results show that the
heavier coal products are more adsorbable on the catalyst and have the priority to react
there.
0 50 100
i I I I I I I 9 I I I I i
.s I I
BS
I I
BS
I I I I
HS
I I I 'iL
BIS i BIS i BS i HS ill
L N
SRC
BIS
IL
I
BS i HS fllL N
SRCcalc li iI BS HS i L iI N
BIS
S~SRC I BIS I BS I I
HS
I I
Figure 4.18 shows the distributions of products and 3H after the liquefaction in the sys-
tem of 3H-labeled gaseous hydrogen and unlabeled tetralin at 400 ~ for 30 min with or
without a catalyst. In the presence of the catalyst, the amounts of 3H incorporation into liq-
uid products are larger than in the absence of catalyst. These increases in incorporation are
due to the hydrocracking of the liquid products using gaseous hydrogen and the hydrogen
exchange between gaseous hydrogen and coal products.
The effect of solvents during the liquefaction was also examined using toluene and
naphthalene solvents in 3H-labeled gaseous hydrogen. The distributions of products and 3H
are shown in Figs. 4.19 and 4.20. Toluene was used as the sample of a solvent which is
less active as a hydrogen donor, and naphthalene was used as the sample which can change
to hydrogen donor, namely tetralin.
Comparing Figs. 4.19 and 4.20 with Fig. 4.18, toluene and naphthalene are shown to be
inferior solvents for the liquefaction without catalyst, but they give good results with a cata-
lyst. In the case of toluene solvent, the amount of 3H transferred to solvent is the least
among the three solvents, but 3H transfer to coal products in the presence of a catalyst is the
largest. In the case of naphthalene solvent, the amount of 3H transferred to solvent is larger
compared with the case of tetralin solvent with the catalyst. It is concluded that, in toluene
solvent, hydrogen atoms used for the coal liquefaction are supplied from the gas phase di-
rectly, but in naphthalene solvent, they are supplied from the gas phase partly through the
solvent. It is assumed that tetralin produced by the re-hydrogenation of naphthalene is de-
hydrogenated immediately by coal and converts to 3H-containing naphthalene. The hydro-
gen transfer cycle of naphthalene --~ tetralin ~ naphthalene is very effective in the lique-
faction in naphthalene solvent. That is why 3H content in naphthalene is high, compared
with the case of tetralin solvent and also why the 3H contents in coal products are less than
in the case of toluene solvent. These results agree with those of Moritomi et al. (1983) and
other workers (Guin et al., 1979; Tsai and Weller, 1979; Rottendorf and Wilson, 1979;
Chow, 1981)
Products %
Catalyst 0 10 20 30 40 50 60 70 80 90 100
Ni-Mo-AI203
None
3H
Ni-Mo-A1203 /
None
m T
LN
i
Residue: ~ PS-BIS: ~ BS-HIS: ~ HS:
Light oil: ~ Naphtha: ~ Gas:
Fig. 4.18 Products and 3H distributions in coal liquefaction in tetralin solvent at 400 ~ under initial pressure of
3H labeled H2 of 5.9 MPa for 30 min. [From Kabe. T., J. Jpn. Petrol. Inst., 39, 349 (1986)]
4.3 Catalysis in Coal Liquefaction 211
Products %
Catalyst 0 10 20 30 40 50 60 70 80 90 100
Ni-Mo-A1203
None
3H
Ni-Mo-A1203
None
Initial H2, 5.9 MPa; Reaction temperature and time, 400 ~ and 30 min.
Fig. 4.19 Products and 3H distributions in coal liquefaction in toluene solvent under 3H-labeled H2.
[From Kabe. T., J. Jpn. Petrol. Inst. 39, 350 (1986)]
Products %
Catalyst 0 10 20 30 40 50 60 70 80 90 100
Ni-Mo-A1203
None
3H
1 I m
Ni-Mo-A1203 N
None
Fig. 4.20 Products and 3H distributions in coal liquefaction in naphthalene solvent under 3H-labeled H2.
[From Kabe. T., J. Jpn. Petrol. Inst., 39, 350 (1986)]
The concentrations of 3H in coal products and solvents given are s h o w n in Fig. 4.21. In
the case of n a p h t h a l e n e solvent, 3H concentration in tetralin is the largest in the solvent
fractions and corresponds to the value which is given w h e n naphthalene is h y d r o g e n a t e d by
two 3H labeled h y d r o g e n molecules. On the other hand, 3H concentrations in u n c h a n g e d
solvents are small. This means that the h y d r o g e n e x c h a n g e b e t w e e n solvents and h y d r o g e n
m o l e c u l e s is m i n o r and that the a m o u n t of the h y d r o g e n e x c h a n g e is less than a few percent
of the solvent h y d r o g e n atoms even in the presence of the catalyst. H o w e v e r , 3H concen-
212 4 Liquefaction of Coal
25
20
• 15 ,':2
Z
~9 10
o
rj
trations in coal products are much larger and, in the case without catalyst, follows the order
toluene > naphthalene > tetralin, indicating the reverse order of the hydrogen donating
qualities of these solvents. The 3H concentrations in coal products follow the same order in
the presence of the catalyst. It suggests that, in toluene solvent, the liquefaction proceeds
with hydrogen spillovers from hydrogen adsorbed on the catalyst, but, in naphthalene sol-
vent, coal is liquefied by hydrogen donation from tetralin converted from naphthalene.
In the absence of a catalyst, 3H concentrations in heavier coal products are larger than
that in lighter coal products. This agrees with Skowronski's results (1984). The reason for
this is not clear but it seems that the heavier products have much more mobile hydrogen
atoms such as in OH or NH, which exchange with hydrogen molecules even without a cata-
lyst.
From the data above, the amount of hydrogen transferred by addition or exchange reac-
tions among gaseous hydrogen, solvent and coal products can be calculated. The calculated
results are shown schematically in Fig. 4.22, in which the solid and dotted arrows indicate
the directions of hydrogen addition and exchange among shown phases, respectively.
The data presented in Fig. 4.22 illustrate that the catalyst slightly decreases the hydro-
gen addition from solvent to coal but it considerably increases hydrogen addition from gas
phase to coal products. The latter increased with reaction time. The increase in hydrogen
consumption is assumed to be equal to the amount used for the hydrocracking of the coal
liquids, because the yields of the liquefaction, which are calculated from the amounts of un-
converted coal, are almost the same in both cases with and without the catalyst. When coal
and/or heavy coal liquids are present, however, the hydrogenation of solvents by gaseous
hydrogen is rather slight even in the presence of the catalyst compared with the case of hy-
drogenation of solvents in the absence of coal. In other words, the hydrogen exchange be-
tween gaseous hydrogen and solvents is suppressed since solvents cannot easily be ad-
sorbed on the catalyst in the presence of the heavier coal products. The hydrogen exchange
between gas phase and coal products, on the contrary, proceeds considerably even in the
absence of catalyst. This result means that there is much transferable hydrogen in the coal
4.3 Catalysis in Coal Liquefaction 213
products which are easy to exchange with gaseous hydrogen but are difficult to exchange
with hydrogen atoms in solvent molecules.
The mechanism of hydrogen transfer during coal liquefaction was determined using 3H
and 14C tracers and the following results were obtained.
(1) In the process of primary liquefaction, coal is thermally decomposed and hydrogenated
by the hydrogen atoms released from hydrogen donor solvent.
(2) In the process of the hydrocracking of primary coal liquids, the coal liquids are hydroc-
racked on the catalyst by hydrogen atoms supplied mainly from gaseous hydrogen.
(3) The hydrogen exchange between solvent and gaseous hydrogen during coal liquefaction
is little even in the presence of a catalyst.
(4) The adsorption of the heavier coal liquids are more easily on the catalyst surface, thus
the hydrocracking of these heavier products take precedence over that of the lighter prod-
ucts.
(5) In naphthalene solvent, only a small amount of coal is liquefied in the absence of cata-
lyst because hydrogen transfer from both solvent and gaseous hydrogen is difficult.
However, the liquefaction in naphthalene in the presence of the catalyst proceeds much
faster, because of the hydrogen transfer cycle of naphthalene -+ tetralin -+ naphthalene due
to the hydrogenation of naphthalene in the presence of a catalyst.
(6) The distinction between hydrogen addition and exchange among gas phase, solvent and
coal products can be made clear by these series of experimental methods using radioisotope
tracers.
In order to elucidate the nature of hydrogen transfer among gas, coal and solvent, 3H
and 14C tracer experiments were also carded out using Datong coal (Kabe et al., 1990c).
Fig. 4.23 shows the effect of liquefaction time on the conversion and product distribution
for Datong coal at 400 ~ in the absence and presence of Ni-Mo/A1203 catalyst. Clearly,
the initial liquefaction rate was rapid, and nearly one third of the coal was converted when
the autoclave was held at 400 ~ for only a few minutes, and the initial rate did not depend
on the catalyst. The conversion reached two thirds after 5 h with or without the catalyst.
214 4 Liquefaction of Coal
30
i
301 (b)
~ q
0
60 ~
r
~,~6 2 0 o o 20 0
4o ~
,.o
~ 10x ~ ~ 10
20 ~ ~ 20 ~
.o .~ 0
~ -----------D-
0 0 0
O 1 2 3 4 5 0 1 2 3 4 5
Reaction time (h) Reaction time (h)
Fig. 4.23 I-] Gas, X Naphtha, A Ligth oil, O HS, V HIS-BS, V BIS-THFS, 9 THFIS.
Changes in product distributions with reaction time for Datong coal liquefaction at 400 ~
a) without, and (b) with a catalyst.
[Reproduced with permission from Kabe. T. et el., Fuel Proces. Technol., 25, 48, Elsevier
(1990)]
However, the catalyst changed the product distribution by accelerating the hydrocracking of
preasphaltenes at first and then hydrogenate asphaltenes to oil. These results suggest that
the catalyst was not able to hydrogenate the coal directly but that the heavier products in
liquids are predominantly adsorbed on the catalyst surface and react with hydrogen. The
same results were observed for Taiheiyo and Wandoan coals (Kabe et al., 1986a, 1987a).
Besson et al. (1986) reported more detailed studies on the effects of Ni-Mo/A1203 and
FeeO3 catalysts on coal liquefaction at 400 ~ in tetralin solvent and reached a similar con-
clusion, i.e., that the catalysts had little effect on the conversion of the coal into THF-solu-
ble products but increased the amount of asphaltenes (toluene soluble products).
The hydrogen distributions among gas phase, solvent and coal products are shown in
Fig. 4.24. Without a catalyst, the hydrogen distribution does not change with reaction time.
When a catalyst was added, hydrogen and coal products increased and that of gas de-
creased. In order to clarify the amount of hydrogen transferred to coal products, the
amounts of hydrogen transferred from the gas phase and solvent to coal products were plot-
ted with elapse of time in Fig. 4.25. Without the catalyst, the amounts of hydrogen initially
transferred to 100 g of coal were 0.21 g from the gas phase and 0.42 g from the solvent, re-
spectively. The hydrogen transfer from solvent to coal was the main reaction in the initial
stage. The amount of hydrogen transferred gradually increased and reached 1.1 and 1.0 g
from the gas phase and the solvent at 5 h, respectively. With the catalyst, the amounts of
hydrogen transferred to coal products were 0.73 g from the gas phase and 0.28 g from the
solvent at 0 h. This indicates that hydrogen transfer from gaseous hydrogen to coal prod-
ucts was rapid, even at the initial stage of the reaction when the catalyst was used. The
amount of hydrogen transferred from gaseous hydrogen to coal products increased to 2.2 g
at 5 h, however, that from the solvent did not increase so much and was only 0.69 g at 5 h.
Since naphthalene produced by hydrogen transfer from tetralin to coal was re-hydrogenated
to tetralin on the catalyst, the amount of hydrogen transferred from the solvent apparently
did not increase. Total amounts of hydrogen transferred to coal were 2.1 and 2.9 g without
and with the catalyst, respectively. Fig. 4.24 also shows tritium distributions among gas
4.3 Catalysis in Coal Liquefaction 215
phase, s o l v e n t and coal products. W i t h o u t the catalyst, the tritium distribution s h o w s that
tritium transfers to coal p r o d u c t s in the initial stage, then tritium in s o l v e n t increases. This
suggests that, as a first step, gas p h a s e tritium transfers to coal f o l l o w e d by the e x c h a n g e
b e t w e e n coal and solvent. W h e n the c a t a l y s t w a s added, tritium w a s t r a n s f e r r e d to coal
m o r e r a p i d l y at the initial stage o f the reaction. T r i t i u m w a s also t r a n s f e r r e d to s o l v e n t at
the initial stage, in contrast to the case w i t h o u t the catalyst. A s s h o w n in Fig. 4.25, h y d r o -
= 60
80
A
a
- ~
L
'
-
(b)
a A
.o - .o
..a
j
.ID
~ 40 _ "~
c~
4 /
.,..~ .~
20
__ ' I I I I I h i I I I I I
0 1 2 3 4 5 0 1 2 3 4 5
Reaction time (h) Reaction time (h)
Fig. 4.24 Changes in balance of hydrogen and tritium in Datong coal liquefaction at 400 ~
(a) without and (b) with a catalyst.
Hydrogen distribution: D gas, A solvent, C) coal; Tritium distribution: 9 gas, A solvent,
9 coal. [Reproduced with permission from Kabe. T. et el., Fuel Process. Technol., 25, 49
(1990)]
2
o
c..)
e~0
.,..~
~D
o ( v
~__A____._--~ -------~-
01 I I I I I
0 1 2 3 4 5
Reaction time (h)
Fig. 4.25 Changes in amounts of hydrogen addition with reaction time in Datong coal liquefaction at 400 ~
H added to coal from gas phese: C) without a catalyst, @ with a catalyst;
H added through tetralin: A without a catalyst, A with a catalyst.
[Reproduced with permission from Kabe. T. et el., Fuel Process. Technol., 25, 49, Elsevier (1990)]
216 4 Liquefaction of Coal
O o
• (a) x (b)
20 .~ 20
300 ~ 300min 440 ~ (0-'- 300 min)
440 ~ 180min |
.o o
230 ~ 120min 9
10 ~ 10-
o
E
I I I
0 20 40 60 80 100 o 20 40 60 80 100
Conversion (%) Conversion (%)
gen was transferred to coal from tetralin to produce naphthalene at the initial stage, and tri-
tium was transferred to the solvent through hydrogenation of naphthalene on the catalyst.
From Fig. 4.24 (b), after 5 h, the tritium distribution among gas phase, solvent and coal
products approached the hydrogen distribution. This indicates that hydrogen exchange
reaches the equilibrium among gas phase, solvent and coal products at 5 h when the catalyst
is added. Fig. 4.26 shows the change in tritium concentrations of each liquefied product
and the solvent with conversion. The experimental points in Fig. 4.26 correspond to those
in Fig. 4.23 (400 ~ 0-5 h), unless otherwise noted. The points with special notation in
Fig. 4.26 show the results at different temperatures. The two horizontal dotted lines in Fig.
4.26 indicate the expected mean equilibrium tritium concentrations for coal, representing
the radioactivity in one gram of coal at the equilibrium in the hydrogen exchange reaction.
If the equilibrium among coal, solvent and gas phase (lower line), and between coal and gas
phase only (upper line), were established in the hydrogen exchange reaction, the equilibri-
um tritium concentrations for coal would be 5 X 103 and 19 X 103 dpm/g-coal, respectively.
The expected concentration for the solvent under equilibrium conditions among the three
components in the system is shown by an arrow on the fight-hand side of the graphs (10 X
103 dpm/g tetralin). Without the catalyst, tritium concentrations in coal products increased
with the conversion of coal. Since tritium is transferred from the gas phase to the solvent
via coal in the absence of the catalyst, the tritium concentration in the solvent began to in-
crease after about 30 wt% of coal was converted. On the other hand, when the catalyst was
added, a large number of hydrogen in coal components exchanged with gaseous hydrogen
at the initial stage of reaction. Hydrogen exchange reached equilibrium among gas phase,
solvent and coal products at the final stage of the reaction consistent with the results given
in Fig. 4.24.
Figure 4.27 shows the product distribution as a function of temperature at a reaction
time of 2 h without the catalyst. In the temperature range 300-440 ~ the conversion of
coal increases with increasing temperature. The hydrocracking to lighter products has little
chance of occurring below 300 ~ Fig. 4.28 shows the amounts of hydrogen transferred
and exchanged at 300, 350 and 400 ~ in the absence of catalyst. The solid and dotted ar-
rows indicate the directions of hydrogen transfer and exchange among shown phases, re-
spectively. Numbers with the arrows indicate the amounts (g) per coal (100 g) of the trans-
4.3 Catalysis in Coal Liquefaction 217
ferred or exchanged hydrogen for 2 h (Kabe et al., 1990c). Since the direct hydrogen ex-
change between gas phase and solvent hardly occurred in the absence of coal and catalyst,
it was ignored in Fig. 4.28.
The amounts of hydrogen transferred and exchanged increased with rise in temperature.
The amount of hydrogen transferred from the gas phase to coal doubled with a rise from
300 to 350 ~ However, the amount of hydrogen transferred from solvent to coal remark-
ably increased with a rise from 350 to 400 ~ These results indicate that gaseous hydrogen
is the main hydrogen donor in coal liquefaction at 350 ~ and that the capability of tetralin
as a hydrogen donor appeared at 400 ~ Since hydrogen in tetralin transfers to coal more
rapidly than gaseous hydrogen at 400 ~ only a slight amount of hydrogen transferred
from gas phase to coal would increase with a rise from 350 to 400 ~
Figure 4.29 shows the relationship between the tritium concentrations of the coal prod-
ucts and the reaction time at 300 ~ In all cases at 300 ~ in the uncatalyzed experiments,
50 100
40 80
0 0
0 0
n~
0 0
30 60
,d
e~O
.,..~
o 20 4o
9
.,...
,.o
lO 20 m
0 I
200 300 400
Temperature (~
9 Residue, 9 Preasphaltene, A Asphaltene, O Oil
Fig. 4.27 Effect of temperature on the fractional weights of products. [Reproduced with permission
from Kabe. T. et el., Fuel Process. Technol., 25, 51, Elsevier (1990)]
Reaction temperature
300 ~ 350 ~ 400 ~
Gas phase Gas phase Gas phase
4 V
o
~7
2
O
8
Q
L)
I I 1
0 2 4 6
Reaction time (h)
9 Residue, V THFS-BIS, V BS-HIS, O HS, [] Tetralin
Fig. 4.29 Tritium concentrations in coal products and tetralin in uncatalyzed reaction of Datong coal at 300 ~
[Reproduced with permission from Kabe. T. et el., Fuel Process. Technol., 25, 52, Elsevier (1990)]
Table 4.5 Tritium Concentrations and Material Balances in Datong Coal Liquefactiona at
300 ~ for 6h
Tritium Tritium Amount of Recovered
concentration balance product ratio
(dpm/g) (%) (g) (%)
Coal product -- 7.47 24.60
Residue 2767 6.65 21.94 85.30
Preasphaltene 3443 0.31 0.83 3.23
Asphaltene 2144 0.26 1.12 4.35
Oil 1770 0.09 0.45 1.75
Light oil 5256 0.09 0.15 0.80
Naphtha 5993 0.07 0.10 0.39
Gas 0 0.00 0.01 0.04
Solvent m 2.68 71.61
Tetralin 210 2.23 70.31 93.18
Naphthalene 4116 0.44 0.96 1.34
Methylindan 210 0.00 0.10 0.14
Decalin 210 0.01 0.24 0.34
Gas 89.83
alnitial conditions: Coal 25.05 g, Tetralin 75.01 g, and H2 1.25 g. [Reproduced with
permission from Kabe. T. et el., Fuel Process. Technol., 25, 53, Elsevier (1990)]
4.4 HydrogenTransferReaction in Coal Liquefaction 219
terial balances in Datong coal liquefaction at 300 ~ and a reaction time of 6 h. From these
data and the hydrogen content of the residue (4.6 wt%), the ratio of exchangeable hydrogen
in the residue hydrogen can be calculated and the exchanged hydrogen at 6 h (approximate-
ly equilibrium value) amounts to 7.8 atom% of hydrogen in the residue. This value may
suggest the ratio of the active hydrogen such as in the form of-OH and -NH to total hydro-
gen in coal. This shows which hydrogen in coal is exchanged. Since oil, asphaltenes and a
part of preasphaltenes are dissolved in the solvent, these are expected to undergo more
rapid hydrogen exchange than the insoluble carbonaceous materials. However, the results
are different. This may suggest that there is more mobile hydrogen in residue. Further, at
300 ~ only the active hydrogen in coal products is exchanged by hydrogen. Although a
detailed analysis of phenolic OH group of Datong coal has not been done, a comparable
analysis of bituminous coals has been reported (Pestryakoc, 1986; Maekawa, 1975).
Bituminous coals which have a chemical composition of C: 75-85% and H: 5.0-5.4 wt%
contain 6-12 atom% of phenolic OH hydrogen for total hydrogen. Kotanigawa et al.
(1979) concluded that the exchange reaction between deuterium gas and phenol took place
rapidly at 350 ~ with ZnO-Fe203 catalyst. OH hydrogen of polycondensed aromatic phe-
nolics is able to exchange at lower temperatures.
1971). It has been recognized that the exchangeability of a compound is strongly relative to
its molecular structure, and further to its capacity as a hydrogen donor or acceptor. Based
on this viewpoint, it is possible to estimate the structural feature of coals by the investiga-
tion of hydrogen exchange. In the studies, the isotope tracer method has been extensively
utilized as an effective means as well as model compounds. For example, in coal liquefac-
tion, this method makes it possible to determine the amount and the structural positions of
hydrogen in coal reacting with hydrogen during liquefaction by labeling reactive sites with
deuterium or tritium. Most of these investigations were performed using deuterium tracer.
Several research groups have used deuterium to investigate the mechanism of coal hydro-
genations and the reaction of coal-related model compounds such as tetralin since 1967 (Fu
and Blaustein, 1967; Franz and Camaioni, 1980; Brower, 1982; Schweighardt et al., 1976;
Cronauer et al., 1982; Skowronski et al., 1984; King and Stock, 1982; Collin and Wilson,
1983). In these researches, intensive effort has been made to obtain a better understanding
of the coal hydroliquefaction mechanism. Such knowledge may lead to the improvement of
coal utilization and coal hydroliquefaction efficiency by elucidating hydrogenation rates
and mechanisms as well as the sites of hydrogen incorporation from the gas and solvent
phases. However, because of the low solubility of coal in solvents and the lack of quantita-
tive data from 2H-NMR, it is difficult in these studies to conduct quantitative analysis of hy-
drogen transfer among the gas phase, solvent and coal.
The investigation of coal liquefaction mechanisms using radioactive tritium as a tracer
has started in recent years. The representative investigations are the quantitative estimation
of hydrogen mobility in the systems consisting of coal and donor solvent, coal and gas
phase hydrogen, and coal and water, reported by Kabe and coworkers (Kabe, 1984, 1986,
1988; Kabe et al., 1983a-c, 1984, 1985a-d, 1986a-c, 1987a-c, 1989a-b, 1990a-e,
1991a-c; Ishihara et al., 1993, 1994, 1995; Yamamoto et al., 1987). The results have
shown that tritium tracer methods have several distinct advantages over the deuterium trac-
er methods, especially for a complicated reaction system. These reports indicate that the
hydrogen mobility of coal and coal-related compounds can be quantitatively analyzed using
the hydrogen exchange reactions occurring between coal, the gas phase and the tetralin sol-
vent, as well as by considering the hydrogen addition reactions.
4.4.2 Behavior of Hydrogen in Coal Liquefaction
The use of catalytic hydrotreatment is an important aspect of the process of Wandoan coal
liquefaction now being developed in Japan. Although the reaction conditions and the na-
ture of the products may present a number of problems in the development of the process,
the fundamental reactions of coal liquefaction in the initial reaction stage are the thermal
decompositions of coal structure under hydrogen atmosphere in the presence of a donor sol-
vent (Curran et al., 1967; Neavel, 1976; Derbyshire and Whitehurst, 1981). When there are
hydrogen atoms, which can stabilize coal radicals made by thermal decomposition, the coal
is converted to coal-derived liquids. If no hydrogen is available, however, the free radicals
recombine to form heavier products (Ohe et al., 1985). Therefore, an understanding of the
hydrogen transfer mechanism during liquefaction is essential for process design.
The liquefaction of Wandoan coal in 3H- and in 14C-labeled solvent was studied (Kabe
et al., 1987b). It was reported that tetralin solvent provided the coal with hydrogen atoms,
and that the presence of a catalyst decreased the addition of hydrogen atoms from solvent to
coal and increased the addition from gaseous hydrogen to coal. Since gaseous hydrogen
was not traced, however, the hydrogen transfer path from the gas phase to coal remains un-
clear.
4.4 Hydrogen Transfer Reaction in Coal Liquefaction 221
coal products. In naphthalene solvent, the degree of liquefaction was obviously low with-
out catalyst (Run 3). However, liquefaction proceeds at a substantial rate in the presence of
the catalyst (Run 4). This suggests that gaseous hydrogen is used for liquefaction in naph-
thalene solvent in the presence of a catalyst.
On the other hand, when decalin is used as the solvent in the presence of a catalyst
(Run 6), the amount, of residue is almost the same as for naphthalene solvent but the prod-
ucts are lighter. Without the catalyst (Run 5), liquefaction in decalin proceeds more exten-
sively than in naphthalene. The amounts of tetralin and naphthalene derived from decalin
show that liquefaction proceeds to a considerable extent accompanying the hydrogen dona-
tion from decalin in the absence of a catalyst.
The last columns (Run 8) of Tables 4.6 and 4.7 show the distributions obtained experi-
mentally without the coal and coal products. The experimental 3H distribution, 83% in the
solvent and 17% in the gas phase, agreed with the calculated values based on the assump-
tion of a complete scrambling of hydrogen atoms of the solvent and molecular hydrogen.
These results show that if coal or coal products are not present, the hydrogenation of
tetralin to decalin and the hydrogen exchange between solvent and molecular hydrogen will
proceed rapidly in the presence of a catalyst. On the other hand, no hydrogenation of sol-
vent or hydrogen exchange occurs in the absence of catalyst (Run 7).
Figure 4.30 shows 3H concentrations in liquefied products and in the solvent. It shows
that 3H concentrations in coal products produced in tetralin and decalin solvents are lower
than those in naphthalene solvent in the absence of a catalyst. But in the presence of a cata-
lyst, the amount of 3H from the gas phase incorporated into coal products is largest in de-
calin solvent except for light oil. It shows that decalin is a good hydrogen donor without a
catalyst, but molecular hydrogen is a better hydrogen donor than decalin when a catalyst is
present. With a catalyst, naphthalene has a similar action to tetralin. In Fig. 4.30, a fairly
low concentration of 3H in each solvent shows that the hydrogen exchange between solvent
and molecular hydrogen is small in the presence of coal and coal products. On the other
hand, the 3H concentration in tetralin converted from naphthalene is high because tetralin
molecules are formed by hydrogenation of naphthalene using molecular hydrogen. The
value of the 3H concentration of tetralin converted from naphthalene solvent was equal to
the value calculated under the assumption that four hydrogen atoms from the gas phase are
added to one naphthalene molecule. The 3H concentration of decalin fraction converted
from tetralin and that of naphthalene converted from tetralin were the same as the 3H con-
centration in tetralin solvent itself, in the presence of the catalyst.
Comparing the uncatalyzed experiments in the three solvents shown in Table 4.6, the
degree of liquefaction, determined from the amount of residue, decreases in the order
tetralin > decalin > naphthalene. This order conforms to that of the hydrogen donating
power of the solvents. However, the order is tetralin > naphthalene -- decalin in the pres-
ence of the catalyst. This shows that the hydrogen donating cycle, naphthalene ~ tetralin
---) naphthalene in naphthalene solvent, is as effective for coal liquefaction as hydrogenation
in decalin solvent in the presence of a catalyst.
Table 4.7 shows that the amounts of 3H in coal products in uncatalyzed experiments
(Runs 1 and 5) are almost the same in both tetralin and decalin solvents. In Run 3 in naph-
thalene without catalyst, the 3H content in the residue is higher than that in tetralin or in de-
calin solvent. In the absence of a catalyst, Fig. 4.30 shows that the amount of 3H incorpo-
rated into the products increases in the order oil < asphaltenes < preasphaltenes < residue,
irrespective of the solvent used. These results agreed with those obtained by Skowronski et
al. (1984) in a coal-deuterium gas system. The amount of 3H transfer from gaseous hydro-
4.4 Hydrogen Transfer Reaction in Coal Liquefaction 223
25
20 _.9.
~ 15 ~ ~Z
_ _
10
0 I I ;L
Residue PA A O LO N T D
Fig. 4.30 Effects of solvents and catalyst on tritium counts of products for Wandoan coal liquefaction at 400 ~
for 30 min. PA: Preasphaltene; A: Asphaltene; O: Oil; LO: Light oil; N: Naphthalene; T: Tetralin; D:
Decalin; 1-Methyl-indan and butylbenzene
[Reproduced with permission from Kabe. T. et el., Fuel, 66, 1327, Elsevier (1987)]
gen to coal products increases substantially in the presence of a catalyst. Fig. 4.30 also
shows that 3H concentration of coal products is almost the same as that in tetralin and naph-
thalene solvents in the presence of a catalyst, and it seems to show that, in naphthalene sol-
vent, a fairly large part of the hydroliquefaction was conducted by tetralin which was pro-
duced from naphthalene. On the other hand, 3H concentrations in coal products produced in
decalin solvent are higher than those in other solvents in the presence of a catalyst. This in-
dicates that direct hydrogenation of coal by gaseous hydrogen and the hydrogen exchange
between hydrogen molecules and coal components are enhanced in decalin solvent. This
also suggests that it is energetically more favorable for liquefied products to react with hy-
drogen dissociated on the catalyst than to be hydrogenated by the decalin itself. From these
results it is concluded that naphthalene behaves as a hydrogen carrier from the gas phase to
coal by being hydrogenated to tetralin with the help of the catalyst.
Here the route of hydrogen incorporation from solvents and gaseous hydrogen to coal
products is discussed. Hydrogen incorporation during coal liquefaction involves two reac-
tions, i.e., hydrogen addition to coal products and hydrogen exchange among the coal prod-
ucts, the solvent and the gas phase. To clarify the correlation of these reactions, the author
and his coworkers attempted to calculate and differentiate these two kinds of hydrogen in-
corporation using the experimental data of Tables 4.6 and 4.7. The results calculated for
Runs 1 to 6 are shown schematically in Fig. 4.31 in which, the solid arrows show the direc-
tions of the hydrogen addition and the dotted arrows represent those of hydrogen exchange
between the phases shown. The numbers with the arrows indicate the amounts (g) per coal
(100 g) of the added or exchanged hydrogen.
Referring to Fig. 4.31, when the catalyst is not used, coal liquefaction proceeds by hy-
drogen addition mainly from the solvent, except for the case of naphthalene solvent, and the
gaseous hydrogen can exchange only with coal products. The catalyst enhances direct hy-
drogen addition from the gas phase to coal products and decreases the amount of hydrogen
224 4 Liquefaction of Coal
made to elucidate the reactivity of tetralin in the presence of coal, little is known about the
behavior of tetralin itself, especially the hydrogen mobility in it under coal liquefaction
conditions. The reaction of tetralin with tritiated hydrogen molecules to estimate the hydro-
gen mobility of tetralin quantitatively using a tritium tracer method is discussed below.
The reaction of tetralin with tritiated hydrogen molecule was performed at 350-400 ~
and the results are shown in Fig. 4.32a (Kabe et al., 1991a). In this temperature range, the
products were 1-methylindan by isomerization, naphthalene by dehydrogenation, and n-
butylbenzene by hydrocracking; the main product was 1-methylindan. With increase in
temperature, the concentration of tetralin decreased and the concentrations of the products
increased. Decalin by disproportionation was not produced at 350--400 ~ This is consis-
tent with Hooper's result, in which the disproportionation of tetralin to decalin and naphtha-
lene does not occur in the absence of coal (Hooper et al., 1979). When the reaction temper-
ature increased to 440 ~ the yields of 1-methylindan, naphthalene and n-butylbenzene in-
creased to 11.6, 4.8 and 5.2%, respectively, and the increase in 1-methylindan was most re-
markable. That the yield of 1-methylindan remarkably increased with a rise from 400 to
~. 1
O
.,..~
0
350 375 400
Reaction temperature (~
~. 4 -
~O 3
O
0 i i
0 120 240 360 480
Reaction time (min)
Fig. 4.32 Effectof reaction temperature and reaction time on product yields in the reaction of tetralin with
gaseous hydrogen (Tetralin: 75 g). /~ Naphthalene; D Methylindan; V Butylbenzene
[From Kabe. T. et al., Ind. Eng. Chem. Res., 30, 1756, (1991)]
226 4 Liquefaction of Coal
440 ~ is also consistent with Hooper's report. Fig. 4.32b shows the effect of reaction time
on the yields of 1-methylindan, naphthalene, and n-butylbenzene. These products increased
with time. When the reaction time was prolonged from 300 to 480 min, yields of products,
especially 1-methylindan, remarkably increased.
Tritium in the gas phase was introduced into tetralin; and the amount introduced was
estimated by the hydrogen exchange ratio (Kabe et al., 1991a). Fig. 4.33 shows the change
in both the hydrogen exchange ratio and the conversion of tetralin with reaction time and
temperature. At 350 ~ the hydrogen exchange ratio was 0.3% even at 300 min. The con-
version of tetralin was also rather low. At 375 and 400 ~ both the hydrogen exchange ra-
tio (open symbol) and the conversion of tetralin (closed symbol) increased with time and
temperature. Some relationship seems to exist between the hydrogen exchange and the
conversion of tetralin. When the reaction time was prolonged to 480 min, the hydrogen ex-
change ratio and conversion of tetralin at 400 ~ were 13.5 and 10.1%, respectively. When
the reaction temperature increased to 440 ~ the hydrogen exchange ratio and conversion
of tetralin at 120 min were 14.3 and 21.7%, respectively. It was reported that, in coal hy-
drogenation, the hydrogen exchange ratio of coal increased remarkably with a rise from 350
to 400 ~ to reach more than 40% at 400 ~ 120 min (Kabe et al., 1990b). The present re-
suits indicate that simple tetralin itself is more difficult to exchange with hydrogen mole-
cules in gas phase than coal.
To estimate the relationship between the hydrogen exchange ratio and conversion of
tetralin, the values of the amount of exchanged hydrogen per amount of converted tetralin
(g-atom/mol) at each temperature were plotted against reaction time in Fig. 4.34. Values at
375 and 400 ~ increased with time and gave a linear relationship. However, the values at
375 ~ were larger than those at 400 ~ at each reaction time. This shows that the hydro-
gen exchange between tetralin and molecular hydrogen strongly depends on temperature,
although the hydrogen exchange may occur at the time when tetralin converts (vide infra).
Further, the values at 375 ~ for 300 min and 400 ~ for 300 min were 25.3 and 16.5 g-
atom/mol, respectively, more than 12 g-atom in 1 mol of tetralin. This shows that tritium is
introduced into not only converted tetralin (products), but also into remaining tetralin.
These results indicate that there is an intermediate such as tetralyl radical which converts to
a product or returns to tetralin and further can exchange with hydrogen molecules in those
routes as shown in Eq. (4.34). On the other hand, the value of AEH/ACT at 350 ~ was
somewhat constant. The mechanism at 350 ~ may be different from that at 375 and 400
~
Table 4.8 shows effects of the amount of tetralin and hydrogen pressure on the hydro-
gen exchange ratio, the conversion of tetralin, and the product distribution at 400 ~ for
120 min. The hydrogen exchange ratio and the amount of converted tetralin were plotted
against the value of hydrogen per tetralin (mol/mol) in Fig. 4.35. The plots of the hydrogen
exchange ratio showed approximately a straight line and increased in proportion to hydro-
gen/tetralin values, while the plot of the amount of exchanged hydrogen showed some scat-
ter. The amount of exchanged hydrogen showed a maximum, about 0.3 g at 30 or 50 g of
tetralin, 60 kg/cm 2. The plots of the amount of converted tetralin also showed a sure
straight line and increased with increase in hydrogen/tetralin.
The amounts of products formed were plotted against hydrogen/tetralin in Fig. 4.36, 1-
methylindan showed a straight line and increased with increase in hydrogen/tetralin. On
4.4 Hydrogen Transfer Reaction in Coal Liquefaction 227
6 -
= 4-
0 o0o.o "~176176
0
0 120 240 360
Reaction time (min)
Fig. 4.33 Effect of reaction temperature and reaction time on hydrogen exchange ratio and conversion of tetralin
(tetralin: 75 g). Hydrogen exchange ratio: O 400 ~ E] 375 ~ 350 ~ Conversion of tetralin: 9
400 ~ 9 375 ~ 9 350 ~ [From Kabe. T. et al.,Ind. Eng. Chem. Res., 30, 1756 (1991)]
30
20
O
d~
[-,
<
10
<
0 i i i
0 120 240 360
Reaction time (min)
Fig. 4.34 Effect of reaction temperature and reaction time on ratio of the amount of exchanged hydrogen (AEH)
to the amount of converted tetralin (ACT). O 400 ~ IS] 375 ~ 350 ~
[From Kabe. T. et al., Ind. Eng. Chem. Res., 30, 1757 (1991)]
the other hand, although 2.3 g of naphthalene was produced in the absence of hydrogen as
shown in Table 4.8 and Fig. 4.36, the amount of naphthalene formed was inhibited by the
presence of hydrogen and showed the tendency to decrease slightly with increase in a hy-
drogen/tetralin molar ratio. The amount of n-butylbenzene formed also increased with an
increase in hydrogen/tetralin. These results mean that the hydrogen-exchange ratio, the
amount of methylindan formed, and the amount of converted tetralin increase with increase
in hydrogen pressure. It is important that the gaseous hydrogen participates in the isomer-
ization which does not accompany the income and the outgo of hydrogen. This may be one
228 4 Liquefaction of Coal
a Reaction temperature, 400 ~ Reaction time, 120 min. briER = H~,drogen exchange ratio, c Np, Naphthalene;
MI, Methylindan; BB, Butylbenzene. dNitrogen atmosphere (1 kg/cm~). [From Kabe. T. et al., Ind. Eng. Chem.
Res., 30, 1757, (1991)]
15 5
,
O
-4 ~0
10-
-3 "~
~
~ - 2 ~
~ - 1 <E
0 t t i i 0
0 1 2 3 4
Hydrogen / Tetralin (mol/mol)
Fig. 4.35 Effect of molar ratio of hydrogen to tetralin on hydrogen exchange ratio and the amount of converted
tetralin at 400 ~ for 120 min.
O Hydrogen exchange ratio; 9 Amount of converted tetralin; A Amount of exchanged hydrogen
(x 10-~ g in scale of amount of converted tetralin).
[From Kabe. T. et al., Ind. Eng. Chem. Res., 30, 1757 (1991)]
of the reactions that make it possible for hydrogen in tetralin to exchange with the hydrogen
molecule. McPherson et al. (1985) suggested that the fact that coal suppressed the forma-
tion of 1-methylindan necessitates a mechanism in which at least one step in the isomeriza-
tion is multimolecular. Penninger (1982) also showed by the kinetics derived from gas-
phase work that gaseous hydrogen participated in the isomerization. Further, it was shown
that the enhancement of ring cracking becomes less significant with increasing concentra-
tion of the hydrocarbon and that the hydrogen-initiated ring cracking is gradually trans-
ferred into a hydrogen donor mechanism as the leading reaction scheme when the concen-
tration of tetralin is increased. A similar phenomenon was observed in the system. In Figs.
4.35 and 4.36, the relationships between the hydrogen/tetralin molar ratio and the amounts
of 1-methylindan and n-butylbenzene formed or the hydrogen exchange ratios approximate-
ly follow straight lines. This means that the increase in tetralin decreases the cracking
products and the hydrogen exchange. Tetralin may be activated by collision with itself or a
4.4 HydrogenTransfer Reaction in Coal Liquefaction 229
i
O
O
E
V
0 I I I I I
0 1 2 3 4
Hydrogen/Tetralin (mol/mol)
Fig. 4.36 Effectof molarratio of hydrogento tetralin on amounts of products at 400 ~ for 120 min.
A Naphthalene; D Methylindan; V Butylbenzene.
[From Kabe. T. et al., Ind. Eng. Chem. Res., 30, 1757 (1991)]
hydrogen molecule to produce one intermediate such as the tetralyl radical in Eq. (4.34).
However, if the intermediate would be quenched by tetralin to form original tetralin, the
conversion of tetralin and hydrogen exchange would be inhibited. The reaction of the inter-
mediate with molecular hydrogen leads to hydrogen exchange in both the conversion and
the reproduction of tetralin. Because molecular hydrogen promotes the conversion of
tetralin, it should not quench the intermediate at least more rapidly than tetralin.
The reactions of tetralin with tritiated hydrogen molecule in the presence of coal were
investigated, and results are shown in Table 4.9. Although coal did not largely affect the
formation of 1-methylindan, n-butylbenzene, or decalin at 300-400 ~ the amount of naph-
thalene formed remarkably increased, especially with a rise from 350 to 400 ~ compared
with that in the absence of coal. It is suggested that the interaction between coal and
tetralin is enhanced in the range of 350-400 ~ This is also observed in the variations in
radioactivity in coal and tetralin (Kabe et al., 1991a). The radioactivity in tetralin remark-
ably increased with a rise from 350 to 400 ~ while that in coal did not change greatly. In
the reaction of coal with tritiated molecular hydrogen without solvent, the radioactivity in
coal increased remarkably in the range 350-400 ~ (Kabe et al., 1990b). When tetralin was
added, it was assumed that tritium initially introduced into coal would be rapidly trans-
ferred to tetralin within this temperature range. On the other hand, the hydrogen exchange
ratio of tetralin was much larger than that in the absence of coal as shown in Table 4.9.
This indicates that coal promotes the hydrogen exchange reaction between molecular hy-
drogen and tetralin to introduce tritium into tetralin. King and Stock (1982) reported that
the hydrogen exchange between coal and tetralin-&2 and naphthalene-d8 was readily re-
versible at 400 ~ and that the reactions were initiated by single-bond homolysis and by
molecule-induced homolysis. Further, McMillen et al. (1987) reported that the hydrogen
transfer from donor solvent to coal model compound proceeded by the radical hydrogen
transfer mechanism. Billmers et al. (1986) also reported that hydrogen migration between
9,10-dihydro positions in anthracene structures was consistent with a free radical mecha-
nism. In these reports, the hydrogen transfer processes are reversible and the hydrogen ex-
change reaction can occur through these radical mechanisms. In our systems, since coal
230 4 Liquefaction of Coal
TL NP MI BB DL HER, %
generates radicals more easily than tetralin, gaseous hydrogen must be more easily activat-
ed on coal surface than tetralin. Tritium transferred into coal would readily exchange with
hydrogen in tetralin through a radical mechanism.
In the presence of coal, it has been reported that hydrogen in naphthalene, which was
initially added, exchanges with gaseous hydrogen (Kabe et al., 1987a). Naphthalene
formed from tetralin was isolated. In spite of the release of hydrogen, it contained tritium
from the gas phase in the absence and presence of coal (Kabe et al., 1991 a). Two pieces of
data in the absence and presence of coal were chosen because the hydrogen exchange ratios
were very similar to each other. The amount of hydrogen exchanged per mole of naphtha-
lene in the absence of coal was 1.85 g. Even in the presence of coal where a large amount
of hydrogen was released, 1.00 g of hydrogen in naphthalene was exchangeable with
gaseous hydrogen. It is suggested that when tetralin changes to naphthalene, hydrogen in
tetralin will become very mobile and be able to exchange with molecular hydrogen.
Skowronski et al. (1984) reported that, in the hydrogen exchange between tetralin-d~2
and hydrogen in coal at 400 ~ for 1 h in a shaken autoclave system, protium was incorpo-
rated into Ha (66%), H~ (23%), and Har (11%) positions in tetralin and that the H~ absorp-
tion of the recovered naphthalene in IH NMR was approximately 7 times as intense as the
H~ absorption. Collin and Wilson (1983) showed from their insensitive nucleus enhance-
ment by polarization transfer (INEPT) and gated spin echo (GASPE) NMR study that, in
the reaction of tetralin with deuterium and coal, the mixture of tetralin consists of molecules
that were nondeuterated and monodeuterated at Ha and/or H~ positions while no evidence
was found for any molecules that were dideuterated at Ha and/or H~ positions. In their
NMR measurement, the intensity of the signal at the H,~ position was larger than that at the
H~ position (Collin and Wilson, 1983). Franz and Camaioni (1980), in a set of pyrolysis
experiments with peresters of 1-tetralyl, 2-tetralyl, l-indanylmethyl, and 2-indanylmethyl,
concluded that as pyrolysis of the 1-tetralyl perester gave no detectable amount of methylin-
dan, formation of 1-methylindan was primarily through the 2-tetralyl radical. These reports
represent that the 1-tetralyl radical appears to be a more important intermediate than the 2-
tetralyl radical in exchange with coal and that, in noncatalylic system without coal, the 2-
tetralyl radical as well as the 1-tetralyl radical would become important. In the hydrogen
exchange of naphthalene resulting from the dehydrogenation of tetralin, the 1-tetralyl radi-
cal would be also important in exchange with coal. In our system with coal, however, the
2-tetralyl radical may be formed competitively with the 1-tetralyl radical to lead to the hy-
drogen exchange at the p-position of naphthalene since 1-methylindan was produced as a
main product.
4.4 HydrogenTransfer Reaction in Coal Liquefaction 231
100
80
O
~- 60
,~ 40
20
Residue Tetralin
Datong Q II
- - 10 Wandoan ~ t-A
X Morwell 0 I--!
"0
5
0
0
V U
Fig. 4.38 Effect of reaction temperature on tritium concentrations in residue and tetralin.
[From Kabe. T. et al., Energy Fuels, 5, 460 (1991)]
hydrogen in Datong coal or Morwell coal exchanged. For Wandoan coal, it was somewhat
small. Since the hydrogen exchange between coal and gaseous hydrogen proceeded rapidly
at the temperatures required for significant coal liquefaction, the exchange seems to be re-
lated to thermally produced radicals.
The change of yields of residue and SRC with reaction time is plotted in Fig. 4.40.
Liquefactions of Datong, Wandoan, and Morwell coals were performed at 400, 400, and
350 ~ respectively. When Morwell coal was liquefied at 400 ~ yields became extreme-
ly large and it was difficult to obtain complete material and tritium balance. To make the
yield of Morwell coal similar to those of Datong and Wandoan, liquefaction of Morwell
coal was performed at 350 ~ At 30 min, the yields of residues of Datong, Wandoan and
Datong 9
o
.,,.~
40 Wandoan
Morwell O
~ 30
~= 20
~Z 10
A
0 --O ..
200 300 400
Reaction temperature (~
Fig. 4.39 Effect of reaction temperature on the hydrogen exchange ratio of coal with gaseous hydrogen.
[From Kabe. T. et al., Energy Fuels, 5, 461 (1991)]
4.4 Hydrogen Transfer Reaction in Coal Liquefaction 233
Morwell coals were 57, 25, and 50%, and the yields of SRC were 42, 61, and 42%, respec-
tively. Although the extent of liquefaction does not necessarily follow the rank or the car-
bon content of coals (Yarzab, 1980), Datong coal, which is the highest rank among three,
was the most difficult to liquefy even at 400 ~ In contrast to these observations, the hy-
drogen exchange reaction showed different results. The hydrogen exchange ratio is plotted
against reaction time in Fig. 4.41. The hydrogen exchange ratio increased with time. After
300 min, the hydrogen exchange ratio of Datong coal was over 50%. Even at 350 ~ the
HER of Morwell coal approached nearly 50%. On the other hand, the HER of Wandoan
coal was small, 30% even after 300 min. These results showed that even though Datong
coal was the most difficult to liquefy, hydrogen in Datong coal were the most mobile
among the three coals. Since Datong coal has the highest rank, it can be presumed to have
the most polycondensed structure. Since the radicals generated in liquefaction can be stabi-
lized in aromatic molecules, they may promote the hydrogen exchange reaction rather than
the hydrocracking reaction. As coal was liquefied, tritium in the gas phase was introduced
into the coal. However, the amount of hydrogen exchanged in coal seems to differ depend-
ing on the kind of coal structure. Fig. 4.42 shows changes in the tritium distribution during
coal liquefaction. Because in the noncatalytic system the amount of hydrogen added into
coal was very small and the hydrogen distribution was nearly constant between the initial
and final stages, it was approximated by straight lines. In Fig. 4.42, horizontal dotted and
solid lines represent the hydrogen distributions of coal (Morwell and Datong 13% ;
Wandoan 17%) and solvent (Morwell and Datong 74%; Wandoan 70%) among three phas-
es, respectively. The arrow in Fig. 4.42 represents the hydrogen distribution of the gas
phase (13%) among the three phases. When the hydrogen exchange reaction approaches
equilibrium among the three phases, the tritium distribution in each phase will approach the
hydrogen distribution in the phase. When Datong coal was used, the hydrogen exchange
between gas phase and coal occurred at the initial stage of the reaction, then tritium was
transferred from coal to solvent. With Morwell and Wandoan coals, the rate of tritium
transfer from gas phase to coal and solvent was approximately equal. In Morwell coal, tri-
100
80
7K-
O
~- 60
40
>,
20
0 I I I I I I
0 1 2 3 4 5
Reaction time (hr)
Fig. 4.40 Changes in yields of residue and SRC with reaction time.
O, A: Datong (400 ~ ~ , / k : Wandoan (400 ~ O, A: Morwell (350 ~
O, ~, O: Residue; A,/k A: SRC
[From Kabe. T. et al., Energy Fuels, 5, 461 (1991)]
234 4 Liquefaction of Coal
60
50
O
40
cD
~x0
.~ 30
~ 20
100
80
~..
9 60
. ,...,
"~ 40
E
~" 20
0 1 2 3 4 5
Reaction time (h)
Fig. 4.42 Change in the tritium distributions during coal liquefaction.
Datong (400 ~ O, II, A; Wandoan (400 ~ (D, ill,/k; Morwell (300 ~ O, IS], A
Coal: O, ~, O; Gas phase: II, [], ~; Solvent: A,/t,, A
Upper solid lines, hydrogen distribution of solvent among three phases; dotted lines, that of coal; arrow,
that of gas phase. [From Kabe. T. et al., Energy Fuels, 5, 462 (1991)]
tium introduced into coal transferred to solvent very slowly, while in Wandoan coal the tri-
tium transfer from coal to solvent was very fast. The reason for this result is not yet under-
stood. The reactivity of hydrogen in coal decreased in the order Datong ~ Morwell
Wandoan, which is consistent with the result from HER in Fig. 4.41. Since tetralin, which
has aromatic and naphthene rings in its structure, can be regarded as a model of one type of
structure in coal, the hydrogen exchange reaction of tetralin with gaseous hydrogen was
also investigated. The hydrogen exchange ratio of tetralin was 0.2% at 350 ~ and in-
creased with increase in temperature. However, the exchange ratio was below 1% even at
400 ~ and the tritium concentration of tetralin was about one tenth that of coal. As shown
4.4 Hydrogen Transfer Reaction in Coal Liquefaction 235
90
O
70 O
O O
m
~ 5o
~= 3o m
1 0 ~ A--
0 T I I I
2 4 6 8
Reaction time (h)
Fig. 4.43 Change in yields of residue and SRC with reaction times at 300 ~ (Kabe et al., 1991b)
Datong: O, A; Wandoan: ~, A; Morwell: O, A
Residue: O, ~, O; SRC: A,/k, A.
[From Kabe. T. et al., Energy Fuels, 5, 462 (1991)]
in Fig. 4.38, a substantial amount of tritium can be introduced into tetralin in the presence
of coal. However, this result shows that tetralin could not be tritiated in the absence of
coal. These results indicate that the exchange reaction of hydrogen in tetralin requires types
of radicals produced from coal which are not produced from the thermolysis of neat tetralin.
It seems that radicals produced in coal react easily not only with gaseous hydrogen but also
with hydrogen in tetralin to cause the hydrogen exchange.
Since it was clarified that the hydrogen exchange reaction proceeded even at 300 ~
coal liquefaction was further investigated at 300 ~ and the results are shown in Fig. 4.43.
Yields of residue and SRC did not change with the elapse of time and Datong coal was
hardly liquefied at 300 ~ Wandoan and Morwell coals were liquefied to give SRC in 20
and 25 wt% yields, respectively. However, these values did not change after 240-360 min,
indicating that hydrocracking reactions proceed slowly at 300 ~ The change in tritium
concentration with time at 300 ~ is shown in Fig. 4.44. The tritium concentration of each
of the three coals approached low constant values below 3000 dpm/g, and that of Datong
coal was the highest among the three. Tritium transfers to coal through both hydrogen ad-
dition and exchange reaction. In order to estimate the hydrogen exchange ratio, the amount
of tritium transferred by hydrogen addition must be subtracted. The hydrogen exchange ra-
tio at 300 ~ is plotted against reaction time in Fig. 4.45. After the amount of hydrogen
added was subtracted, the hydrogen exchange ratio increased in the order of D a t o n g -
Wandoan ~ Morwell. The largest amount of tritium was transferred to Datong coal; how-
ever, since the amount of hydrogen added to Datong coal was larger than that added to
Wandoan and Morwell coals, the hydrogen exchange ratio of Datong coal became small.
The hydrogen exchange ratio for Datong, Wandoan, and Morwell approached constant val-
ues, 4.5, 5.0 and 7.8%, respectively. The HER for Morwell coal was the largest and there-
fore the hydrogen exchange at 300 ~ may be related to the exchange of hydrogen in func-
tional groups such as -OH and -NH. Although a detailed analysis of such active hydrogen
has not been done, the comparable analysis of bituminous coals has been reported
(Pestryakov, 1986; Maekawa, 1975). Bituminous coals which have a chemical composition
236 4 Liquefaction o(Coal
3
O
.,..~
I I I I
0 2 4 6 8
Reaction time (h)
Fig. 4.44 Change in tritium concentrations with reaction time at 300 ~
[From Kabe. T. et al., Energy Fuels, 5, 462 (1991)]
~" 8 Datong 9 O ~
O
.2 Wandoan 9
*~ 6 . Morwell 2 / / / ~ fib
C9
4
cD
;;m
I I I
0 2 4 6
Reaction time (h)
Fig. 4.45 Effect of reaction time on the hydrogen exchange ratio of coal with gaseous hydrogen at 300 ~
[From Kabe. T. et al., Energy Fuels, 5, 462 (1991)]
of C 75-85% and H 5.0-5.4wt% contain 6-12 atom% of phenolic OH hydrogen for total
hydrogen. Yokoyama et al. (1967) reported that high-rank coals, which have a chemical
composition of C 75-84% and H 5.8-6.4% (daD, contain 3-9 atom% of phenolic (OH) hy-
drogen and carboxylic acid (COOH) hydrogen for total hydrogen, while low-rank coals
which have a chemical composition of C 61-70% and H 5.3-6.0% (daf), contain 12-14
atom% of those. Kotanigawa et al. (1979) reported that the exchange reaction between
deuterium gas and aromatic hydrogen in phenol took place rapidly at 350 ~ with ZnO-
Fe203 catalyst and that no such exchange reaction occurred in the absence of catalyst. They
did not refer to the exchange reaction between deuterium gas and hydrogen of the hydroxy
group in phenol.
The reaction of phenol with tritiated gaseous hydrogen was carried out at 340 ~ for 2
h in the absence of a catalyst; 8.8% of the hydrogen in phenol underwent tritium exchange.
4.4 HydrogenTransferReaction in Coal Liquefaction 237
Since it can be assumed that only hydrogen in the hydroxy group in phenol is exchange-
able, this indicates that 53% of the hydrogen in the hydroxy group in phenol exchanged
with gaseous hydrogen at 340 ~ for 2 h. Further, the reaction of aniline with tritiated
gaseous hydrogen was also performed at 300 ~ for 2 h in the absence of a catalyst; 13.7%
of the hydrogen in aniline underwent tritium exchange. Since it can be assumed that only
the hydrogen in the amino group in aniline is exchangeable, this indicates that 48% of the
hydrogen in the amino group in aniline exchanged with gaseous hydrogen at 300 ~ for 2 h.
These results support the suggestion that, in the reaction of coal with gaseous hydrogen,
OH and NH hydrogen in polycondensed aromatic compounds were exchanged at lower
temperatures.
4.4.4 Effect of Solvent
Solvents play an important role in coal liquefaction because they can be used as hydrogen
donors and dissolve some portion of the coal (Whitehurst et al., 1980). Solvents with naph-
thene tings, such as tetralin, mainly serve as donor solvents, while those with two or more
aromatic tings, such as methylnaphthalene, dissolve a larger amount of coal than those with
aliphatic structures because coal mainly consists of condensed aromatic structures.
Naphthenic solvents, such as decalin, may have poorer ability to donate hydrogen or to dis-
solve coal than tetralin or methylnaphthalene. Since the liquefaction includes hydrogena-
tion and hydrocracking of coal, with hydrogen in the gas phase and solvent, a number of at-
tempts have been made to clarify the hydrogen transfer mechanism in coal liquefaction in
the presence of solvents (Billmers et al., 1986; McMillen et al., 1985; Murakata et al.,
1993). Billmers et al. (1986) suggested a free radical mechanism following kinetic experi-
ments in model reactions of coal liquefaction. McMillen et al. (1985) reported the impor-
tance of solvent radicals in the hydrogen transfer reaction between the coal model and sol-
vent. In these researches, tetralin was used as a hydrogen donor solvent, and decalin
(Murakata et al., 1993) and methylnaphthalene (Oga et al., 1985; Sato et al., 1992), which
have poor ability as hydrogen donors, were used as solvents.
A more useful method to trace hydrogen transfer mechanisms in coal liquefaction is to
utilize isotopes, such as deuterium and tritium tracers. A deuterium tracer was effective in
tracing reactive sites in coal and coal model compounds; however, there are few examples
which enable quantitative analysis of hydrogen mobility in coal because of the poor solubil-
ity of coal products and the difficulty of quantification of the deuterium tracer (Fu and
Blaustein, 1967; Franz and Camaioni, 1981b; Brower, 1982; Schweighardt et al., 1976;
Cronauer et al., 1982; Wilson et al., 1984; Collin and Wilson, 1983; Skoweonski et al.,
1984). Further, hydrogen transfer mechanisms in the presence of various solvents have not
yet been sufficiently clarified using the deuterium tracer. For example, Benjamin et al.
(1982) studied the hydrogen exchange reaction of a group of aromatic compounds in recy-
cled solvents with diphenylmethane-d2 (Ph2CD2), deuterated pyrene or D2 gas under lique-
faction conditions, assuming that reactivity toward hydrogen exchange is related to hydro-
gen shuttling. They reported that methyl substituted aromatics, such as methylnaphthalene
and toluene, underwent extensive exchange reactions, while non-substituted aromatics,
such as naphthalene, biphenyl ether, showed little observable exchange with three deuterat-
ed reagents. They concluded that the methyl substituted aromatic and hydroaromatic com-
pounds in the recycled solvent make the most important contribution to hydrogen shuttling
and hydrogen transfer. However, the detailed mechanisms and the position of the hydrogen
exchange were not discussed. Recently, it has been reported that tritium and 14C tracer
techniques are effective in tracing quantitatively the hydrogen in coal liquefaction (Kabe et
238 4 Liquefaction of Coal
al., 1987b, 199 la-b, Ishihara et al., 1993). In these works, it was shown that quantitative
analysis of hydrogen mobility in coal can be determined by hydrogen exchange reactions
between the coal, gas phase and solvent, as well as by hydrogen addition. In order to inves-
tigate the effect of the kind of solvent on hydrogen transfer between the coal, gas phase and
solvent, the reaction of tetralin, decalin and 1-methylnaphthalene with tritiated gaseous hy-
drogen was investigated in the absence and presence of coal to estimate hydrogen mobility
10
~ 6
O
O
4
0 , -- , , , , ,A. . . .
280 300 320 340 360 380 400 420
Temperature (~
Fig. 4.46 Effect of temperature on the yields of products from solvents in the absence of coal.
Products from tetralin: O Naphthalene; 9 n-butylbenzene; 9 l-methylindan
Products from decalin: A Naphthtalene; I, Tetralin
Products from 1-methylnaphthalene: [] Naphthalene
[From Kabe. T. et al., Prepr., ACS Div. Petrol. Chem. (1994)]
30
~-
O 20
O
.,,.,
10
0
280 300 320 340 360 380 400 420
Temperature (~
Fig. 4.47 Effect of temperature on the hydrogen exchange ratio of solvents in the presence and
the absence of coal.
In the presence of coal: O Tetralin; A Decalin; V-I 1-Methylnaphthalene
In the absence of coal: 9 Tetralin; I, Decalin; 9 1-Methylnaphthalene
[From Kabe. T. et al., Prepr., ACS Div. Petrol. Chem. (1994)]
4.4 HydrogenTransfer Reaction in Coal Liquefaction 239
from decalin was observed, the yields were very small, as shown in Fig. 4.49. The amount
of hydrogen addition from tetralin to coal at 300 ~ was 0.13 g, similar to that from decalin
240 4 Liquefaction of Coal
100
80
O
60
O
O
.,..~
~. 40
O
20-
0 , I i I i I , I , I i l i
to coal at 400 ~ i.e., 0.11 g. Coal conversion with tetralin at 300 ~ was 35%, which was
close to that with decalin at 400 ~ i.e., 45%, indicating that such an a m o u n t of hydrogen
f r o m solvent to coal can c o n v e r t m o r e than one third of W a n d o a n coal. A significant
a m o u n t of naphthalene was f o r m e d from 1-methylnaphthalene. The yield of naphthalene
f r o m 1 - m e t h y l n a p h t h a l e n e r e m a r k a b l y i n c r e a s e d with t e m p e r a t u r e and r e a c h e d 17% at
400 ~ The fact that the conversion of coal in m e t h y l n a p h t h a l e n e was higher than that in
decalin may be due to the difference in the solubility of coal in m e t h y l n a p h t h a l e n e and de-
calin, or the addition of methyl radicals formed from m e t h y l n a p h t h a l e n e to coal radicals.
30
es 20
O
~, 10
, I , ! , ~ ,
0 i i ~ -
Table 4.11 Tritium Distribution after Reaction of Gaseous Hydrogen in the Presence of
Solventsa
Temperature Rgas Rcoal Rsolvent
(~ Solvent (dpm) (dpm) (dpm)
300 Tetralin 937202 62798 27581
Decalin 906860 90492 2648
1-Methylnaphthalene 897886 79247 22868
350 Tetralin 727018 159213 113768
Decalin 788071 190152 21777
1-Methylnaphthalene 842397 112959 44644
400 Tetralin 576002 181586 242412
Decalin 603196 225842 170962
1-Methylnaphthalene 458413 143554 398033
a Reaction time, 120 min. Total radioactivities were normalized to 106 dpm. Tritium re-
covery, 100 +__5% [Reproduced with permission frorfi Ishihara, A. et al., Fuel, 74, 66,
Elsevier (1995)]
Table 4.12 Hydrogen Transfers among Coal, Gas Phase and Solvent"
Amount of hydrogen added Amount of hydrogen exchanged
Temperature Gas to coal Solvent to coal Solvent Coal
(~ Solvent (g) (g) (g) (g)
300 Tetralin 0 1.27 X 10 - l b 3.56 X 10 -2 1.17 X 10 -1
Decalin 0 9.78 X 10-5c 3.45 X 10 -3 1.21 X 10 -1
Methylnaphthalene 0 3.96 X 10- 5d 3.11 X 10- 2 1.39 X 10-1
350 Tetralin 0 3.41 X 10-1 b 1.89 X 10-1 4.54 X 10 -1
Decalin 0 6.53 X 10- 3c 6.26 X 10- 2 3.18 X 10-1
Methylnaphthalene 0 5.88 X 10- 2d 6.47 X 10- 2 2.29 X 10-1
400 Tetralin 4.04 X 10- 2 5.26 X 10-1 b 5.09 X 10-1 8.50 X 10-1
Decalin 5.96 X 10 -2 1.11 X 10 -lc 3.34 X 10 -1 7.16 X 10-1
Methylnaphthalene 6.38 X 10- 2 1.96 X 10-1 d 1.06 X 10~ 1.38 X 10~
aCoa130 g: Solvent 75 g; 5.9 MPa. Initial amounts of hydrogen in the gas phase were 1.21, 1.18 and 1.22 g in the
cases of tetralin, decalin and methylnaphthalene, respectively. Initial amounts of hydrogen in solvent were 6.82,
9.78 and 5.28 g for tetralin, decalin and methylnaphthalene, respectively.
bAmount of hydrogen added with formation of naphthalene.
cAmount of hydrogen added with formation of naphthalene and tetralin.
din the case of methylnaphthalene, hydrogen addition from coal to solvent occurred to from naphthalene and
methane. [Reproduced with permission from Ishihara, A. et al., Fuel, 74, 67, Elsevier (1995)]
242 4 Liquefactionof Coal
demethylation of toluene proceeded via radical chain mechanism, and that the reaction of
toluene with the hydrogen atom to form benzene and a methyl radical was the rate-deter-
mining step. It has also been reported that, in hydrocracking of toluene at 455-490 ~ and
0.14 MPa, dealkylation of toluene proceeds via the radical chain mechanism (Gonikberg
and Nikitenkov, 1954). A similar mechanism has been reported for demethylation of
methylnaphthalene (Sato et al., 1993). Ogata et al. (1983) reported that the hydrogen atom
which was formed by the addition of hydrogen sulfide, promotes dealkylation of methyl-
naphthalene. Further, Ogo et al. (1985) showed that dealkylation of methylnaphthalene
proceeds by a radical chain mechanism, where radicals or hydrogen atoms are formed sec-
ondarily or by thermolysis of biphenyl. Taking into account these reports, the radical chain
mechanism can be assumed in the case of methylnaphthalene. Possible mechanisms of hy-
drogen exchange and dealkylation with methylnaphthalene are shown in Eqs. (4.35)-(4.43)
Hydrogen exchange reactions of coal proceed through the route shown in Eqs.
(4.35)-(4.37). The reactions in Eqs. (4.35)-(4.37) are common among tetralin, decalin and
methylnaphthalene. The tritium radical formed will react with methylnaphthalene to form
tritiated naphthalene and a methyl radical, as shown in Eq. (4.38). The methyl radical re-
acts with another methylnaphthalene to form methane and a naphthylmethyl radical, which
may react with tritiated coal to form tritiated methylnaphthalene and a coal radical, as
shown in Eqs. (4.39) and (4.40). Dealkylation of methylnaphthalene can be explained by
Eqs. (4.38)-(4.40) reasonably. In hydrogen exchange of methylnaphthalene through these
routes, however, the HER does not exceed 4% of hydrogen in methylnaphthalene when it is
calculated on the basis of conversion of methylnaphthalene, Under conditions where
methylnaphthalene decomposes, methylnaphthalene will easily form a naphthylmethyl radi-
cal through the reaction with the coal and tritium radical as shown in Eqs. (4.41) and (4.42).
A naphthylmethyl radical formed in such reactions can be tritiated through the routes
shown in Eqs. (4.40) and (4.43). The remarkable decomposition of methylnaphthalene and
the high HER value of methylnaphthalene at 400 ~ can be explained by considering the
radical chain mechanism described above. If it is assumed that hydrogen exchange pro-
ceeds through the free radical mechanism, the relative ease of the formation of radicals
would be related to the strength of a bond. For example, the bond dissociation energies of
C-H in the benzene and benzyl positions of toluene are 469 and 356 kJ mol-', respectively.
The latter hydrogen will be easier to exchange than the former. Concerning the hydrogen
exchange reaction of tetralin, extensive studies using deuterium have been reported. As
mentioned in section 4.4.2, it was proposed that the 1-tetralyl radical appears to be a more
important intermediate than the 2-tetralyl radical in exchange with coal. In the hydrogen
exchange of naphthalene resulting from the dehydrogenation of tetralin, the 1-tetralyl radi-
cal would also be important in exchange with coal. Possible mechanisms of hydrogen ex-
change of tetralin in the presence of coal are shown in Eqs. (4.44)-(4.47).
1-Tetralyl radicals are formed by the reaction of tetralin with a coal radical or a tritium
atom, as shown in Eqs. (4.44) and (4.45). The formed tetralyl radical will react with a triti-
ated hydrogen molecule or a tritium atom in tritiated coal to form tritiated tetralin, as shown
in Eqs. (4.46) and (4.47). Although tetralin may easily form a tetralyl radical, its lifetime is
shorter than that of a naphthylmethyl radical, because the formation of naphthalene from a
tetralyl radical, as shown in Eq. (4.48), occurs more easily in the presence of coal than the
formation of tritiated tetralin, as shown in Eqs. (4.46) and (4.47). Therefore, the HER of
tetralin in the presence of coal did not increase with a rise from 350 to 400 ~ as much as
that of methylnaphthalene. Decalin does not form such radicals as the naphthylmethyl or
tetralyl radical, which are stabilized by aromatic rings. Therefore, the remarkable increase
4.4 HydrogenTransferReactionin CoalLiquefaction 243
CH3 T CH3
T
4- CH3" (4.38)
CH3 CH2-
@ 4- CH3. - @ 4- CH4 (4.39)
CH2, CHzT
@ + Coal-T - @ + Coal. (4.40)
CH3 CH2,
+ Coal-H (4.41)
CH3 CH2,
+ T. ' ~ + T-H (4.42)
CH2, CH2T
4- T2 " ~ " 1 + T. (4.43) .
+ T. ~ + T-H (4.45)
T
@ + T-H @ + H- (4.46)
244 4 Liquefaction of Coal
T
~ ] + Coal-T " ~ ~ ] + Coal. (4.47)
in tetralin or decalin HER with a rise from 350 to 400 ~ did not occur.
Hydrogen exchange of hydrogen in coal with tritiated gaseous hydrogen was also esti-
mated. Since the HERs of solvents markedly increased in the presence of coal, it can be as-
sumed that tritium in the gas phase may transfer to a solvent through coal. Based on this
assumption, the HER of coal in the presence of solvent was estimated and plotted against
temperature in Fig. 4.50. The reaction of coal with tritiated gaseous hydrogen in the ab-
sence of a solvent was also performed, for comparison with that in the presence of a sol-
vent. Tritium balance and the amount of hydrogen exchanged are listed in Table 4.13. The
HER of coal in the absence of a solvent was also plotted in Fig. 4.50. The HER of coal in
inactive decalin was very similar to that in the absence of a solvent. Since tetralin can easi-
ly form a tetralyl radical under the same conditions, the HER of coal in tetralin was slightly
higher than that of decalin. However, tetralin did not promote to an extreme degree the hy-
drogen exchange reaction among the gas phase, coal and solvent for the reasons described
above. In contrast, the value of methylnaphthalene at 400 ~ deviated largely from the oth-
er curves. Hydrogen exchange in the presence of methylnaphthalene can be regarded as a
reaction proceeding through the radical chain mechanism described in Eqs. (4.35)-(4.43).
Different from the cases of tetralin and decalin are the reactions of the naphthylmethyl radi-
cal in Eqs. (4.40) and (4.43). Especially, the remarkable deviation of methylnaphthalene at
400 ~ in Fig. 4.50 may be related to the reaction of Eq. (4.43), in which tritium can be in-
corporated from the gas phase to methylnaphthalene directly.
Table 4.13 Tritium Distribution and Hydrogen Transfer after Reaction of Coal with Tritiated
Gaseous Hydrogen in the Presence of Solvents"
Amount of Amount of
Temperature Rgas Rcoal hydrogen hydrogen
(~ (dpm) (dpm) exchanged (g) added (g)
260 993261 6378 1.03 X 10- 2 0.0000
300 896987 103013 1.84 X 10- ~ 0.0000
360 775033 224967 3.35 X 10-1 0.1006
385 705166 294834 5.02 X 10- ~ 0.1176
410 650260 349740 7.81 X 10- ~ 0.0516
a Reaction time, 120 min. Total radioactivities were normalized to 106 dpm. Tritium recov-
ery, 100 ___5%. Initial amount of hydrogen in gas phase, 1.6 g. Initial amount of hydrogen
in coal. 1.86 g. [Reproduced with permission from Ishihara, A. et al., Fuel, 74, 65, Elsevier
(1995)]
4.4 Hydrogen Transfer Reaction in Coal Liquefaction 245
100
o 80
0
~0
"~ 60
N 40
.~ 2o
I , I
0
200 300 400 500
Temperature (~
Fig. 4.50 Effect of temperature on the hydrogen exchange ratio of coal.
O Tetralin; A Decalin; E] 1-Methylnaphthalene; 9 Solvent free
[Reproduced with permission from Kabe. T. et al., Prepr. A CS Div Petrol. Chem. (1994)]
because the complex nature of coal and its derived products prevent a thorough understand-
ing of the reaction mechanism.
Here, the effect of H2S and H20 on the hydrogen exchange in the liquefaction using a
model compound tetralin is discussed (Godo et al., 1997b, 1998c). Tetralin is one of the
most interesting and convenient model compounds because it has an aromatic ring and a
naphthene ring in its structure and can serve as an effective hydrogen donor solvent.
The reactions of tetralin with tritiated hydrogen were performed under the conditions of
3 5 0 4 0 0 ~ in the presence of H2S. The products yields are plotted against the reaction
time in Fig. 4.51. The reaction products from tetralin were 1-methylindan (MI) by isomer-
ization, naphthalene (NP) by dehydrogenation and n-butylbenzene (BB) by hydrocracking.
They increased monotonically with reaction time. The yields of MI, NP and BB in the pres-
ence of H2S at 400 ~ for 300 min were 2.3, 0.7 and 0.9%, respectively. Decalin was not
formed under by disproportionation these conditions. The results in the absence of H2S
were also plotted in Fig. 4.51. The yields of MI, NP and BB at 400 ~ for 300 rain were
2.1, 0.8 and 1.1%, respectively. The amount of each product in the presence of H2S was
close to that in the absence of HaS.
In the reaction of tetralin with tritiated hydrogen in the presence of H2S, the conversion
and hydrogen exchange ratio of tetralin were plotted against reaction time in Figs. 4.52 and
4.53, and compared with those in the reaction in the absence of H2S. As shown in Fig.
4.52, the conversions of tetralin in the presence and absence of H2S at 400 ~ for 300 min
were 4.1 and 3.9 %, respectively. These values were very close to each other. As shown in
Fig. 4.53, the hydrogen exchange ratios of tetralin in the presence and absence of H2S in-
creased gradually with time and reached 40.4 and 4.6%, respectively, at 400 ~ for 300
min. The hydrogen exchange ratio in the presence of HaS was about 10 times higher than
that in the absence of H2S at 375 ~ and 400 ~ and reached 71% at 400 ~ for 600 min.
3
0
9
2 / /
/
~D
.,..~
0
~ I
-.-~~
~ I , I ,
0 200 400 600 800
Reaction time (min)
Fig. 4.51 Effectof reaction time on product yields at 400 ~
In the presence of H2S:
O: Naphthalene; A: n-Butylbenzene; I1: 1-Methylindan.
Reaction in the presence of H20:
(1: Naphthalene; A: n-Butylbenzene; []: 1-Methylindan.
In the absence of H2S and H20:
C): Naphthalene; A: n-Butylbenzene; [S]: 1-Methylindan.
[From Godo. M. et al., Energy Fuels, 11,472 (1997)]
4.4 Hydrogen Transfer Reaction in Coal Liquefaction 247
10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8 ," ,i"
/,
9 //
6
0
t/ /"
.2 4 / / /"
/" / ,
D" / /" s.,l I
o ,,0/'/,"
2 / 9
0 ...... :' ' " ' ' '' ' ' ' ' ' '
0 200 400 600 800
Reaction time (rain)
Fig. 4.52 Effect of reaction time on the conversion of tetralin.
Reaction in the'presence of HzS: h-: 350 ~ I1:375 ~ A:400 ~
Reaction in the presence of H20: []: 375 ~ A: 400 ~ ~:425 ~
Reaction in the absence of H2S and HzO: []: 375 ~ A: 400 ~ 0:425 ~
[From Godo. M. et al., E n e r g y F u e l s , 11,472 (1997)]
8 / 80
/
O
t~ 6 6O
~D
=
~D
4 ~) ,/,,'" // 40
O
0.00
--0.02
--0.04
X|
',,,,
,~ -0.06
-0.08
(a)
--0.10
--0.12
' 2()0 ' 4()0 ' 6()0 ' 800
Reaction time (min)
0.0 0.0
-0.5 --0.5
X |
-0.10 --1.0
(b)
. . . . . - 1.5
--0.15 0 2(}0 4()0 600 800
Reaction time (min)
Fig. 4.54 First-orderplots of conversion of tetralin and hydrogen exchange ratio.
a) Conversion of tetralin, b) Hydrogen exchange ratio.
Reaction in the presence of HzS:+: 350 ~ I1:375 ~ A: 400 ~
Reaction in the presence of H20: ~: 375 ~ 400 ~ ~: 425 ~
Reaction in the absence of H2S and H20: F-l: 375 ~ A: 400 ~ 0:425 ~
[From Godo. M. et al., Energy Fuels, 11,473 (1997)]
and 400 ~ were not significantly different. On the other hand, as shown in Table 4.15, the
rate constants of hydrogen exchange reactions in the presence of H2S were about 12-15
times higher than those in the absence of H2S. From the Arrhenius plots, activation ener-
gies for the tetralin conversion in the absence and presence of H2S were determind. They
were 3 3 _ 1 and 35__+1 kcal/mol, respectively. These values were very close to each other,
suggesting that the conversion of tetralin in the absence and presence of HzS proceeds via
on the same reaction mechanism, and that the presence of H2S did not affect the activation
energies. Similarly, activation energies for hydrogen exchange reactions were determind.
Activation energies of the hydrogen exchange in the absence and presence of HiS were
30___ 1 and 34___1 kcal/mol, respectively. These values were slightly different from each
other, suggesting that the hydrogen exchange of tetralin in the absence and presence of H2S
proceed via different mechanisms.
The reactions of tetralin with tritiated hydrogen were performed under the conditions of
4.4 HydrogenTransfer Reaction in Coal Liquefaction 249
375-425 ~ in the presence of H20. The product yields are also plotted against the reaction
time in Fig. 4.51. The yields of MI, NP and BB at 400 ~ for 300 min in the presence of
H20 were 0.9, 0.2 and 0.4%, respectively. These values were less than half those in the ab-
sence of H20. Under every condition in the absence and presence of H20, the ratios among
MI, NP and BB did not differ significantly.
In the reaction of tetralin with tritiated hydrogen in the presence of H20, the conversion
and hydrogen exchange ratio of tetralin were also plotted against reaction time in Figs. 4.52
and 4.53. The conversions of tetralin and hydrogen exchange ratio in the presence of H20
at 400 ~ for 300 min were 1.7 and 1.3%, respectively. The hydrogen exchange ratio at
375-425 ~ in the presence of H20 was about one third of that in the absence of H20.
First-order plots of these data for the c o n v e r s i o n and h y d r o g e n e x c h a n g e are also
shown in Fig. 4.54. As shown in Tables 4.14 and 4.15, both the rate constants of the
tetralin conversion and hydrogen exchange in the presence of H20 were smaller than those
in the absence of H20 at 375-425 ~ Similarly, from the Arrhenius plots of the rate con-
stants in the presence of H20 shown in Table 4.14, the activation energy of tetralin conver-
sion in the presence of H20 was determined and was 35_+ 1 kcal/mol. This value was very
close to that in the absence of H20, suggesting that the conversion of tetralin in the pres-
ence of H20 proceeded via the same reaction mechanism as that in the absence of H20 de-
pended. Similarly, the activation energy of the hydrogen exchange in the presence of H20
was determined from Table 4.15 and was 2 4 _+ 1 kcal/mol. In contrast to the results from
the conversion of tetralin, this value was different from that in the absence of H20. This re-
250 4 Liquefaction of Coal
suit shows that the hydrogen exchange of tetralin in the presence and absence of H 2 0 pro-
ceeds via different mechanisms.
The reactions of tetralin, decalin and naphthalene with tritiated hydrogen were per-
formed at 400 ~ in the presence of H2S. Although tritium was introduced to the solvents
by hydrogen addition and hydrogen exchange reactions, most of the tritium was introduced
through hydrogen exchange. The product yields from solvents in the presence and absence
of H2S are plotted against reaction time in Figs. 4.55 and 4.56. In the reaction with decalin,
tetralin by dehydrogenation and some unidentified products were formed. The yield of
0.5
0.4
.= 0.3
m
0
0.2
.,..~
J
J
0.1
O
0.0
0 100 200 300 400
Reaction time (min)
Fig. 4.55 Effect of reaction time on the yield of tetralin in the reaction of decalin at 400 ~ H2: O" HE/H2S"
[Reproduced with permission from Godo. M. et al., Fuel, 77, 949, Elsevier.(1998)]
= 3
0 ! w i
Fig. 4.56 Effect of reaction time on the yield of tetralin in the reaction of naphthalene at 400 ~ H:: O; H2/H2S: 9
[Reproduced with permission from Godo. M. et al., Fuel, 77, 949, Elsevier. (1998)]
4.4 HydrogenTransfer Reaction in Coal Liquefaction 251
r~
=
~D 3
>
O
O
=
O 2
~D
>
o ~
0 100 200 300 400
Reaction time (min)
Fig. 4.57 Effectof reaction time on the conversion of solvents at 400 ~
H2: O tetralin; A decalin; D naphthalene
Hz]HzS: 9 tetralin; 9 decalin; 9 naphthalene
[Reproduced with permission from Godo. M. et al., Fuel, 77, 950, Elsevier (1998)]
tetralin from decalin in the presence of HzS at 400 ~ for 300 min was 0.2%, and that in the
absence of HzS was also 0.2%. In the reaction with naphthalene, tetralin by hydrogenation
was formed and the yields of tetralin in the presence and absence of HzS at 400 ~ for 300
min were 3.1 and 0.7%, respectively. HzS only affected the hydrogenation of naphthalene.
In the reaction of solvents with tritiated hydrogen in the presence and absence of HzS,
the conversions and hydrogen exchange rations (HERs) of solvents were plotted against re-
action time in Figs. 4.57 and 4.58, and compared with those in the reaction in the absence
of HzS. As shown in Fig. 4.57, the conversions of tetralin, decalin and naphthalene in the
presence of HzS at 400 ~ for 300 min were 4.1, 1.0 and 3.1%, and those in the absence of
HzS at 400 ~ for 300 min were 3.9, 0.9 and 0.7%, respectively. These values of tetralin
and decalin were very close to each other, suggesting that HzS did not participate in the
rate-determining step of the conversions of tetralin and decalin to produce the products.
However, the conversion of naphthalene in the presence of HzS was about four times higher
than that in the absence of HzS. This shows that the hydrogenation of naphthalene is pro-
moted remarkably by HzS.
As shown in Fig. 4.58, HERs for tetralin, decalin and naphthalene in the presence and
absence of HzS gradually increased with time. The HERs for tetralin reached 40.4 and
4.6%, at 400 ~ for 300 min, respectively, in the presence and absence of HzS. HERs for
decalin reached 11.7 and 1.5%, respectively. The hydrogen exchange ratios of tetralin and
decalin in the presence of HzS were about 10 times higher than those in the absence of H2S.
However, HERs for naphthalene in the presence and absence of HzS were 22.9 and 7.7%, at
400 ~ for 300 min, respectively. HERs for naphthalene in the presence of HzS was about
three times higher than that in the absence of HzS. The extent of the promotion effect of
HzS on the hydrogen exchange in naphthalene was nearly equal to that of the conversion of
naphthalene, and was less than those of the hydrogen exchange in tetralin and decalin. As a
result, in the presence for HzS, the HER for tetralin was the highest among the three sol-
vents. Under coal liquefaction conditions, it is well known that HzS is produced by the re-
252 4 Liquefaction of Coal
50
40
O
.,..~
30
= 20
~ 10
0 i B~------'---'r-""'-5"-'-'~, i i
0 100 200 300 400
Reaction time (min)
Fig. 4.58 Effect of reaction time on the hydrogen exchange ratio of solvents at 400 ~
H2: O tetralin; A decalin; I-1 naphthalene
Hz/H2S: 9 tetralin; 9 decalin; 9 naphthalene
[Reproduced with permission from Godo. M. et al., Fuel, 77, 950, Elsevier (1998)]
duction of pyrite in coal. It is suggested that compared with decalin and naphthalene the
hydrogen mobility of tetralin is the highest under coal liquefaction conditions 9
The rate constants of conversion and hydrogen exchange in the presence and absence
of H2S were determined from the slopes of first-order plots and summarized in Tables 4.16
and 4.17. In these tables, k2/kl shows the ratio of the rate constants in the presence of H2S
Table 4.16 Rate Constants for Conversion of Tetralin and Formation of Decalin and
Naphthalene (min-~)
Reaction Solvent
[Reproduced with permission from Godo. M. et al., Fuel, 77, 951, Elsevier (1998)]
Reaction Solvent
Tetralin Decalin Naphthalene
Absence of H2S" kl 1.65 X 10 -4 5.07 X 10 -5 2.66 X 10 -4
800000
= Ca)
"~ 600000
400000
o
200000
.....
0 ,~
Feed 0 200 400 600 800
Reaction time (min)
800000
"~ 600000
=
~ N (b)
400000
200000
, r ir"~~~ll 7 i n I i
Feec~ 0
0
200 400 600 800
Reaction time (min)
800000
(c)
"~ 600000
=
400000
o
200000
".~,
[--.
0
Feed 0 200 400 600 800
Reaction time (min)
Fig. 4.59 Effect of reaction time on the tritium concentration in the presence of H/S.
Reaction Temperature: a) 350 ~ b) 375 ~ c) 400 ~
O tritiated hydrogen; A HzS; [-] tetralin.
[From Godo. M. et al., Energy Fuels, 11,474 (1997)]
and in the absence of H2S. This value represents the extent of the promotion effect of H2S
on the conversion and hydrogen exchange reaction. The promotion effects of H2S on the
hydrogen exchange in solvents increased in the order of naphthalene < decalin < tetralin.
254 4 Liquefaction of Coal
800000
(a)
600000 "----0--
400000
O
200000
/
,~
J
o , II- I , I ~ ,
800000
~. (b)
~, 600000
O-
400000
O
~ 200000
o~
o _ _ _
800000
(c)
600000
oQ.,,.,..
O
400000
O
E 200000
.... j
0 |
Feed 0 100 200 300 400
Reaction time (min)
Fig. 4.60 Changes in the tritium concentration with reaction time in the presence of H20.
Reaction Temperature: a) 375 ~ b) 400 ~ c) 425 ~
O tritiated hydrogen; A H2S; F-] tetralin.
[From Godo. M. et al., Energy Fuels, 11,475 (1997)]
Further, k2/kl of the conversion of naphthalene is close to that of the hydrogen exchange.
Changes in the tritium distribution in the presence of H2S among gaseous hydrogen,
solvents and H2S at 400 ~ with reaction time are shown in Fig. 4.59a,b,c. In these figures,
4.4 HydrogenTransfer Reaction in Coal Liquefaction 255
solid horizontal lines show the equilibrium value which was calculated on the assumption
that hydrogen atoms were completely scrambled between gaseous hydrogen, solvents and
HzS. The tritium initially included in gaseous hydrogen decreased monotonically with the
passage of time. On the other hand, the tritium concentration in solvents increased monoto-
nically with time. The tritium concentration in HzS of the each solvents system increased
rapidly, showed the maximum value beyond the calculated equilibrium value at 0 min and
reached a very close value to that in gaseous hydrogen, and decreased with the same ten-
dency as gaseous hydrogen molecules. This result shows that the hydrogen exchange reac-
tion between gaseous hydrogen and H2S proceeded rapidly at the initial reaction stage and
reached equilibrium between the two phases. At 300 min, the tritium distribution among
gaseous hydrogen, solvents and HzS approached the calculated equilibrium value.
Variation in the tritium distribution in the presence of H20 among gaseous hydrogen,
tetralin and HzO at 375, 400 and 425 ~ with reaction time are shown in Fig. 4.60 (a-c).
Since the rate of hydrogen exchange between gaseous hydrogen and tetralin in the presence
of H20 was far smaller than that in the presence of HzS, the decreases of tritium in gaseous
hydrogen and the increases of tritium in tetralin solvent were slower than those in the pres-
ence of HzS. The tritium concentration in H20 at every temperature increased gradually
and was close to that in gaseous hydrogen at 300 min. However, the increasing rate of tri-
tium concentration in H20 was obviously slower than that in HzS.
It was assumed that the hydrogen exchange for tetralin in the absence of H2S proceeded
via a tetralyl radical formed by unimolecular dissociation, which also acted as an intermedi-
ate in the conversion of tetralin, as shown in Eqs. (4.24) and (4.25) (Skowronski et al.,
1984; Kabe et al., 1991a; Ishihara et al., 1995; Godo et al., 1997b, c). If the tetralyl radical
is quenched by hydrogen radical (H.), which is formed from original tetralin by the thermal
dissociation, the hydrogen exchange of tetralin will not proceed. Ttie reaction of the tetra-
lyl radical with tritiated hydrogen leads to hydrogen exchange in the absence of H2S.
The tritium distributions of HzS in Fig. 4.59 (a-c) show that the hydrogen exchange re-
action between tritiated gaseous hydrogen and HzS proceeds rapidly under the coal lique-
faction conditions. A possible mechanism of the hydrogen exchange reaction between
gaseous hydrogen and H2S is shown in Eqs. (4.50) and (4.51). In Eq. (4.49), HzS produces
H. and HS. radicals by the thermal dissociation under these reaction conditions. The
formed H. will react with a tritiated hydrogen molecule and produce a tritium radical, as
shown in Eq. (4.50). The reaction of a tritium radical (T.) with HS- leads to hydrogen ex-
change between tritiated gaseous hydrogen and HzS as shown in Eq. (4.51).
The HER between tetralin and tritiated hydrogen in the presence of HzS was about ten
times higher than that in the absence of HzS. In the case of hydrogen exchange of tetralin
in the presence of HzS, H- produced from tetralin via Eq. (4.24) or from H2S via Eq. (4.49)
changes into T- through Eqs. (4.50) or (4.52). Eq. (4.52) is considered to be much faster
than Eq. (4.50), because the bond dissociation energy of hydrogen sulfide (370 kJ/mol) is
much smaller than that of hydrogen (432 kJ/mol). Stenberg et al. (1983) reported the hy-
drogen donor ability of HzS by the stoichiometry of the biphenyl conversion and the favor-
ably low bond dissociation energies of HzS compared with H2. The bond energy of H2 is
greater than that of most C-H bonds whereas that for H2S is not. It is assumed that the con-
centration of T. increases remarkably in the presence of HzS. The most likely and energeti-
cally feasible early initiation step is the addition of T. into tetralin, promoting the formation
of a hydrotetralyl radical and the hydrogen exchange in an aromatic ring, as shown in
Eq.(4.53). The hydrogen exchange reaction as shown in Eqs. (4.45) and (4.46) may also
occur in the presence of HzS. In the reaction of tetralin, decalin was not formed by hydro-
236 4 Liquefaction of Coal
+ To - ~ + HT (4.54)
H
. ~ H + Ho
+ HT (4.55)
Both conversion and hydrogen exchange reaction in naphthalene were promoted about
four times in the presence of H2S. It was suggested that the rate-determining steps of the
conversion and hydrogen exchange reaction were the same and that H2S promoted the rate
determining step. The product in the conversion of naphthalene was tetralin by hydrogena-
tion. Tetralin would be produced by the multi-steps of hydrogenation, as shown in Eq.
(4.55). Tritium was introduced to the tetralin by hydrogenation. However, the amount of
tritium added was very small in comparison with the amount of tritium exchanged.
In the presence of H2S, atritium radical is easy to generated and therefore the tritiated
naphthalyl radical is also easy to be generated. The tritiated naphthalyl radical reacts with
hydrogen atom to lead to dihydronaphthalene, which converts into tetralin. Further, HS.
also abstracts the hydrogen atom in the tritiated naphthalyl radical reversibly to produce tri-
tiated naphthalene, promoting the hydrogen exchange. Therefore, it is assumed that the
common rate-determining step of the conversion and the hydrogen exchange of naphthalene
is the formation of the naphthalyl radical. This mechanism would explain the same degree
of the promotion effect of H2S on the conversion and hydrogen exchange.
The hydrogen addition and exchange reactions between tetralin and gaseous hydrogen,
and the formation of naphthalene and 1-methylindan from tetralin are considered to proceed
via a radical reaction mechanism (Kabe et al., 1990d; Ishihara et al., 1995; Godo et al.,
4.4 HydrogenTransfer Reaction in Coal Liquefaction 257
T
~ + To . ~ + H. (4.56)
IHo
+HT H
HH T
1997b), where a tetralyl radical was an intermediate in the hydrogen exchange and conver-
sion of tetralin. Therefore, the formation of the tetralyl radical in this system was assumed
to be the rate-determining step for both the hydrogen exchange and the conversion of
tetralin.
In the system of tetralin and gaseous hydrogen, tetralin may collide with not only itself
but also tritiated hydrogen. It is possible that tritiated hydrogen affects the step of the for-
mation of a tetralyl radical. If tritiated hydrogen affects the formation of the tetralyl radical,
the hydrogen exchange ratio (HER) and conversion of tetralin will change by the partial
pressure of gaseous hydrogen. As we know, it is difficult to control the reaction pressure in
the reactions using the autoclave because the effective reaction pressure changes depending
on reaction temperature. Further, an autoclave experiment takes longer than ca. 30 minutes
to heat the system to 400 ~ the effective temperature for coal liquefaction. Hydrogen
transfer and conversion of tetralin occur while the autoclave is heated to the set tempera-
ture. This may not be a major problem for comparative experiments but would be unsatis-
factory for kinetic studies, especially in the short reaction time. In contrast, a flow type re-
action system can control the reaction pressure strictly and is suitable for the short reaction
time.
Here, the hydrogen exchange reactions between tetralin and gaseous hydrogen were in-
vestigated in a flow type reactor to estimate the effect of reaction pressure and the kinetics
in the short reaction time (Godo et al., 1997c). The reactions of tetralin with tritiated hy-
drogen were performed under the conditions 400-450 ~ and 2.5-9.8 MPa for 25-420 sec.
The yields of products at 425 ~ are plotted against the reaction time in Fig. 4.61. In this
temperature range, the reaction products from tetralin were 1-methylindan by isomerization
as shown in Fig. 4.61a), naphthalene by dehydrogenation as shown in Fig. 4.61b), and n-
butylbenzene by hydrocracking as shown in Fig. 4.61c); the main product was 1-methylin-
dan. Decalin was not formed under these conditions. This is consistent with the previous
result in the reaction using the autoclave (Godo et al., 1997b), in which the disproportiona-
tion of tetralin to decalin does not occur. The products increased monotonically with resi-
dence time, and the yields of methylindan, naphthalene and butylbenzene at 425 ~ for 450
reached 0.13, 0.05 and 0.06%, respectively. As shown in Fig. 4.61 a), the plots of methylin-
dan at 2.5, 4.9 and 9.8 MPa showed approximately the same straight line and were not in-
fluenced by the reaction pressure. A similar result was also obtained for naphthalene as
shown in Fig. 4.6 l b). The results in the reaction using the autoclave reported are also plot-
ted in Fig. 4.61. Those reactions were conducted under the conditions at initial pressure 5.9
MPa for 0-300 min. In these figures, the ratios between the reaction time and the yields of
products in the reactions using the autoclave are the same as those in the reactions using the
flow reactor. So the relationship between the reaction time and the product yield can be
238 4 Liquefaction of Coal
0.15 15.0
.=, []
0.10 10.0
0 0
-~ 0.05
~D
5.0 _ ~D
0.00 t : I : I , I , I , 0.0
~, 0.06 6.0
~D
-fi 0.04
[]
7 4.0 "fi
0 O
0.02 2.0 __.
.,.9,
0.00 i , I , 0.0
c [] []
~0.06 6.0
0
0.02 2.0
~D
.,..q
i , I , I , I ,
0.00 ( 0.0
0 100 200 300 400 500
Residence time (s)
compared within the same figure. The yields of methylindan and naphthalene in the reac-
tion using the autoclave were slightly less than those of the flow reactor. However, these
values were not affected significantly by the type of reactor. In the reaction using the auto-
clave, although the initial pressure was 5.9 MPa, the effective reaction pressure was about
14.7 MPa at 425 ~ by thermal expansion of gas phase. The result shows that the yields of
methylindan and naphthalene are not affected considerably by the reaction pressure in the
range from 2.5-14.7 MPa. In contrast, the yields of butylbenzene in the reaction in the
4.4 Hydrogen Transfer Reaction in Coal Liquefaction 259
flow reactor increased with rise of reaction pressure, and the formation of butylbenzene de-
pended on the reaction pressure. The yield of butylbenzene in the reaction using autoclave
also increased monotonically with reaction time because the reaction pressure of autoclave
was nearly equal when the reaction temperatures were the same, and were close to that in
the reaction of a flow reactor at 4.9 MPa.
The amount of tritium introduced into tetralin from gas phase is represented by the hy-
drogen exchange ratio (HER). Fig. 4.62 shows the change in HER with reaction time. The
HER of tetralin increased gradually with lapse of time and reached 0.11% at 425 ~ for 450 s.
However, at 400 ~ the hydrogen exchange ratio was only about 0.04 % for 520 s.
Further, the plots of HER could be shown with the same line in the pressure range from 2.5
MPa to 9.8 MPa. In addition, HER in the reaction using the autoclave was almost the same
as that in the flow reactor. The result is similar to that of the conversion of tetralin into
methylindan and naphthalene.
Reaction time • 10 -4 (s)
0 1 2 3 4 5
0.15] , i , i , i , t , 15.0
.o .- D O
..a
0.10 10.0
Ca)
Fig. 4.62 Effect of residence time on the hydrogen exchange ratio of tetralin at 425 ~
[From Godo. M. et al.,AIChE, 43, 3108 (1997)]
First-order plots of these data for the formation of methylindan and naphthalene, and
hydrogen exchange are shown in Fig. 4.63. All plots approximately fit a linear relationship,
indicating that these reactions could be treated as first-order reactions. The formation rate
constants of methylindan from tetralin could be determined from the slopes of first-order
plots. These constants are shown in Table 4.18. In the same way, the formation rate con-
stants of naphthalene, and the rate constants of hydrogen exchange were obtained and listed
in Table 4.18. From the Arrhenius plots, activation energies of the tetralin conversions into
methylindan and naphthalene, and hydrogen exchange between tetralin and tritiated hydro-
gen were determined and were 3 2 _ 2, 33 --_+2 and 33 +__2 kcal/mol, respectively.
Activation energies in the reaction using the autoclave are also determined and are listed in
Table 4.19. These values were very close to each other, suggesting that the conversion of
tetralin into methylindan and naphthalene, and hydrogen exchange reaction of tetralin pro-
ceeds via the same reaction mechanism.
260
0.000
--0.001
x
,7,, --0.002
,1 OA
--0.003
I m
0.0000 I ,
--0.0002
-0.0004
,,,,,.t
--0.0006
-0.0008
,-~ , I m I m
--0.000
--0.001 A
X !
--0.002
.d
--0.003
(c)
m I , I ,
--0.004
0 200 400 600
Residence time (s)
Fig. 4.63 First-order plot of conversion of tetralin and HER. (Godo et al., 1997c)
a) conversion of tetralin into methylindan; b) conversion of tetralin into naphthalene; c) HER.
[From Godo. M. et al., A1ChE, 43, 3109 (1997)]
Table 4.18 Rate Constants in the Reaction of Tetralin with Tritiated Hydrogen
Temperature Conversion of tetralin Hydrogen exchange
(~ of tetralin
Into methylindan Into naphthalene
(Xl0 -6 sec -l) ( x l 0 -6 sec -1) (x10 -6 sec -1)
Table 4.19 Summaryof Yields from Various Direct Coal Liquefaction Processes Involving Hydrogenation
Process SRC II H-coal ExxonEDS Kohloel BCL NEDOL
Coal used Illinois Illinois Illinois Bituminous Morwell Tanitoharum
Scale of operation (t/d) 2.5 250 200 200 50 150
Hydrogenation temperature (~ 370 450 410 460-470 450 455
Pressure conditions (MPa) 18 20 15 30 15 17
Yields (%coal)
Heterogas 11.7 12.5 12.0 11
Hydrocarbon gas 6.4 10.5 5 21
Naphtha 8.4 13.3 14 16
Mid-distillate 6.4 19.7 10 36
Heavy oil 40.0 31.0 11 22
Residues (ash and 33.0 37.1 48 16
unconverted coal)
Hydrogen consumption 3 4 4 6 5
wt% coal feed
Reference Newman Newman Neavel Langhoff NEDO/NBCL Wasaka
1985 1985 1961 1982 1994 1999b
It is assumed that the formation of methylindan and naphthalene, and the hydrogen
transfer from the tetralin, proceeded via a tetralyl radical, which acted as an intermediate in
the conversion and the hydrogen exchange of tetralin (Kabe et al., 1990d; Ishihara et al.,
1995; Godo et al., 1997b). In the system of tetralin and gaseous hydrogen, the conversion
into methylindan and naphthalene, and hydrogen exchange reaction of tetralin were not
changed by the reaction pressure, indicating that gaseous hydrogen does not affect the for-
mation of a tetralyl radical at least in the present reaction system using a flow reactor.
Tetralin may be activated by collisions and become a tetralyl radical and a hydrogen atom
according to Eq. (4.24). Some of the hydrogen atoms produced from tetralin would react
with tritiated hydrogen molecular, leading to the formation of the tritium atom, as shown in
Eq. (4.50). If the tetralyl radical is quenched by tetralin to form original tetralin, the con-
version of tetralin and hydrogen exchange would be inhibited. The reaction of the tetralyl
radical with the tritium atom leads to hydrogen exchange in the reproduction of tetralin, as
shown in Eq. (4.57). The hydrogen exchange between tetralyl radicals and tritiated hydro-
gen (Eq. (4.46)) can proceed via radical hydrogen transfer reaction depending on the con-
centration of tetralyl radicals, which also controls the formation of methylindan and naph-
T
@ + .T 9 @ (4.57)
+ H. - ~ (4.58)
thalene in Eqs. (4.26) and (4.58). Therefore, the formation of the tetralyl radical by unimol-
ecular scission in this system may be the rate-determining step for both the conversion of
tetralin into methylindan and naphthalene, and the hydrogen exchange. At 400 ~ the con-
version and hydrogen exchange ratio of tetralin were very low. If coal is included in the
262 4 Liquefaction of Coal
system, a tetralyl radical may be formed easily. However, if coal which forms a radical is
not included in the system, a tetralyl radical is difficult to form. Then the unimolecular
scission of tetralin barely occurs at 400 ~
Hooper et al. (1979) reported that butylbenzene was formed by the thermal dissociation
of tetralin, as shown in Eq. (4.59). While the plot of the yields of butylbenzene showed
some scatter, the effect of the reaction pressure was recognized at every temperature. An
alternative route to form butylbenzene appears in Eq. (4.27), where tetralin reacts with hy-
drogen atom.
4.5 Process of Coal Liquefaction
Coal liquefaction processes generally have as the principal objective the manufacture of
transport fuels, and many have the ambitious target of producing premium fuels that can be
directly substituted for the pump grades currently obtained from petroleum. The same phi-
losophy of direct substitution can be adopted for the petrochemical industry using feed-
stocks for the massive aromatics/olefins plants currently producing most of the precursors
for today' s synthetic materials. Methods for the direct liquefaction of coal have been devel-
oped in a number of countries and the processes in the United States, Germany and Japan
are described here.
4.5.1 C o a l L i q u e f a c t i o n P r o c e s s e s in the U S A
A. Solvent Refined Coal (SRC-II)
This process was originally developed by the Pittsburg and Midway Coal Company for the
production of pure carbons from coal. It was later adapted as a method for desulfurizing
the coals that were available from vast reserves in the US Appalachian coal basin. The
American Clean Air Act of 1977 limited the sulfur content in the Appalachians to 2-5%
sulfur. The SRC product is a pitch-like product and needs further hydrotreatment to make
useful distillates. In the latest version of the SRC process being developed by the Catalytic
Corporation (Fossil Energy, DOE/PC/50041-(79)) at their Wilsonville pilot plant, the SRC
primary product is being hydrocracked to produce low boiling distillates. A line diagram of
this process is shown in Fig. 4.64. Briefly, pulverized coal, slurried with recycle oil, is
Ash
Hydrogen -~ concentrate
I
Slurry High pressure Light solvent Critical solvent
Coal
preparation hydro-extraction recovery de-ashing
I I Solvent
recycle
I
Clean SRC solution
Hydrogen Gases
I
I Catalytic Product I Distillate
hydrocracking solvent
product
J ebullating bed separation I
,qm
Recycle solvent
I
Fig. 4.64 Two-stage SRC coal liquefaction. [Reproduced with permission from Davies, G.O. et al., Critical
Report on Applied Chemistry Volume 9 Chemicals from Coal: New Developments, 102, Backwell Sci.
Pub., (1985)]
4.5 Processof Coal Liquefaction 263
passed with high pressure hydrogen to a digestor where the coal is almost totally dissolved.
After separation of the gas for recycle the digest is cleaned by critical solvent de-ashing.
After de-ashing the SRC material is hydrocracked in an ebullating bed of a catalyst. A typi-
cal set of product yields from Illinois # 6 coal is given in Table 4.19. In the process, a por-
tion of ash containing pyrite (FeS2) as a catalyst with recycle oil was cycled.
B. H-coal Process
This process was developed by the Hydrocarbon Research Incorporation from the H-oil
technology used commercially for the beneficiation of heavy petroleum residues. The heart
of the process, illustrated in Fig. 4.65, is the ebullating bed hydrocracker where a slurry of
coal in recycle oil is catalytically reacted with high pressure hydrogen. The high active Co-
Mo particle catalyst prepared was used. The up-flow of the slurry expands the catalyst bed,
while the flow of hydrogen induces a washing-machine mixing action. The net result is
free movement of the catalyst pellets that prevents plugging by ash and carbon deposits,
produces isothermal reaction conditions and facilitates the addition and removal of catalyst
in the reactor. The hydrocracked products pass to a separation system where light distil-
lates are recovered as product and the heavier products are split by vacuum distillation into
heavy oil for recycle and an ash-laden pitch that can be used as a fuel for power generation.
The process was developed at the Trenton Laboratories with units processing up to 3 ~ t of
coal per day (Comolli et al., 1978). Data from these units were used to design a 250 t/d
plant that was built at Catlettsburg, Kentucky. This large plant has been operating since
1981 and as the work has been sponsored by the US government. The results have been re-
ported in a number of US Department of Energy documents (Fossil Energy, DOE/ET
10143-19/37). A yield summary from a typical run at Catlettsburg using a Kentucky coal
is given in Table 4.19.
Reactor
Coal (f " Naphtha
__~ Hydrotreating
and
r
reforming E:• Gasoline
chemicals
Naphtha ?
~ Hydrotreating
::g /t Recycle I , and ::::::::.~ Mid-distillates
tube "1 i distillates
reforming
Lt Jj
Ammonia(9 kg)
High-Btu-Gas
Hydrogen (3200cu.ft.)
Fig. 4.65 H-coalprocess. [Reproducedwith permissionfrom Davies, G.O. et al., Critical Report on Applied
Chemistry Volume 9 Chemicals from Coal: New Developments, 105, Backwell Sci. Pub. (1985)]
264 4 Liquefaction of Coal
Catalytic
[ Solvent
i hydrogenation Solvent
I
Coal _~ Slurry ~_~
t
H2
Gas
preparation preparation
7"
H2 ~ Liquefaction ~ Distillation - Liquid
products
I Heavy
bottoms
slurry
H20 _1
Air ~[ Flexicoking
-I
I ~ Fuel gas
Ash residue
Fig. 4.66 Exxon EDS process. [Reproduced with permission from Davies, G.O. et al., Critical Report on Applied
Chemistry Volume 9 Chemicals from Coal: New Developments, 111, Backwell Sci. Pub. (1985)]
4.5.2 C o a l L i q u e f a c t i o n P r o c e s s e s in G e r m a n y
German coal liquefaction processing is being developed by Ruhrkihle Oel und Gas Gmbh,
and is based on the original liquefaction work (IG process) by Bergius and Billiviller
(1918) and Pier (1929). The process (New IG process) has been described by Peters (1978)
and Romey (1981) and has reached a large pilot plant (250t/d) scale of operation at the
Bottrop plant near Essen (Langhoff et al., 1982). The process is illustrated in Fig. 4.67.
Briefly, coal is slurried in heavy recycle oil and is passed with high pressure hydrogen
(30 MPa) and a cheap throw-away catalyst (Fe-sulfur catalyst) to a simple tubular reactor
where, at high temperature (460 ~ in the presence of high pressure hydrogen, the coal is
hydropyrolyzed to produce distillates. The distillates are fractionated in a series of pressure
let-down vessels and a vacuum distillation is used finally to separate the unreacted coal and
ash from the heavy oil needed for recycle. Typical yields from the Bottrop plant are given
in Table 4.19. The plant at Bottrop has been operating since November 1981 and it was re-
ported by Langhoff (1982) that more than 1000 h of operation had been achieved by March
1982. The upgrading of the Kohloel syncrude is being investigated by VEBA OEL in
bench-scale units. The quality of the various distillate products is similar to those obtained
from SRC and H-coal syncrudes.
4.5 Process of Coal Liquefaction 265
Coal + catalyst
Gas
,..._
"- I
I [ Light oil
txl ~ Middle oil
~00 bar
i
~75 ~ [ ~ Heavy oil
Preheater , I Vacuum
"-~--'Zt I L istillati~
Reactors ]
r 1
Heavy
Hydrogen I Gasification I residues+ash
Ash ~ I
Recycle oil
Fig. 4.67 German coal liquefaction process. [Reproduced with permission from Davies, G.O. et al., Critical
Report on Applied Chemistry Volume 9 Chemicals from Coal: New Developments, 113, Backwell
Sci. Pub. (1985)]
4.5.3 C o a l L i q u e f a c t i o n P r o c e s s e s in J a p a n
The Agency of Industrial Science and Technology (AIST) of the former Ministry of
International Trade and Industry (MITI) of Japan started the Sunshine Project in 1974 to
develop new energy technologies (Yoshida, 1997; Office of New Sunshine Project, 1995).
The Moonlight Project was started in 1978 to develop energy conservation technologies.
Both projects successfully maintained Research and Development (R&D) schedules for en-
ergy subject areas with the close cooperation of industry, government and academic organi-
zations. These projects have been steadily providing effective results for basic technology,
practical applications and application to peripheral fields. The New Energy and Industrial
Technology Development Organization (NEDO) has been established to develop coal liq-
uefaction technology. NEDO started basic research for coal liquefaction development in
1980 and subsequently developed the NEDOL coal liquefaction process (Wasaka, 1999a;
Wasaka and Ibaragi, 2000). The economics study of the NEDOL coal liquefaction plant,
for commercial scale-plants, was conducted after a 150 t/d pilot plant operation had been
completed. For coal liquefaction technology, R&D is being carried out on brown coal and
bituminous coal liquefaction technology. These are technologies for liquefying coal and
manufacturing transportation fuel that can be used in place of petroleum.
Recycle solvent
Secondary hydrogenation
Ballmill [_..... Recycle gas
I Slurry ~ ~c~ Lightoil )
making I I rePrSs:iU~g1
CLB recycle feed pump
~ SepTat~
Solvent recycle from ATM fractionator
C~ I Settler Fixedbed reactor
I~ Sludge Fractionator
Hydrotreatedde-ashedoil (HDAO) recycle CLB run down I
Fig. 4.68 Brown coal liquefaction (BCL) process. [Reproduced with permission from Shimasaki. K. et al., J.
Jpn Inst. Energy, 78, 809 (1999)]
the analysis of the results obtained by the operation of the pilot plant, the conceptual design
of a 6000 t/d demonstration plant and an economic evaluation of the process were complet-
ed by the end of FY 1993 (NEDO/NBCL, 1994). Typical yields are given in Table 4.19.
Atmospheric 9Gas
Separators~ _ ,., tower
Catalyst I
Coal ~ Hydr~
Reactors ~ L
~V~ ~--~
j --Naphtha
- Gas oil
Pulverizer I Letdown
-4-~l-~JPreheater valve Vacuumtower I . Residue
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . j
- Naphtha
Hydr~ ~ara ~ "~-~
solvent i tors
~ ~ ~ IHydr~
Stripper i l-_C..'), Recycle
" ] ~ T ,,l,:~! solvent
! Reactor Preheater
9. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . j
Solvent hydrogenation
Fig.4.69 A flowdiagramof the NEDOLprocess. [Reproducedwithpermission
fromWasaka.S. et al.,J. Jpn. Inst. Energy, 78, 802 (1999)]
in many countries the overall cost of producing light distillates from coal is still consider-
ably higher than their production from petroleum.
There are numerous problems to be solved when considering future commercial coal
liquefaction processes. First, there is the matter of hydrogen production. Much hydrogen is
consumed in any coal liquefaction process corresponding to the production scale and the
nature of the liquefied product. The issue of hydrogen manufacture in the process remains
open to debate. Most of the methods currently under consideration involve the partial oxi-
dization of the coal or use of residues after liquefaction.
Among the conversion processes, the liquefaction process is most heavily influenced
by the nature of the coal. However, there exists controversy regarding the chemical and
macromolecular structure of coal despite recent advances, as described in Chapter 2. Thus,
the exact roles of catalyst and solvent in coal liquefaction remain ambiguous. Design in the
coal liquefaction process is based mostly on experience.
This Page Intentionally Left Blank
5
Gasification of Coal
5.1 Introduction
Coal may become the major energy source among fossil resources in 21st century because
of its abundantly availability throughout the world (Mochida and Sakanishi, 2000). Even
though coal has a relatively high heating value of approximately 8000-14000 Btu/lb, its
solid state has been a major reason for its inconvenience of use as a fuel. Also among the
fossil resources, coal has the highest contents of carbon and contaminants, such as sulfur,
nitrogen and other trace elements, which release global warming gases and pollutants such
as CO2, NOx, SOx and Hg. The efficient and clean utilization of coal remains a challenging
task and continues to be pursued extensively for its longer availability and least load on the
global environment. With coal as the largest energy reserve left, energy specialists are
banking on coal gasification to meet the energy needs of the future.
The term coal gasification has been given to any process in which coal or char is react-
ed with an oxidizer at high temperature to produce a fuel-rich product. The aim is to
achieve the highest possible conversion of the carbon present in the starting material.
Exceptions are those processes that need coke to meet the energy requirement. Most coals
can be gasified in relatively high yields; however, younger coals, lignite to slightly caking
hard coals, are preferentially employed. Air, oxygen, steam, carbon dioxide, or mixtures of
these components are usually employed in the process as gasifying agents. Carbon monox-
ide, hydrogen and methane, which are combustible gases, are the predominant products of
gasification.
Coal gasification technology can be utilized in four energy systems of potential impor-
tance:
(a) Production of gaseous fuel for electricity generation.
(b) Manufacturing substitute natural gas for use as pipeline gas supplies.
(c) Production of synthesis gas for subsequent production of alcohols, gasoline, plas-
tics, etc.
(d) Generation of gaseous fuel (low or medium Btu) for industrial purposes.
Mixtures of carbon monoxide and hydrogen at various ratios in the total gas mixture
are necessary for many of the syntheses. Chemicals to be directly processed from synthesis
gas as well as other utilization of synthesis gas are illustrated in Fig. 5.1. This figure in-
cludes processes that have already been commercialized or are currently under develop-
ment.
The rates and degrees of conversion for various gasification reactions are functions of
temperature, pressure, gas composition and the nature of the coal being gasified. The rate
of reaction is always higher at higher temperatures, whereas the equilibrium of the reaction
may be favored at either higher or lower temperatures. The effect of pressure on the rate
depends on the specific reaction. Gasification reactions like the carbon-hydrogen reaction
270 5 Gasification of Coal
--IIMethane II
_1
v I Ethylene I
"J] Pr~ lene [
~[I Methanol II
v_lI Ethanol I
Coal jI l Coal I
~1 Gasification I ~lll Butanol 11
1
-I Ethylene glycol I
~l Acetic acid [
-I I
=Ill Fuel oil II
-~]1Ammonia/urea [I
Commercialized
~[I Reducing gas [I
I I
Under development ~111Electricity II
Fig. 5.1 Utilization of synthesis gas. [From JICA. I, 207 (1989)]
to produce methane are favored at high pressures (70 atm) and relatively low temperatures
(760-930 ~ whereas low pressures and high temperatures favor the production of carbon
monoxide and hydrogen.
The basic technical processes of coal gasification are distinguished into (1) gasification
in the stationary or slowly moving fixed-bed (fixed-bed gasification); (2) reaction in the flu-
idized-bed; (3) coal gasification using the entrained-bed principle; and (4) reaction in a
molten bath of salt or metal (molten-bath gasification). Another important method of clas-
sification is by the applied combustion and gasification temperature, and whether the ash
from the solid fuel is discharged in liquid form (slag) or as a solid (ash).
Supply of heat is an essential element in the gasification. The heat supply can be dis-
tinguished between autothermal and allothermal processes, as schematically described in
Coal Coal
Oxygen I
Gas Oas
Steam Steam
Ash Ash
autothermal allothermal
Fig. 5.2 Performance of heat supply in autothermal and allothermal gasification.
[From JICA. I, 207 (1989)]
5.2 Productionof Gases InvolvingGasification 271
Fig. 5.2. In the authothermal process, the heat for the endothermic gasification is supplied
by the combustion of part of the feedstock with oxygen-containing gasifying agents, while
in the allothermal process the heat necessary to achieve process temperature levels is sup-
plied by external sources. The autothermal processes are subdivided according to whether
the reactants are fed cocurrently or countercurrently. The essential advantage of the latter
method lies in the more favorable heat utilization. One drawback is the mixing that occurs
in certain circumstances of the gasification agent gas with the product gas, which contains a
relatively large amount of methane and higher hydrocarbons. On the other hand, the al-
lothermal processes can be characterized according to the way in which heat is introduced:
via heat transport with solid or gaseous heat carriers, via heat conductance through the
walls of the reaction chamber, or via an immersion heater. In the last case, a heat exchanger
is present in the reaction chamber. However, this type of heat supply is technically very in-
volved.
The type of coal being gasified is also important to the operation. Only a suspension-
type gasifier can handle any type of coal; if caking coals are to be used in a fixed- or flu-
idized-bed, special measures must be taken so that the coal does not agglomerate during
gasification. In addition, the chemical composition, volatile content and moisture content
of coal affect its processing during gasification.
70 /
/
/
60- CO~.
X ~ - ~i ~ ' ~ ] /
// I: Fluidized bed
50 ,, I" (Winkler)
1 /
40 I
/
/ ,~I2 / ,, II: Entrained
Suspension
(Koppers)
20 ,'
10~( 5
60 ~'~ CO/~
I f
" C02
/'CH2>~.. "~..-.<" ~ ,
70
60 -~ }0
b~ar . i //f f
I /"
50 CO ,./
"" /,H20 ' / ,"
40 xqr" / H2
3O c02 \ "\ / / /
~ .--.~--------,--.-.~
20 ~ ' - - ~
10
j~.1 lj CH4
600 800 1,000 1,200
Equilibrium temperature (~
Fig. 5.3 Gas equilibrium composition at several temperatures.
[From JICA. I, 182 (1989)]
The first five reactions are the most prominent in the gasification processes. Reactions
(5.3) and (5.4) are combustion of carbon which is produced by the pyrolysis of coal at the
primary gasification. These reactions are exothermic.
In gasification, steam is usually added as a reactant to control high temperatures result-
274 5 Gasification of Coal 9
ing from combustion of carbon and to increase the heating value of the product gas, as
shown in Eq. (5.5). This reaction, which is called "water gas reaction", is endothermic, and
must rely on the heat release from the combustion of carbon for energy requirements.
Further, the gas composition at equilibrium at above 1100 ~ (Fig. 5.3) shows carbon
monoxide and hydrogen as dominant components, while at lower reaction temperatures,
there is more carbon dioxide.
In order to produce a fuel-rich gas and to consume the remaining carbon, a much slow-
er endothermic reaction Eq. (5.6), must occur. This reaction occurs preferentially at higher
temperatures of above 900 ~
In a coal/steam/oxygen system, homogeneous reactions (with one another) of the ini-
tially formed gas components are also very important, e.g., Eqs. (5.7) and (5.8). The first of
these reactions causes the rapid consumption of oxygen, increases gas temperature and
forces the requirement for slow, heterogeneous carbon reaction with CO2. The second,
slightly exothermic water-gas shift reaction also produces CO2 from CO and tends to con-
trol the final product distribution.
The methanation reactions (the last two reactions) favor high pressure and low temper-
atures (see Fig. 5.3), but in most cases a methane concentration higher than equilibrium is
predicted because methane is also formed during devolatilization. Methane formation in-
creases the thermal efficiency of gasification and the heating value of the product, but it is
not favorable for the production of synthesis gas.
The rate of trace elements such as sulfur and nitrogen is also important. Sulfur in coal
is converted primarily to H2S under the reduction conditions of gasification. Approximately
5% to 15% of the sulfur is converted to COS, somewhat complicating the gas clean-up step.
High temperatures and low pressure are favored in the conversion of nitrogen in coal to N2,
while the opposite conditions are favored in the conversion of some of the nitrogen to NH3.
Small amounts of HCN are also formed. Tars, oils and phenols, if they are not destroyed,
contain some of the oxygen, nitrogen and sulfur from the coal as more complex molecules.
The course of the gasification process cannot be predicted on the basis of thermody-
namic data alone. Kinetics play a major role and are in turn heavily influenced by material
and heat transfer phenomena. Fig. 5.4 shows the consecutive steps in simplified form for
HzO G
~.f Diffusion via hydrodynamic
CO H2 boundary layer
, Boundary layer
9 <.2..(...
Chemical reaction
C + H20 = CO + H2
Fig. 5.4 Transport mechanisms in coal gasification. [Reproduced with permission from B. Cornelis,
Chemicals from Coal : New Processes, 9, John Wiley & Sons. lnc. (1987)]
5.4 GasificationProcesses 275
1200 ~
III II I
1/T
Fig. 5.5 Three characteristic temperature ranges for reaction between gases and porous solids.
[Reproduced with permission from B. Cornelis, Chemicals from Coal : New Processes, 10, John Wiley
& Sons. Inc. (1987)]
the gasification process. The mass transport of gasification agents to the coal grain and re-
moval of gasification products from the coal grain occur simultaneously on the solid coal
through the diffusion-retarding boundary layer (step 1), pore diffusion (step 2) and the
chemical reaction of the actual gasification process (step 3) (Cornelis, 1987).
The rate-limiting step in the coal gasification can be chemical (adsorption of the reac-
tant, reaction, desorption of products) or gaseous diffusion (bulk phase or pore diffusion of
reactants or products). It has been established that the reaction between gases and porous
solid involves three temperature zones (Cornelis, 1987), as illustrated in Fig. 5.5. In zone I,
which occurs at low temperatures below approximately 1000 ~ the process is very heavily
dependent on temperature because of the temperature dependence of the chemical reaction
(step 3). Above 1000 ~ to about 1200 ~ (zone II), the activation energy is lowered to
half, i.e., it is characterized by control of to both chemical'reaction and pore diffusion. At
the higher gasification temperatures the rate of gasification is only slightly dependent on
temperature. In this zone, the reaction only occurs on the surface of the particle due to the
fast intrinsic reaction. The gasification agents (oxygen and water) do not diffuse signifi-
cantly into the particle and the reaction is controlled by diffusion through the boundary lay-
er on the particle. This has three consequences for the technical control of the gasification
reaction:
i) reduction in grain size, that is, the use of finely grained coal dust, where increase in
specific surface leads to a particularly rapid reaction;
ii) rise in the gasification pressure increases productivity of a given reactor volume;
iii) reduction of the diffusion-retarding boundary layer is possible by increasing the
relative velocity between gas and coal grain.
The technical control of the gasification process should allow for this and permit the re-
quired high relative motion of coal and gasification agent by the selection of a suitable gasi-
tier.
Cial \ Gas
Steam & __
:_.,~-~'._/~7- oxygen
I I I
Steam 500 1000 1500
oxygen
Ash Temperature (~
Fig. 5.6 Principleof the fixed-bed gasification process.
[FromJICA. I, 183 (1989)]
Lurgi gasification. The most important fixed-bed process is the Lurgi gasifier. The
older version of the Lurgi process is a dry ash gasification, which significantly differs from
the more recently developed slagging process.
A schematic of a Lurgi gasifier with its equipment is depicted in Fig. 5.7. The un-
washed, but graded, coal of the preferred granulation (5-30 mm) is introduced batchwise
into the pressure chamber of the gas generator via a pressure lock. At this point, the coal is
evenly distributed over the cross section of the reactor by means of the coal distributor, the
supply of coal to which is so great that the actual gas generator is continuously charged.
Caked coals are broken up again in the caking zone by the vertical movement of the wedge-
shaped water-cooled stirring arms located at the coal distributor. Outlet conduits that ex-
tend from the coal distributor to the bottom ensure sufficient space for the vertical move-
ment of the charge.
The gasification takes place at 30 bar with an oxygen-steam mixture that is injected
5.4 Gasification Processes 277
2. Gasifier condition
a. Operating temperature (~ 700-900 1,500 1,000
b. Pressure (bar) 20-30 20-30 Atmospheric
c. Ash removal type Dry Slagging Dry
3. Gas properties
a. Composition (vol.%-dry)
CO 19.0 61.0 28.6
H2 39.0 28.0 15.0
CO2 30.0 2.6 3.4
CH4 11.0 7.6 2.7
N2 & others rest rest rest
b. Heating value (kcal/Nm 3) 2,247 3,290 1,500
4. Gasifying medium Air/oxygen & steam Oxygen & steam Air & steam
into the reactor through the rotary screen. The bed of ash lying on the screen serves to uni-
formly distribute and preheat the gasification agent. The necessary sensible heat for the en-
dothermic gasification reaction is produced in the brief gasification zone. The coal gasifi-
cation and degasification zones follow. The resulting raw gas then dries (in countercurrent)
the coal moving to the bottom. The solid remaining ash is discharged from the reactor by
means of a lock.
In fixed-bed gasification, coal granules pass through the following zones in the gas pro-
ducer: drying zone, degassing zone, gasification zone and combustion zone. Coal dust and
tar are next removed from the raw gas exiting the reactor by means of a water scrubber; the
gas is then further cooled in the waste heat boiler.
5.4.2 Fluidized-bed Gasifier
This gasifier allows intimate contact between gas and solids and provides longer residence
times. It uses finely pulverized coal particles (0.2 to 0.4 mm) and the gas flows up through
the bed. It exhibits liquid-like characteristics. Due to the ascent of particles and separation,
a larger coal surface area, promotes chemical reaction. Dry ash is removed continuously
from the bed. The gasifier is operated at such a high temperature that it can be removed as
agglomerates, which results in improved carbon conversion. Such beds also have limited
ability to handle caking coals. Fig. 5.8 shows the process principle for a fluidized-bed gasi-
fication and Table 5.2 shows the typical performance of several fluidized-bed gasifiers.
278
Coal
I
Tar recycle
Drive ~ / x ~ / ~
~ ~ - - ~ <-'-'//-'~ Steam
~,1 z...._ff , [ , , ~ , _ _ / ~ Washing
' ' cl/x Itondenser
Screen ~
drive Water jacket
Ste
oxygen / Ash
lock
r-'~_____Gas Gas
Char
Coal
(< 0C5almm)"S~;~~ Oxygen &
( i i
2. Gasifier condition
a. Operating temperature (~ 900-1,000 1,050 980
b. Pressure (bar) 20 7-20 80
c. Ash removal type Dry Agglomerate Dry
3. Gas properties
a. Composition (vol.%-dry)
CO 15.1 22.0 25.0
Ha 15.8 14.0 30.0
CO2 12.1 6.0 25.0
CH4 3.6 3.0 19.0
N2 52.7 Rest Rest
Others 0.7 Rest Rest
b. Heating value (kcal/Nm 3) 1,280 1,370 3,270
4. Gasifying medium Air & steam Air/oxygen & steam Oxygen & steam
5. Specific consumption
a. Oxygen (Nm3/t-coal) Air 1800 NA 200-225
b. Steam (t/t-coal) 0.71 NA 1.1
8. Process developer CMRC Inst. of Gas Tech. (IGT), Inst. of Gas Tech. (IGT),
EPDC Chicago, Illinois USA Chicago, Illinois USA
NEDO Japan
|
|
!
|
|
i
|
|
:. -". 9r..- ".. |
|
|
|
. Gas Gas',
Ash
I I
and Table 5.3 shows the typical performance of several fluidized-bed gasifiers.
Texaco Process. Fig. 5.10 is a schematic process flow of the pressurized, down flow,
entrained Texaco gasifier. The feed coal is carefully pulverized to a certain grain size for
the coal dust gasification, then suspended in wet rod mills (mixing ratio of 1.5-0.8, depend-
ing on the coal). The slurry water consists of recycled condensate from raw gas cooling to-
gether with make-up water. Carbon that is not converted in the gasifier can be recovered
and recycled to the gasifier feed via the slurrying operation.
The coal/water slurry is fed to the reactor by means of high pressure pumps and enters
the reaction chamber via the annular gap of the burner together with oxygen. The gasifica-
tion takes place rapidly at temperatures over 1260 ~ under which conditions the coal is
converted primarily to H2, CO and CO2 with no liquid hydrocarbon being found in the gas.
The water in the coal slurry not only serves to convey the coal to the gasifier, but also mod-
erates the gasifier temperature so that excessively high temperatures are not experienced.
The crude raw gas leaving the gasifier at 1260-1480 ~ contains a small quantity of
unburned carbon and a significant portion of molten ash. Depending on the end use, this
gas stream can either be directly quenched in water (to cool the gas and remove solidified
ash particles) or be cooled in a radiant and convection boiler for sensible heat recovery (via
high-pressure saturated steam generation) prior to water scrubbing.
The raw gas initially releases some of the heat in the waste heat boiler. This heat is
used for steam generation. The solid particles present in the raw gas are removed by scrub-
bing. Nongasified or partially gasified fuel is recycled to the process. The pretreated raw
5.4 Gasification Processes 281
~
Oxidant
Makeup
asifier High-pressure
water
steam
[ I .~Convection Particulate-free
CoalI - -[...k...__..._d
- - - - -Mi-l- ~ I------,[ , k,.~...,) : , cooler synthesis gas
Recycle
m [~_/J
I Radiation
]cooler,
]
I
[
~
scrubber[ [
L__.__...A I
solids SlurrYtankT Slurrypump~l) ~ ]'~-i'J
-.-- r - - ~ '['-~
gas is then cleansed of carbon dioxide and hydrogen sulfide (including carbon oxysulfide)
in a chemical gas washing. The resulting CO: H2 ratio is always somewhat higher than 1:1.
It depends on the feedstock employed: highly carbonized fuels produce a considerably
greater share of carbon monoxide. The process is basically suitable for all solid fuels, a
particularly attractive feature being that the gasification plant can also be operated with oil
or natural gas.
5.4.4 M o l t e n B a t h Gasifier
In this gasifier, coal is fed together with steam or oxygen into a molten bath of salt or metal
operated at 1000-1400 ~ Ash and sulfur are removed as slag. The MIP, Kellogg and
ATGAS processes are examples of this type of gasifier. Fig. 5.11 shows the principle of
the molten iron bath process, and typical performance characteristics of this process are
shown in Table 5.4.
Molten Iron Bath (MIP) Process. Coal gasification in a molten iron bath is a relatively
new process, the logical extension of basic oxygen furnace (BOF) steel-making technology.
The most promising of the development efforts is the Sumitomo/KHD cooperative pro-
gram. Sumitomo Metal Industries Ltd. (SMI), Japan, and KHD Humboldt Wedag AGI
Germany, started the development individually, SMI with the top-blowing technology and
KHD with the bottom-blowing technology. The top-blowing technique provides for oxy-
gen to be blown onto the iron bath at supersonic velocity through a specially designed lance
together with CO2 as the carder gas. As in the BOF process, a dimple is created at the point
where the jet hits the bath surface. This is where 02 and coal are blown and come into con-
tact with the molten iron. The reactions of injected oxygen and coal at the dimple are spe-
cific features of this process. Gasification predominantly occurs within the iron bath. The
coal impinging on the high temperature molten iron is cracked almost instantaneously into
the carbon dissolving in the molten iron bath.
The gasification proceeds very rapidly with the formation of CO by the reaction of oxygen
and steam with the dissolved carbon. It is estimated that the following reaction mainly
takes place inside the molten iron bath.
Synthesis
Woler
Oxygen
steam
Feed
system
Transporting
gas ,Gasifying vessel
(Reactor)
Fig. 5.12 Molten iron bath coal gasification. (MIP Plant) [FromJICA. I, 197 (1989)]
er content of sulfur compounds in the gas and zero tar formation are decisive advantages of
the MIP coal gasification technology.
1 dW 1 dx
r . . . . (5.16)
W dt ( 1 - x) dt
where r is the normalized gasification rate, t is time, W is the char mass at time t (dry ash
free, daf), x is char conversion,
Wo-W
x = ~ (5.17)
Wo
1 dW dx
rP ~ _ _ _ _ -- _ _ (5.18)
Wo dt dt
When considering the reactivity of coal gasification, the behavior of the conversion
versus time (x versus t) and that of the reaction rate versus conversion (r versus x or r" ver-
sus x) are studied as char is reacting with CO2 or H20 at gasification temperatures (800-
1200 ~
In general, the curves of conversion against time show a similar shape, regardless of
coal type and experimental conditions, as can be seen from the literature (for example, Lee
and Kim, 1995; Shufen and Ruizheng, 1994; Kasaoka et al., 1987; Yang and Watkinson,
1994; and Alvarez et al., 1995). The curve is almost a straight line up to x -- 0.75.
On the other hand, there is no agreement regarding the variation of the gasification rate
with conversion. Some authors have found a maximum rate for conversions between 20%
and 60% (Dutta et al., 1977; Yang and Watkinson, 1994; Chi and Perlmutter, 1989;
Gavalas, 1980; Bhatia and Permutter, 1980; and Hamilton et al., 1984). Others found a lin-
ear decrease in gasification rate as reaction advances (Chin et al., 1983; and Adschiri et al.,
1986). Finally, some studies regarding coal reactivity during gasification (for example, as
given in studies by Matsui et al., 1987, and Schmal et al., 1983) do not consider gasification
rate variation with conversion and only report the gasification rate at a specific value of
conversion.
the reactivity of coal char or carbon in gasification and oxidation reactions. The technique
used to determine the concentration of carbon active sites is oxygen chemisorption. The
following discussion is aimed to review the importance of this technique for monitoring the
reactive sites in coal char and provide an understanding of the detailed interactions between
gas and solid.
c(o) ~ c o (5.21)
Cf--F- C(O)--F- 02 ~ C02 ..ql_C(O) (5.22)
c(o) + c(o2) ~ co2 + c(o) (5.23)
2C(O) ~ CO2 + Cf (5.24)
where Cf is a free active site on carbon, presumably edge carbon, and C(O2) and C(O) rep-
resent molecular and atomic oxygen adsorbed on a site.
Generally, each researcher has proposed that the reactivity is proportional to the ASA
measured by his own method, although the ASA is usually measured at temperatures lower
than the reaction temperature. Cheng and Harriott (1986) attemped to clarify the relation-
ship between chemisorption rate and reactivity. Fig. 5.13 shows the Arrhenius plots of
gasification rates and chemisorption rates measured at temperatures lower than the reaction
temperature. It can be seen that the controlling step of the gasification reaction changes
287
9 k, at high temperature
9 k, at low temperature
zx k adsorption
tD
0
\
tD
t..,
r~
\
~D
0 \ ~k A
tD \A
\
A
\
A
i i i i'
Fig. 5.13 Arrhenius plot of O2 gasification rate and adsorption rate of O2 for a demineralized active carbon.
(Cheng and Harriott, 1986; modified by Miura. K. et al.) [Reproduced with permission from Miura,
K., et al., Fuel. 68, 1469, Elsevier (1989)]
o CALGON RC
A VW65
9 DARCO G60
9 MALLINKROTT. USP
xl [] AMOCO PX-21
9 NUCHAR SA
Tong et al.
9 6
eq
\!
•
9 m
,\:
6
Fig. 5.14 Arrhenius plot of O2 gasification rate per unit active surface area for several demineralized
active carbons. (Cheng and Harriott, 1986; modified by Miura et al.)
[Reproduced with permission from Miura, k., et al., Fuel. 68, 1470, Elsevier (1989)]
288 5 Gasificationof Coal
with temperature. The adsorption rate extrapolated to higher temperatures coincides with
the gasification rate: this is the region of adsorption control. The gasification rate at lower
temperatures is judged to be controlled by the desorption rate because the gasification rate
extrapolated to lower temperatures coincides with the desorption rate calculated from the
adsorption rate and equilibrium constant. From this consideration, the relation between re-
activity and chemisorption was clarified. Cheng and Harriott (1986) replotted their data as
an Arrhenius plot of the gasification rate per unit ASA for several samples, as shown in Fig.
5.14. This figure shows that the difference in reaction rate among the samples is within two
to three fold when the rate is normalized by ASA. Thus ASA seems to be a good index of
reactivity for noncatalytic gasification. However, the ASA in Fig. 5.14 was evaluated at
temperatures below the reaction temperature.
To employ ASA as a reactivity index, a reliable method for its estimation must be es-
tablished. Up to now, each researcher has employed his own method. Furthermore, the
distribution in the strength of active site may have to be taken into account, as Waters et al.
(1986) suggest. The effect of physisorption must be completely eliminated, because 02 ph-
ysisorption is significant even at 200 ~ (Allardice, 1966). Causton and McEnanay (1985)
suggested using the temperature programmed desorption method in v a c u o to measure ASA,
since this technique completely eliminates the effect of physisorption. They reported that
the value of ASA is constant at adsorption temperatures ranging from 100 to 250 ~ This
may be a reliable method to measure ASA for noncatalytic gasification. Unfortunately,
however, the relation between the reactivity and ASA was not discussed.
5.6.2 The Role of Oxygen Chemisorption in Catalytic Gasification Reaction
In the field of coal gasification, it is well known that salts of alkali and alkaline earth metals
as well as transition metals are active catalysts for gasification reactions. Hippo et al.
(1979) and Hengel and Walker (1984) investigated the catalytic effect of Ca using deminer-
alized lignites impregnated with Ca salts. Radovic et al. (1983a and b) confirmed that the
active species of calcium is CaO and that the reactivity decreases with the severity of the
pyrolysis conditions, and attributed this to the sintering of CaO. Furthermore, they found
that the turnover frequency per active CaO molecule was constant irrespective of the pyrol-
ysis conditions. The following mechanism was considered for gasification catalyzed by
CaO (Miura, 1989):
Cf + CaO + 02---) Ca02 -q--Cf(O) (5.25)
2 C f ( O ) ~ CO2 + Cf (5.30)
where Cb is the carbon atom on the basal plane of the graphite-like structure, and Ca02 is
5.6 Reactivityof Coal Char 289
assumed to be either the superoxide or peroxide of Ca. Eq. (5.28) represents the surface
diffusion of an oxygen atom adsorbed on the carbon of the basal plane to an active site. For
noncatalytic conditions, this cycle reduces to one similar to the cycle proposed by Ahmed
and Back (1985). Dispersion of CaO was estimated through measurement of the crystallite
size of CaO by XRD, but the use of chemisorbed CO2 was also proposed. Solano et al.
(1987) demonstrated that the amount of chemisorbed CO2 at 200 ~ is a satisfactory mea-
sure of CaO dispersion.
Alkali metals such as Na and K have long been known to catalyze the gasification reac-
tion of carbon. Attention has now focused on the role of adsorbed oxygen in catalytic gasi-
fication. The nature of the oxide at the carbon surface, in particular the ratio of oxygen to
metal, has been discussed in relation to the gasification mechanism.
As early as 1954, Sato et al. measured the amount of adsorbed oxygen on chars im-
pregnated with K and Na salts as part of a study of char combustibility. They reported that
atoms of K or Na act as active sites for oxygen adsorption, adsorption is dissociative and
graphite adsorbs less oxygen than chars even if it is impregnated with the same amount of
K or Na. This last observation suggests the importance of the edge carbon for the forma-
tion of active sites for oxygen adsorption on carbonaceous materials.
More recently, Yokoyama et al. (1980) gasified an activated carbon impregnated with
K2CO3 in a batch recirculating flow reactor to measure the amount of oxygen trapped by the
char during gasification. They found the gasification rate to be proportional to the amount
of oxygen trapped and theorized that a redox cycle between K and K20 promotes the gasifi-
cation. Ratcliffe and Vaughn (1985) proposed that the amount of CO2 chemisorbed at 300
~ corresponds to the number of active potassium sites. According to these investigators,
the number of sites determines the rate of gasification because the turnover frequency per
active site remains constant during gasification. In other words, the gasification rate is di-
rectly proportional to the amount of chemisorbed CO2. Hashimoto et al. (1986) investigat-
ed the steam gasification of a carbon black impregnated with Na and K salts and found that
the rate was proportional to the amount of oxygen trapped on the char. Their results are
shown in Fig. 5.15. From FTIR analysis, they argue that oxygen appears to be trapped as K-
O-C or Na-O-C on the surface. Several researchers claim the existence of a M-O-C struc-
ture from either direct measurement (Mims and Pabst, 1983) or O/M value (Delannay et al.,
1984; Sams and Shadman, 1986; Saber et al., 1986). Yuh and Wolf (1983 and 1984) also
reported the existence of K-O-C and Na-O-C bonds from FTIR measurements.
In addition to studies of the role of oxygen, several contributions Freriks et al. (1981),
Mims et al. (1982), and Cerfontain and Moulijn (1983) have been concerned with the iden-
tification of the active species in the alkali metal catalyst. The mechanism of alkali metals-
catalyzed gasification has been reviewed by McKee (1983), Kapteijn et al. (1984) and
Moulijn et al. (1984). The mechanism presented by Wigmans et al. (1983) seems the most
reasonable for steam gasification based on the observations noted above.
M + H20 ~ M ( O ) + H2 (5.33)
where M is K or Na, M(O) is the M-O-C structure, and C(O) is chemisorbed oxygen. A
290 3 Gasification of Coal
. . . . I . . . . I '
i
0 NaECO3 /
A K2CO3 / O ~
9 NaC1
._, 9 KC1 / " -
~'~'~ ~
x / v KNO3
A Ir -
t3/ZX
/
[] KOH _
v,tQ / ~ Pure
A/zxO
/
~-i i i i I i i i i I i
00 0.5 1.0
Oxygen trapped on the char
no(mol/kg-fixedcarbon)
Fig. 5.15 Relationshipbetween steam gasification rate of a carbon supporting alkali metal salts and amountof
chemisorbed oxygen at 30 ~ [Reproducedwith permission from Hashimoto, K. et al., Fuel. 65, 491,
Elsevier (1986)]
rate equation can be derived from this mechanism, in which the rate is proportional to the
amount of oxygen trapped by K, i.e., the moles of chemisorbed oxygen correspond to the
moles of active K (Hashimoto et al., 1986). On the other hand, Saber et al. (1986) proposed
a mechanism which assumes different reduced states of metal, as illustrated in the follow-
ing reaction steps:
M 2 0 - C + H20 ~ ( M O ) 2 - C q- H2 (5.37)
where M 2 0 - C and (MO)2-C are the reduced and oxidized forms of alkali catalyst respec-
tively. This mechanism is similar to the one proposed for CO2 gasification, in which oxy-
gen trapped by M accelerates the gasification reaction. Thus, the amount of chemisorbed
oxygen or CO2 represents an index to the reactivity of char when gasification is catalyzed
by Ca, K or Na.
Table 5.5 Change in Selectivity of Steam Gasification of Coal Chars and Carbon (s)
with Change in Reactivity for Catalyzed Steam Gasification
Steam P. Catalyst (s) Change in CO2/CO value
No. Sample (s) Reactor T (~ (atm) supported with increase of reactivity
10-1 I I I
o Coal (Taiheiyo)
- [] Coal (Sufco) -
9 Coal (Soldier
'~.~ Canyon)
10 -2 "~ ~k k [] Coal (Wa/larah)
"',,, ~ 9 Na on a carbon -
~ 5 zx'-,~ , black (CB)-
~ 2 -
~ ,, ~ K on CB
,, ~ -
o
,~ 10 -3
9
5
Data of Chin e
zx Active carbon ( A C ) " , ~ , -
-,\
9 K on AC ',
10 .4 _
Fig. 5.16 Arrhenius plot of CO formation rate for several samples during steam gasification.
[Reproduced with permission from Miura, K. et al., Fuel. 65, 411, Elsevier (1986)]
292 ~ Gasification of Coal
I I I
O Coal (Taiheiyo)
2 [- ~x [] Coal (Sufco) -
~ , , 9 ",,
..~ 10 -2 _ ", 9
2
~s,
10 3 -- ",
A AC [ Chin
~',, A 9 K on AC I et al.
5 - ,,
-~ Ptt2o = 1.0 atm
2- ", -
"a,
10 -4 I I I
9 10 11 12
1/TX 104 (K -~)
Fig. 5.17 Arrhenius plot of CO2 formation rate for several samples during steam gasification.
[Reproduced with permission from Miura, K. et al., Fuel. 65, 41 l, Elsevier (1986)]
Doraiswamy, 1983; Jenkins et al., 1973) to remove the effect of the gas phase reaction. In
general, the results indicate that the CO2/CO ratio increases with increasing reactivity. Rate
data for CO and CO2 formation, measured in differential reactors, are plotted in Arrhenius
type diagrams in Figs. 5.16 and 5.17, respectively (Miura et al., 1986). CO formation rates,
rco, are represented by a single correlation, irrespective of the coal, carbon or catalyst. Fig.
5.16 suggests a large activation energy for CO formation, but CO2 formation rates, rco2,
vary significantly depending on the source of coal and the catalyst or its loading (Fig. 5.17).
Furthermore, the slopes of the Arrhenius plots are smaller than those for re 9 From these
figures, it is concluded that CO formation proceeds by a noncatalytic reaction while CO2
formation is catalyzed by Ca, Na, K and other metals. On the other hand, the mechanism
presented by Saber et al. (1988) indicates that both CO and CO2 are formed catalytically
through Eqs. (5.38) and (5.39), respectively. Carbon dioxide, however, can be the main
product if the equilibrium of Eq. (5.39) favors the formation of CO2. Conclusions concern-
ing this question are complicated by the water-gas shift reaction, which produces CO2 from
CO. This shift is catalyzed by potassium on the carbon (Htittinger et al., 1986) and conse-
quently it is difficult to distinguish the contribution of the shift reaction from that of the cat-
alytic reaction (Eqs. (5.34) or (5.39)) even if a differential reactor is used.
ferences in performance between coals, rank-dependent parameters have not always been
found to be adequate.
Some authors have published work on the reactivities of various chars in gasification
with air (Jenkins et al., 1973), CO2 (Hippo and Walker, 1975), H2 (Tomita et al., 1977) and
H20 (Solano et al., 1979). They concluded that the reactivity of char generally increases
with a decrease in the rank of parent coal in an oxidizing gas atmosphere. However,
Takarada et al. (1983), in a study on the effect of a catalyst on different types of coals,
found that noncatalytic gasification reactivity of lower rank coals is not always larger than
that of higher-rank coals.
To provide some conclusive results on the effect of coal rank, it is necessary to collect
the gasification reactivity of many coals with a wide range of carbon content. This was also
done by Takarada et al. (1985a), who examined the reactivities of 34 coal chars of varying
rank in steam gasification. They found that the reactivities of chars derived from caking
coals and anthracites (carbon content > 78 wt%, daf) were very small compared with those
from noncaking (lower rank) coals. The reactivities of low rank coal chars did not correlate
with the carbon content of the parent coals.
Miura et al. (1989) reviewed the data in the literature for the rate of gasification of 68
coals with steam, CO2 and oxygen to clarify the factors controlling this process. Their re-
suits for char reactivity in gasification by steam, CO2 and oxygen versus carbon content in
parent coals are shown in Figs. 5.18, 19 and 20, respectively. The data for steam gasifica-
tion are very scattered for low rank coals (%C ~ 80), but the reactivity was higher than that
obtained for higher rank coals. When carbon content was more than 80%, reactivity data
were less scattered, but decreased as rank increased. For the CO2 and 02 reactions, the
O I ' I ' I
4- T= 800 ~
PH2O ~ 0 . 5 a t m
3A A
A
A
_ 0 OA 0
x 2
A
oo
0
I-
AA
o Cl 90
70 80 90
Carbon content in coal (wt%, daf)
Fig. 5.18 Relationshipbetween steam gasification rate and carbon content in parent coal.
[Reproduced with permission from Miura, K. et al., Fuel. 68, 1465, Elsevier (1989)]
294
T=900 ~
Pco: = 1 atm
15-
~2 10 - A A 9
• A
8
O
A
O x
5 A
A x
0
A0 x
9 0
o
,
0
60 70 80 90
Carbon content in coal (wt%, daf)
Fig. 5.19 Relationship between CO2 gasification rate and carbon content in parent coal.
[Reproduced with permission from Miura, K. et al., Fuel. 68, 1465, Elsevier (1989)]
14 I
T = 500 ~
12
OO
OO
10
9 9
Tr~ 8_0 9
O
x
6 -
O
9
4 - m
2 -
O
0 I
60 70 80 90
Carbon content in coal (wt%, daf)
Fig. 5.20 Relationship between 02 gasification rate and carbon content in parent coal.
[Reproduced with permission from Miura, K. et al., Fuel. 68, 1466, Elsevier (1989)]
5.7 Factors Affecting the Reactivity of Coal Char during Gasification 295
O Raw coal
Demineralized coal
2-8
9 9
_.---,
x 9
() TC
CT WD 0 /
0
70 80 90
Carbon content in coal (wt%, daf)
Fig. 5.21 Relationship of steam gasification rate of coal chars and demineralized coal chars with carbon content
in coal. [Reproduced with permission from Miura, K. et al., Fuel. 68, 1467, Elsevier (1989)]
296 5 Gasification of Coal
' I '
O Raw coal
Q Demineralized coal
,,..,0 MW
/'" JR OCT
!
I
l
I /
/ !
/
I ! /
I / /
f
I
/
I /~
/
TC I / /(..)
,-~ I / i
! / /
I / /
X I t !
k,
I
# I
'
I
, CT) RL
# / / I _
i I I I 0
, I I I
:1 I I
9 o ii/ / :
0 1 2
,(-2X 10 (kg/kg-coal, daf)
Fig. 5.22 Relationship of steam gasification rate of coal chars and demineralized coal chars with amount of mois-
ture adsorbed at 30 ~ [Reproduced with permission from Miura, K. et al., Fuel. 68, 1467, Elsevier
(1989)]
the reactivity of steam gasification decreases greatly with demineralization without affect-
ing the moisture holding capacity (.(2) of the parent coals, in other words, demineralization
should not change the pore structure of coal. Furthermore, the X-ray diffraction (XRD)
pattern was found to change little on demineralization (Hashimoto et al., 1987), indicating
that demineralization removes coal minerals without greatly affecting the microstructure of
coal.
The above result therefore suggests that the gasification reactivity of chars of lower
rank coals (C < 80%) is controlled by the catalytic activity of coal minerals. On the other
hand, the gasification rate of chars of higher rank coals seems to be the noncatalytic reac-
tion of carbon, controlled mainly by the intrinsic reactivity of the char. This suggestion
does not necessarily mean that gasification of the chars of higher rank coals is not affected
by the catalytic activity. If alkali metals or alkali earth metals are added to these coals be-
fore charring, the gasification rate is greatly enhanced (Takarada et al., 1985a).
An alternative approach in the investigation of mineral matter effects is to load the se-
lected element of interest onto a coal matrix. Based on the study of the gasification of chars
of demineralized coals loaded with Ca or Mg, Hippo et al. (1979) and Hengel and Walker
(1984) demonstrated that the reactivity of lignite chars is controlled mainly by the catalytic
effect of Ca associated with carboxyl groups. Takarada et al. (1985b) reported that the re-
activity of steam gasification for chars of lower rank coals is proportional to the amount of
Ca and Na ion exchanged by ammonium acetate for the lower rank coals. Other investiga-
tors (Fung and Kim, 1984; Morales et al., 1985; Fujita et al., 1982) have also observed the
catalytic activity of Ca, Mg, Na. K, etc. These results clearly show that the reactivity of the
chars of lower rank coals is controlled by the amount, state and distribution of coal miner-
als.
Interactions with existing mineral matter must also be envisaged as an additional com-
plicating factor. This may be avoided by using a model carbon with a very low mineral
5.7 FactorsAffecting the Reactivityof Coal Char during Gasification 297
matter content, such as one produced from a polymer resin (Gonenc et al., 1990; Li et al.,
1994). In this way, the effect of calcium has been found to increase up to a "saturation lev-
el" of about 4% w/w, and to depend strongly on the degree of dispersion within the car-
bonaceous matrix (Gonenc et al., 1990; Li et al., 1994).
5.7.3 Thermal History of Char
Several factors concerning pyrolysis may affect the reactivity, such as the volatile content,
the temperature at which coal is pyrolyzed (Zhang et al., 1996), the extension of the pyroly-
sis, the heating rate and the gas atmosphere at which the pyrolysis occurs (Miura et al.,
1989).
Generally, gasification reactivity decreases with the severity of conditions employed
for preparing char (Tp), i.e., higher Tp, longer holding time at Tp, and lower heating rate.
These effects are generally larger for lower rank coals than for higher rank coals. Serio et
al. (1987) reported that reactivities differ by a factor of 1000 among Zap lignite chars pre-
pared using heating rates between 0.5 and 20 000 K s-1 and final temperatures between
400 and 900 ~ On the other hand, van Heek and Muhlen (1987) reported differences of
only a factor of 2 for chars prepared from a bituminous coal between 700 and 900 ~ they
found no difference for chars prepared from an anthracite between 700 and 800 ~
Kasaoka et al. (1987) reported differences of a factor of 1.5 to 10 for chars prepared be-
tween 900 to 1300 ~ from ten coals ranging from 61.1 to 93.4 wt% C.
It is known that a graphite-like structure develops as the severity of the pyrolysis condi-
tions increases, leading to a decrease in active carbon sites. Furthermore, metals that act as
catalysts (Ca, Na, K, etc.) lose their activity by sintering, formation of intercalated com-
pounds or stable aluminosilicates, or through vaporization (Kasaoka et al., 1987; Radovic et
al., 1983 and 1984; Wigmans et al., 1983a, b). Therefore, the decrease in reactivity with in-
creasing severity of pyrolysis is caused by decreases in the number of carbon active sites
and in the catalytic activity.
Radovic et al. (1983 and 1984) tried to distinguish the two factors for a lignite. To ob-
serve only the decrease in carbon active sites with pyrolysis conditions, they gasified chars
prepared from a demineralized North Dakota lignite (Dem-char). Next, they prepared chars
from a demineralized and Ca-loaded lignite (Dem+Ca-char). These chars were prepared as
model samples to examine only the catalytic activity. In air at 0.1 MPa, the reactivity of the
Dem+Ca-char was more than 30 times larger than the Dem-char when the chars were pre-
pared by rapid pyrolysis (heating rate 104 K s-1) in an entrained-flow reactor at 1275 K.
The effect of pyrolysis conditions on reactivity differed greatly between the chars. For ex-
ample, between 975 and 1475 K, for a residence time of 1 h using slow pyrolysis (10 K
min-1), the reactivity of Dem-char decreased by a factor of only about six, whereas that of
Dem+Ca-char decreased by a factor of 100. Severe pyrolysis conditions enhance CaO sin-
tering in the Dem+Ca-char, and therefore cause a decrease in its dispersion. From these
studies, Radovic et al.(1983a, b, 1984a, b, c) state that the gasification reactivity of lignite
chars depends on the concentration of inherent catalytic sites.
For chars prepared under much milder pyrolysis conditions, the number of carbon ac-
tive sites may be more significant. Khan (1987) reported that char prepared from a high
volatile bituminous coal at 500 ~ was much more reactive than high-temperature chars.
The reactivity of the char was even higher than that of the parent coal. The significantly
greater reactivity of such low-temperature chars is attributed to the greater hydrogen con-
tent of the chars. Hydrogen-rich regions of coal char are preferentially oxidized, leaving
behind highly reactive "nascent" carbon sites.
298 5 Gasificationof Coal
addition of 0.1% oxygen to the hydrogen stream accelerated the formation of methane con-
siderably. On the other hand, Zoheidi and Miller (1987) reported that partial combustion of
carbon black prior to exposure to hydrogen enhances the methane formation rate, while
high temperature pretreatment (degassing) drastically reduces reaction rate. These results
arise from the addition or removal of active surface oxygen groups and from thermal an-
nealing of the carbon active sites. Indeed, via pH measurements it was found that partial
combustion at 400 ~ fixes acidic groups on the carbon surface (Zoheidi and Miller, 1987).
Gasified and degassed carbons contain a predominance of basic groups.
50 D _ _ 50 0.106
50 50 -- D -- 0.061
30 50.3 11.5 8.2 m 0.047
[Reproduced with permission from Goyal, A. et al., Am. Chem. Soc. Div. Fuel Chem. Prepr., 27 (1), 57 (1982)]
C(O) -~ CO (5.41)
The main characteristic of this mechanism is the inhibition by the CO produced which will
shift reaction (5.40) to the left.
5.8 Factors Affecting Reaction Rates 301
However, recently Moulijn and Kapteijn (1995) considered that the inhibitory mecha-
nism does not fully explain the reduction of gasification rate by H2. Experimentally they
found that gasification reaction stops almost completely when hydrogen concentration is
more than 50%. This suggests that H2 is part of two different mechanisms during gasifica-
tion, a reversible reaction which agrees with the Langmuir-Hinshelwood kinetics and an ir-
reversible reaction which leads to the deactivation of the active sites. The proposed mecha-
nism is:
Cf + H20 ~ C ( O ) + H2 (5.42)
5.8.2 Pressure
Although the incidence of the partial pressure of the reactive gases in the char gasification
rate has been exhaustively studied (Dutta et al., 1977; Yang and Watkinson, 1994; Chin et
al., 1983; Adanez and DeDiego, 1993; Goyal et al., 1989; Chi and Perlmutter, 1989;
Agarval and Sears, 1980; Tanaka et al., 1995), there is a dearth in the open literature con-
cerning the influence of the total system pressure.
Schmal et al. (1983) found that total pressure affects gas composition during steam coal
gasification at 850-1000 ~ in a fluidized bed. High pressures do not alter the system H2
concentration to any extent-from 61% at 0.1 MPa, to 58% at 1.0 MPa, but it increases
methane concentration from 1.1% (0.1 MPa) to 2.0% (1.0 MPa). The CO/CO2 ratio also
decreases for higher pressures.
In the same investigation, it was found that the gasification rate increases with total
pressure and that its effect is more marked in the low pressure region. For example, the re-
activity doubles its value when pressure is raised from 0.1 to 0.5 MPa. However, pressure
values above 1.0 MPa do not produce a significant increase in gasification rate.
5.8.3 S a m p l e Size
When char reactivity is studied, laboratory experiments are generally done in such a way
that diffusive restrictions can be avoided. The analysis is done by plotting char conversion
against time for different particle sizes. Diffusion restrictions should be considered when
burn-off curves begin to level off for larger particles. This means that particle size should
be small enough so that no difference can be found in reactivity if a smaller particle is used
for reactivity studies.
The particle size where diffusion restrictions can be neglected depends on coal type.
Kovacik et al. (1991) found diffusion restrictions for subbituminous and bituminous coals
at 900 ~ and particle size (about 74-105)/.tm. On the other hand, Matsui et al. (1987) did
not find diffusion restrictions when working with subbituminous 710-/~m coal particles.
Chin et al. (1983) worked with coal particles up to 1000/~m without finding any diffusion
effect. Such differences have made every research team change the particle size in order to
find the experimental conditions where diffusion restriction can be neglected.
Even though gasification processes of various types are in operation in industries world-
wide, the research and development work now in progress to produce advanced environ-
302 5 Gasificationof Coal
In Table 6.1, Actinomycetes are found everywhere due to flying of spores and are
caught frequently in the screening of coal-degradable microbes. Many reports of aerobic
bacteria have been presented, and aerobic bacteria such as Pseudomonas and Rhodococcus
304 6 MicrobialDepolymerizationof Coal
often have the ability to degrade aromatic compounds oxidatively, although it has been re-
ported that Pseudomonas cepacia degraded water-soluble coal nonoxidatively (Gupta et al.,
1990; Crawford and Gupta, 1991a). On the other hand, there are few reports on anaerobic
bacteria, although it is assumed that they degrade water-soluble coal reductively into low
molecular compounds.
In Table 6.1, yeast well known in alcohol fermentation have not been found in any re-
port for application to coal depolymerization. In a broad sense, fungi include yeasts, fungi,
alias "mold" are full of vigor in the development of mycelia in vegetative phase which do
not form "mushrooms" and mushrooms. Genera such as Aspergillus and Penicillium of
fungi are found everywhere and have versatile abilities. Various species have been studied
extensively. Recently, it has been found that lignin-degradable fungi, alias "white-rot fun-
gi" of Basidiomycetes, which are able to form mushrooms, have the ability to depolymerize
coal to convert into low-molecular matter (Catcheside and Ralph, 1999, Fakoussa and
Hofrichter, 1999).
Then, how can microorganisms able to depolymerize recalcitrant macromolecular coal
be screened? Enrichment culture is well known as a method for screening microbes, for ex-
ample, culture including coal as the sole source of carbon using soil or waste water includ-
ing coal as habits is transferred repeatedly into the same fresh media. Thus, some microbes
survive and are enriched. These surviving microbes are isolated and each is cultured purely
and tested for its ability to degrade coal or coal-related compounds.
6.1.2 A p p r o a c h to M i c r o b i a l D e p o l y m e r i z a t i o n o f C o a l
As shown in the structure model of coal proposed by Wiser (1975) (see Fig. 2.7 in Chapter
2), coal is a complicated macromolecule that consists mainly of 1-5 aromatic tings includ-
ing one or two naphthenic and heterorings with O, N and S and are bridged by ether bonds.
These structures are different in degree of coalification, as shown in the structure model
proposed by Schumacher (1997). Recently, it was shown that even bituminous coal, a typi-
cal coal, has 3-10 long chains of methylene groups (Shinn, 1984). Moreover, considerable
amounts of longer ethylene groups are present in low-rank coal, as reported by Hu et al.
(2000). Based on the above reports, a structural model of coal including various unit struc-
tures and bridge bonds is shown in Fig. 6.1, where different bonds attacked microbially are
marked with arrows and numbers. Macromolecules having such a complicated structure
are unlikely to be degraded by just one kind of microorganism.
Accordingly, microbial depolymerization of coal will be practicable by a strategy of
collecting microorganisms able to attack different bonds in Fig. 6.1 using a mixture of en-
zymes.
As shown in the structure model of coal proposed by Wiser (1975), coal is a complicat-
ed macromolecule. It is constructed mainly of 1-5 aromatic tings. Moreover, it includes
one or two naphthenic rings and heterorings with O, N, S. These tings are bridged by ether
bonds, methylene groups and others.
Another strategy is the use of lignin-degrading fungi. Fungi capable of degrading re-
calcitrant lignin which have a complicated structure like coal play an important role in the
material cycle in nature. Lignin is a complicated macromolecule consisting of phenyl-
propane derivatives as basal structual units as proposed by Nimz (1974). As shown in the
figure, lignin is a copolymer of guaiacyl and syringyl units connected with many fl-O-4
bonds. Noting this structure, studies on the screening of microorganisms able to degrade
lignin dimer have been reported (Crspedes et al., 1992; Rhoads et al., 1995). Structures of
lignin monomer and dimer are shown in Fig. 6.2.
305
OC"3oIC.3""
Fig. 6.1 Microbial attack on different bonds of low rank coal.
(~, (~), (~): Ring cleavage of aromatic hydrocarbon
(~): Decomposition of biphenyl
(~), (fi), (~): Ring cleavage of aromatics substituted with oxygen-including groups
(~): Ether bond cleavage
(~): Cleavage of methylene group
Monomers
Guaiacyl unit Syringyl unit
CH2OH CHEOH
CH2 CH2
I I
CH2 CH2
OCH3 OCH3
......................
ortho
cleavage [•"COOH
~t~,,,COOH
o.
on___~ } -'--~ TCA cycle = CO2 + H20
"OH ~ O H
OH
...................... meta
cleavage OH
Benzene Catechol
The study using phenanthrene as a model compound of coal by Rogoff and Wender
(1957) may be the first attempt in the microbial degradation of coal. Later, microorganisms
able to degrade multi-ring aromatic hydrocarbons such as biphenyl (Catelani et al., 1971,
1973; Gibson et al., 1973; Abe et al., 1995; Kimura et al., 1996), naphthalene (Kiyohara et
al., 1994; Yang et al., 1994), anthracene (Kabe, Y., 1993), fluorene (Yang et a1.,1994),
pyrene (Bouchez et al., 1997) and others were found, and the degradation pathway, its relat-
ed enzymes and genes have been reported. Among microorganisms having the ability to
degrade aromatic hydrocarbons, a number of bacteria, especially pseudomonads, have been
reported. For naphthalene with two rings, the first ring is decomposed by ortho cleavage to
form salicylic acid, which is further decomposed by meta cleavage. For phenanthrene with
three rings, the first ring is decomposed by ortho cleavage to form 1-hydroxy-2-naphthoic
acid and subsequently decomposed by the pathway of naphthalene degradation, but most
bacteria can not decompose further after opening of the first ring. Therefore, it will be ef-
fective to make consortia of bacteria having different abilities as in nature. In addition,
while most microorganisms are not viable in organic solvents, it has been reported that bac-
teria tolerant to not only solvents such as hexane and cyclohexane but highly toxic solvents
such as benzene and toluene were found in deep sea and these could more effectively de-
compose aromatic hydrocarbons such as biphenyl and naphthalene in these solvents (Abe et
al., 1995).
6.3 Depolymerizationof Coal 307
1991; H61ker et al., 1995, 1999; Kabe et al., 1995, 1999; Catcheside and Ralph, 1999; G6tz
and Fakoussa, 1999; Laborda et al., 1999; Weber et al., 2000), and various mechanisms of
solubilization have been proposed (Runnion and Combie, 1990; Catcheside and Ralph,
1999; H61ker, et al., 1999). The main factors in the formation of these black drops are the
solubilization of coal by alkaline matter produced by microorganisms (ionization of coal by
increased pH) and chelating agents (ionization of coal by chelating of Fe 3+, Ca 2+, Mg 2+, etc.
bonded with coal), but not the action of enzymes produced by the microorganisms in most
cases.
Microbial solubilization of solid coal in culture medium is observed by the appearance
of a brownish color, and can be monitored by the measurement of UV-VIS absorption spec-
tra (Kabe et al., 1995, 1999; Laborda et al., 1999; H/31ker et al., 1999). In Fig. 6.4, UV-VIS
absorption spectra of four kinds of coal powder (Argonne Premium Coal Program) solubi-
lized by the fungus Aspergillus sp. strain are shown (Kabe et al., 1999). This figure shows
that microbially solubilized products of coal (spectra with inoculation) are water-soluble
macromolecules similar to water-soluble components from coal (spectra without inocula-
tion). Laborda et al. (1999) revealed that black drops were formed from Spanish lignite
powder on mycelia mat of Aspergillus sp. strain, and comparing IR spectra of these black
drops with those of native coal, methyl and methylene groups decreased, while hydroxy and
carboxy groups increased. Weber et al. (2000) obtained results similar to those of Laborda
et al. (1999). COOH and CH2 in German brown coal solubilized by the fungus Trichoderma
atroviride were determined quantitatively using FT-IR spectroscopy, and it was shown that
the solubilized matter was humic acid (Weber et al., 2000).
Low-rank coal solubilized by microorganisms is as yet impractical, but as described by
Faison (1991), solubilized products have potential for use as anti-oxidizing agents, surfac-
IL UF
<
1.0 -
0
POC ND
<
1.0 - ~
0 I I I
200 300 400 500 200 300 400 500
Mnm Mnm
Fig. 6.4 UV-VIS spectra of coal dissolved from four kinds of coal powder into liquid cultures by Aspergillus sp.
strain FKS11. (Kabe et al., 1999)
Incubation period: 4 weeks, coal: Coals of the Argonne Premium Sample Program (IL: Illinois #6, UF:
Upper Freeport, POC: Pocahontas #3 and ND: Beulah zap), measurement conditions of spectra: 1/5 dilu-
tion at pH 10, (~): inoculation, (~: non-inoculation.
[Reproduced with permission from Bao Qing Li. et al., Prospects For Coal Science In The 2U Century,
I, 326, Shanxi Science & Technology Press (1999)]
6.3 Depolymerizationof Coal 309
tants, components of resin and adhesives, immune supplements, chelating agents, soil stabi-
lizers, chemicals, gas and liquid fuels, etc.
The following investigation is of interest in the attempt to utilize solubilized products.
Ftichtenbusch and Steinbtichel (1999) reported that liquid products microbially solubilized
could be utilized as nutrition for the bacteria Pseudomonas oleovorans, Rhodococcus ruber
and that their metabolites, polyhydroxyalkanoates, were biosynthesized. These can be uti-
lized as biodegradable plastics. Another application of solubilized coal is continuous two-
stage processes combining aerobic cultures with anaerobic cultures, producing liquid fuels
such as ethanol, 1-propanol, acetic acid, etc. (Ackerson et al., 1990). There is also an appli-
cation to surfactant for CWM (coal-water mixture) as a direct utilization of microbiological
solubilized products (Saiki et al., 1991).
6.3.2 Depolymerization of Coal Humic Acid
As mentioned in the last section, water-soluble macromolecules are produced, but low mol-
ecular substances are seldom produced by microbial action to solid coal. If accompanied
by the production of low molecular substances, the microorganism also has the ability to
depolymerize the solubilized macromolecules, in addition to the ability to solubilize coal.
However, in culture systems where both reactions proceed from solid coal successively, it
is difficult to distinguish the depolymerization from the solubilization of coal. Thus, most
investigators have attempted differentiating the depolymerization of water-soluble coal
(coal humic acid) from the solubilization described in the last section. In most investiga-
tions of the depolymerization of coal, coal humic acid is used as an alkali-soluble acid-pre-
cipitate prepared by alkali extraction from low-rank coal such as lignite and brown coal.
Coal humic acids give broad spectra gradually decreasing absorbance without a special ab-
sorption band over the ultraviolet to visible region in UV-VIS absorption spectra as shown
in Fig. 6.5 (Kabe et al., 1999), and bands around 3000--3500 cm -1 corresponding to O-H
and C-H stretching vibration and bands around 1660 cm-1 corresponding to C=O stretching
vibration of aromatic COOH are characteristic of their IR spectra, as shown in Fig. 6.6
(Kabe et al., 1999). Solubilized coals (humic acid) used in Figs. 6.5 and 6.6 were prepared
collecting precipitates by acidification from alkali extracts after nitric acid-oxidation of
weathered Illinois # 6 coal, and sample B is soluble in pH 7--5.5 and precipitated in pH 5.5,
and sample C is soluble in pH 5.5--1.5 and precipitated in pH 1.5. Upon action of
Aspergillus sp. Strain FKS1 to these solubilized coals B and C, decolorization was ob-
served as shown in Fig. 6.5, and it was assumed that the carboxy-containing aromatic ring
was cleaved to form aliphatic carboxylic acid, because the band around 1660 cm-1 shifted
to around 1720 cm-1, and reduction also simultaneously took place, because absorption in-
tensity of CH2 near 2900 cm-1 increased slightly, as shown in Fig. 6.6.
For investigations on the microbial depolymerization of humic acid, there are reports
on the discovery of microorganisms able to depolymerize soil humic acid by Burges and
Latter (1960), and later findings of a white-rot fungus able to degrade forest soil humic acid
(Hurst and Burges, 1962; Haider and Martin, 1988; Blondeau, 1989), streptmycetes (Monib
et al., 1981; Kontchou and Blondeau, 1990), and various other bacteria (Gordinko and
Kunz, 1984). Among these microorganisms, white-rot fungi, members of basidiomycetes
and lignin-degraders, have the highest ability to degrade soil humic acid. Phanerochaete
chrysosporium well known as the most excellent lignin-degrader among white-rot fungi has
been used for the depolymerization of coal humic acid by many investigators (Haider and
Martin, 1988; Blondeau, 1989; Wondrack et al., 1988, 1989; Ralph and Catcheside, 1994,
1999; Ralph et al., 1996).
310
r~
<
1.0
0 m
r~
d~
<
_ O
1.0
0
200 300 400 500 600 700
&/nm
Fig. 6.5 UV-VIS spectra of solubilized Illinois #6 coal B (top) and C (bottom) before and after incubation with
Aspergillus sp. strain FKS1. (Kabe et al., 1999)
Incubation period: 2 weeks, measurement conditions of spectra: 2/25 dilution at pH 10, (~): before incu-
bation, (~): after incubation.
[Reproduced with permission from Bao Qing Li. et al., Prospects For Coal Science In The 21 st Century,
I, 325, Shanxi Science & Technology Press (1999)]
BQ
B|
c~
c@
I I I I I I I I
Wavenumber cm-!
Fig. 6.6 IR spectra of acid precipitates obtained from solubilized Illinois #6 coal B (top) and C (bottom) before
and after incubation with Aspergillus sp. strain FKS 1. (Kabe et al., 1999)
Incubation period: 2 weeks, measurement conditions of spectra: 2/25 dilution at pH 10, @: before incu-
bation, @: after incubation. [Reproduced with permission from Bao Qing Li. et al., Prospects For Coal
Science In The 21" Century, I, 325, Shanxi Science & Technology Press (1999)]
6.3 Depolymerizationof Coal 311
6.3.4 E n z y m e s I n v o l v e d in the D e p o l y m e r i z a t i o n of C o a l
This section summarizes information about the isolation and action of extracellular en-
zymes produced by lignin-degrading fungi, which are very effective in the depolymeriza-
tion of coal humic acid, as described above. Lignin-degrading fungi are called white-rot
fungi because they cause whitening of wood due to cellulose left by the degradation of
lignin, and most of these fungi are basidiomycetes, which form mushrooms. Enzymes in-
volved in the depolymerization of lignin as well as coal humic acid are mainly three kinds
of extracellular enzymes: lignin-peroxidase (LIP), manganese peroxidase (MnP) and lac-
case. These enzymes depolymerize and decolorize lignin, but in plants they occasionally
polymerize excess nutrients in order to accumulate them as lignin and change plants to a
brown color. Thus, depolymerization and decolorization by basidiomycetes usually take
place under conditions of limited carbon and nitrogen. The ability to produce these three
enzymes varies depending on the kind of fungus, but there are also fungi species in which
one fungus produces all three enzymes (Temp et al., 1999).
A. Lignin Peroxidase
Lignin peroxidase (LIP) produced by the basidiomycete Phanerochaete chrysosporium is
the most frequently investigated, and much information is available (Kirk et al., 1986;
Harvey et al., 1986; Candeias and Harvey, 1995; Tien and Kirk, 1988; Yoshida et al.,
1996a, 1996b; Chung and Aust, 1995). LiP is the enzyme which oxidates veratryl alcohol
(3,4-dimethoxybenzyl alcohol, VOH) to veratric aldehyde (3,4-dimethoxybenzaldehyde,
VCHO) with H202 as oxidizing agent, and it is a protoporphyrin IX-containing glycopro-
tein. In culture, H202 is generated when glucose is oxidated by glucose oxidase. In addi-
tion, veratryl alcohol VOH is produced endogenously in culture. In the course of the oxida-
tion of this VOH to aldehyde, the cation radical VOH "§ is produced. It then attacks bridge
linkage and aromatic ring and radical chain reaction takes place in lignin and coal humic
acid, resulting in their depolymerization. Here, veratryl alcohol plays the role of mediator
for the depolymerization. As the mediator may or may not be veratryl alcohol and aromatic
compounds near this structure also play such a role, relatively low molecular substances in-
cluded in lignin and coal humic acid may be mediators. The amount of H202 generated in
culture is important. In view of this problem, Li and Chen found that the addition of hydro-
carbons such as hexadecane into the culture medium significantly enhanced glucose oxi-
dase activity (Li and Chen, 1994).
B. Manganese Peroxidase
Basidiomycete Phanerochaete chrysosporium also produces the peroxidase MnP in addi-
tion to LiP (Wariishi, et al., 1992). MnP is an enzyme similar to LiP, but different in re-
quiting manganese as substrate and as mediator. When Mn 3+ is reduced to Mn 2+ by MnP,
other mediators such as thiol, lipid and unsaturated fatty acids are simultaneously oxidated,
and the resulting cation radicals can sever linkage such as ether bond bridging between
multi-ring aromatics cannot be attacked by MnP.
C. Laccase
Laccase produced by basidiomycetes is also an important enzyme for catalyzing the de-
polymerization of lignin and coal humic acid. Laccase is called polyphenol oxidase or phe-
nolase, and it is the enzyme which oxidates polyphenols to quinones with 02 as oxidants,
and fungi-producing laccase is a protein containing four copper atoms. Yoshiyama investi-
6.4 EnvironmentalRemediation 313
gated the conditions for enhancing laccase producibility of Coriolus versicolor on a jar fer-
menter scale (Yoshiyama and Itoh, 1994). Nishizawa isolated and purified laccase from the
white-rot fungus Trametes sanguinea and characterized it (Nishizawa et al., 1995).
Furthermore, Sheel et al., (1999) demonstrated evidence of expression of genes coding lac-
case produced by three basidiomycetes able to depolymerize humic acid. Studies on the
mechanism of depolymerization of lignin and others by laccase are behind those for LiP
and MnP. Although laccase resembles them in severing bonds of macromolecules by radi-
cal reaction via mediator, there remain differing views regardingly the subject. In any case,
this ability is a very attractive area of investigation.
D. Arylalcohol Oxidase
Of the lignin-degrading enzymes, LiP oxidizing veratryl alcohol with H202 is well known,
and arylalcohol oxidase AAO, an enzyme oxidizing veratryl alcohol with O2 also exists. In
1988, Waldner found that AAO was produced by the white-rot fungus Bjerkander adusta
(Waldner et al., 1988; Muheim et al., 1990). Whereas the enzyme LiP oxidating with H202
is produced in a relatively later phase of culture, AAO is produced earlier and decomposes
lignin rapidly. Later, it was found that AAO was also produced by Phanerochaete
chrysosporium (Asada et al., 1995).
E. Hydrolase
In addition to the above enzymes catalyzing oxidation, hydrolase also enables the depoly-
merization of lignin and coal. An enzyme that depolymerizes coal humic acid non-oxida-
tively by the bacterium Pseudomonas cepacia appears to be esterase, a kind of hydrolase
(Crawford and Gupta, 1991). Lipase produced by Ps. cepacia is commercially available.
As lipase is a kind of hydrolase or esterase, and relatively high heat-resistant and tolerant to
organic solvents, it is a useful enzyme for organic syntheses such as application to biosyn-
thesis of diesel oil (Iso et al., 2001). Lipase highly tolerant to various organic solvents has
been isolated from Fusarium heterosporum and purified (Shimada et al., 1993). Thus, es-
terase of not only bacteria but also those of imperfect fungi such as Fusarium and
Trichoderma atroviride are involved in the depolymerization of lignin or coal (Laborda et
al., 1999; H61ker et al., 1999; Weber et al., 2000).
A
Abdel-Baset, M. B., Yazab, R. F., Given, P. H., Fuel, 57, 89, Elsevier (1978).
Abe, A., Inoue, A., Usami, R., Moriya, K., Horikosi, K., Biosci. Biotech. Biochem., 59, 1154 (1995).
Ackerson, M. D., Johnson, N. L., Le, M., Clausen, E. C., Gaddy, J. L., Appl. Biochem. Biotechnol., 24/25, 913
(1990).
Adanez, J. and DeDiego. R. F., Int. Chem. Eng., 33, 656 (1993).
Adschiri, T., Shiraha, T. Kojima, T. and Furuzawa, P., Fuel, 65, 1688, Elsevier (1986).
Agarval, A. K. and Sears, J. T., Ind. Eng. Chem. Proc. Des. Dev., 19, 364 (1980).
Agarwal, P. K., Genetti, W. E., Lee, Y. Y.,Am. Chem. Soc. Div. Fuel Chem. Prepr., 29 (2) 94 (1984).
Ahmed, S. and Back, M. H., Carbon, 23, 513 (1985).
Aida, T. and Squires, T., G. Prepr., Am. Chem. Soc., Div. Fuel Chem., 30, 95 (1985).
Aida, T., Fuku, K., Fujii, M., Yoshihara, M., Maeshima T., Squires, T.G., Energy Fuels, 5, 79 (1991).
Aida, T., J. Jpn. Fuel Soc., 70, 820 (1991).
Aida, T., Sato, T., Sekiguchi, G., Adschiri, T., Arai, K., Fuel, 81, 1453, Elsevier (2002).
Albright, L. F., Crynes, B. L., Corcoran, W. H., Pyrolysis, Theory and Industrial Practice, Academic Press, (1983).
Allan, J. and Larter, S. R., Advances in Organic Geochemistry (ed. Bjoroey, M., 1981), p. 534, John Wiley and
Sons, Inc. New York (1983).
Allardice, D. J., Carbon, 4, 255 (1966).
Alonso, A. M., Birmejo, J., Granda, M., Tascon, J. M. D., Fuel, 71, 611 (1992).
Alvarez, T., Fuertes. A. B., Pis, J. J., Ehrburger, P., Fuel, 74, 729, Elsevier (1995).
Amett, E. M., Mitchell, E. J., Murty, T. S. S. R, Gorrie, T. M., Schleyer, P. V. R., J. Am. Chem. Soc., 92, 2365
(1970).
Amett, E. M., Joris, L., Mitchell, R., Murty, T. S. S. R., Gorric, T. M., Schleyer, P. V. R., J. Am. Chem. Soc., 92,
74 (1970).
Amett, E. M., Mitchell, E. J., Murty, T. S. S. R., J. Am. Chem. Soc., 96, 3875 (1974).
Amono, A., Horie, O., Sato, Y., Katayose, T., J. Jpn. Petrol. Inst., 15, 41 (1972).
Asada, Y., Watanabe, A., Ohtsu, Y., Kuwahara, M., Biosci. Biotech. Biochem., 59, 1339 (1995).
Asante, K. O. and Stock, M. L.,J. Org. Chem., 51, 5452 (1986).
Attar, A. and Hendrickson, G. G., Coal Structure, (ed. Meyers, R. A.), Academic Press, New York, p 131 (1982).
Autrey, S. T., Alborn, E. A., Franz, J. A., Camaioni, D. M., Energy Fuels, 9, 420 (1995).
Autrey, T., Linehan, J. C., Camaioni, D. M., Kaune, L. E., Watrb, H. M., Franz, J. A., Catal. Today, 31, 105
(1996).
B
Baba, M.,An Introduction to Fuel, p. 47, Gihoto (1965).
Babu, S.P., Knight, R.A., Onischak, M., Wootten, J.M., Duthie, R.G., Longanbach, J.R., Paper presented at the
7th U.S./Korea Workshop on Coal Utili. Tech., Pittsburgh, (1990).
Bacaud, R., Besson, M., Djega-Mariadassou, G., Prepr.,Am. Chem. Soc., Div. Fuel Chem., 38 (1), 1 (1993).
Badger, G. M., Jolad, S. D., Spotswood, T. M.,Aust. J. Chem., 17, 771 (1964).
Baldwin, R. M., and Vinciguerra, S., Fuel, 62, 498, Elsevier (1983).
Bale, H. D. and Schmidt, P. W., Phys. Rev. Lett., 53, 596 (1984).
Barton, W. A., Lynch, L. J., Webster, D. S., Fuel, 63, 1262, Elsevier (1984).
Barton, W. A., Lynch, L.J., Proc. 6th Australian Coal Sci. Conf., Newcastle, 65 (1994).
Bate, K. and Harrison, G., Fuel, 71, 289, Elsevier (1992).
Benjamin, B. M., Raaen, V. F., Maupin, P. H., Brown, L. L., Collins, C. J., Fuel, 57, 269, Elsevier (1978).
Benjamin, B. M., Hagaman, E. W., Raaen, V. F., Collins, C. J., Fuel, 58, 386, Elsevier (1979).
Benjamin, B. M., DoUglas, E. C., Mesmer, S., Prepr., Am. Chem. Soc., Div. Fuel Chem., 27, 807 (1982a).
Benjamin, B. M., Douglas, E. C., Mesmer, S., Prepr., Am. Chem. Soc., Div. Fuel Chem., 27, 1 (1982b).
Benjamin, B. M., Douglas, E. C., Mesmer, S., US DOE Rep. (1983).
Benson, R. E., Thermochemical Kinetics, 2nd Ed., Wiley-Interscience, New York (1976).
316 References
C
Calf, G. E. and Garnett, J. L.,Adv. Heterocycl. Chem., 15, 137 (1973).
Camaioni, D. M., Autrey, S. T., Franz, J. A., J. Phys. Chem., 97, 5791 (1993).
Cameron, R. E. and Miller, R. M., Int. Symp. Microbiol. Ecolog., 1, Dunedin, New Zealand, Springer (1977).
Candeias, L. P. and Harvey, P., Phys., 316, 733 (1995).
Cao, J. R. and Back, M. H., Carbon, 23, 142 (1982).
Carlson, G. A., Energy Fuels, 6, 771 (1992).
Cartz, L and Hirsch, P. B., Philos, Trans. R. Soc. London Ser. A, 252, 559 (1960).
Cartz, L., Diamond, R., Hirsch, P. B., Nature, 177, 500 (1956).
Catcheside, D. E. A. and Mallett, K. J., Energy Fuels, 5, 141 (1991).
Catcheside, D. E. A. and Ralph, J. P., Appl. Microbiol. Biotechnol., 52, 16 (1999).
Catelani, D., Sorlini, C., Treccani, V., Experienta, 27, 1173 (1971).
Catelani, D., Colombi, A., Sorlini, C., Treccani, V., Biochem. J., 134, 1063 (1973).
Causton, P. and McEnanay, B., Fuel, 64, 1447, Elsevier (1985).
References 317
D
Davidson, R. M., Molecular structure of coal, Report number ICTIS/TR 08, IEA Coal Research, London (1980).
Davidson, R. M., Molecular structure of coal, in : Coal Science, Vol. 1 (eds. Gorbaty, M. L. Larsen, J. W. Wender,
I.), Academic Press, New York p. 83 (1982)
Davis, A., Mitchell, G. D., Derbyshire, F. J., Rathbone, R. F., Lin, R., Fuel, 70, 352, Elsevier (1991).
Davis, A., Derbyshire, F. J., Mitchell, G. D., Schobert, H. H., Enhanced Coal Liquefaction by Low-Severity
Catalytic Reactions, Report to U. S. Dept of Energy DOE-PC-90910-F1, p. 175 (1989).
318 References
Davis, A., Rathbone, R. F., Lin, R., Quick, J. C., Org. Geochem., 16, 897 (1990).
Davis, G. O., Coal Refining by Solvent Extraction, Chem. Ind., Aug. 1978.
Davis, G. O., Production of toluene and xylenes from coal. Toluene, the xylenes and their industrial derivatives.
Elsevier Chem. Eng. Monogr., 15.
Davies, G, O., in "Chemical feedstocks by the direct liquefaction of coal", (ed. Payne, K. R.), 1985.
Davis, K. P. and Garnett, J. C., J. Chem. Soc., Chem. Commun., 79 (1975).
DeCanio, E. C., Edwards, J. C., Scalzo, T. R., Storm, D. A., Bruno, J. W., J. Catal., 132, 498 (1991).
Deevi, S. C. and Suuberg, E. M., Fuel, 66, 54, Elsevier (1987).
Delannay, F., Tysoe, W. T., Heinemann, H., Somonjai, G. A., Carbon, 22, 401 (1984).
Dela Rosa, L., Pruski, M., Lang, D., Gerstein, B., Solomon, P., Energy Fuels, 6, 460 (1992).
Derbyshire, F. J. and Whitehurst, D. D., Fuel, 60, 655, Elsevier (1981).
Derbyshire, F. J., de Beer, V. H. J., Abotsi, G. M. K., Scaroni, A. W., Solar, J. M., Skrovanek, D. J., Appl. Catal.
27, 117 (1986a).
Derbyshire, F. J, Davis, A., Epstein, M., Stansberry, P., Fuel, 65, 1233, Elsevier (1986b).
Derbyshire, F. J., Davis, A., Lin, R., Prepr, Am. Chem. Soc., Div. Fuel Chem., 34 (3), 676 (1989a).
Derbyshire, F. J., Davis, A., Lin., R., Energy Fuels, 3, 431 (1989b).
Derbyshire, F. J., Marzec, A., Schulten, H-R., Wilson, M. A., Davis, A., Tekely, P., Delpuech, J-J., Jurkiewicz, A.,
Bronnimann, C. E., Wind, R. A., Maciel, G. E., Narayan, R., Bartle, K., Snape, C., Fuel, 68, 1091, Elsevier
(1989c).
Derbyshire, F. J., Energy Fuels, 3, 273 (1989).
Derbyshire, F. J., CHEMTECH, 20, 439 (1990).
Derbyshire, F. J., Fuel, 70, 276, Elsevier (1991).
De Rosset, A. J. and Tan, G. Hilfman, Hydrocarbon Processing, May 1979.
De Rosset, A. J. et al., UOP Inc. US DOE Contract EF-77-C-012566.
Detoni, S., Hadzi, D., Smerkolj, R, J. Chem. Soc. (A), 2851 (1970).
Dickinson, E. M., Fuel, 64, 704, Elsevier (1985).
Diez, M. A., Alvarez, R., Gonzalez, A. I., Menendez, R., Moinelo, S. R., Bermejo, J., Fuel, 73, 139, Elsevier
(1994).
Doolan, K. R., Mackie, J. C., Fuel, 64, 400, Elsevier (1985).
D'Orazio, F., Tarczon, J. C., Halperin, W. P., Eguchi, K., Mizusaki, T., J. Appl. Phys., 65, 742 (1989).
Drago, R. S. and Epley, T. D., J. Am. Chem. Soc., 91, 2883 (1969).
Dryden, I. G. C., Fuel., 37, 444, Elsevier (1958).
Dryden, I. G. C., Fuel., 41, 55, Elsevier (1962).
Dutta, S., Wen, C. Y., Belt, R. J., Ind. Eng. Chem. Proc. Des. Dev., 16, 20 (1977).
E
Eggert, C., Temp, U., Dean, J. F. D., Eriksson, K. L., FEBS Lett., 391, 144 (1996).
Ehrburger, P., Mahajan, O. P., Walker, P. L., J. Catal., 43, 61 (1976).
Ehrburger, P. and Lahaye, J., Fuel, 63, 494, Elsevier (1984).
Eisenhut, W., Chemistry of Coal Utilization, (ed. Elliot, M.A.), John Wiley & Sons, Inc., New York, p.1 (1981).
Eismann, F., Becker, F., Kushk, U., Stottmeister, U., Appl. Microbiol. Biotechnol., 46, 604 (1996).
Entel, J., J. Am. Chem. Soc., 76, 3646 (1954).
Eply, T. D. and Drago, R. S., J. Am. Chem. Soc., 89, 5770 (1967).
Evance, D. G., Fuel, 52, 186, Elsevier (1973).
Evans, E. L., Jenkins, J. L., Thomas, J. M. Carbon, 10, 637 (1972).
Exxon Donor Solvent (EDS) Coal Liquefaction Process Development, US Report DOE FE 2893-61, March 1981.
F
Faison, B. D., Critical Rev. Biotechnol., 11,347 (1991).
Fakoussa, R. M., Doctoral dissertation, Rhein Friedrich- Wilhelms Univ., Bonn ( 1981).
Fakoussa, R. M., Frost, P. J., Appl. Microbiol. Biotechnol., 52, 60 (1999).
Fakoussa, R. M., Hofrichter, M., Appl. Microbiol. Biotechnol., 52, 25 (1999).
Farcasiu, M. and Smith, C., Prepr., Am. Chem. Soc., Div. Fuel Chem., 35, 404 (1990).
Farcasiu, M. and Smith, C., Energy Fuels, 5, 83 (1991).
Farrar, T. C. and Becker, E. D., Pulse and Fourier Transform NMR, Academic Press, New York (1971).
Faulon, J-P., Hatcher, P. G., Wenzel, K. A., Prepr., Am. Chem. Soc., Div. Fuel Chem., 37 (2), 900 (1992).
Fischer, F. and Gluud, W., Ber. Deut. Chem. Ges., 49, 1460 (1916).
Fischer, F., Broche, H., Strauch, J., Brennstoff-Chem., 5, 299 (1924).
Fischer, F., Broche, H., Strauch, J., Brennstoff-Chem., 6, 33 (1925).
Fitzer, E., Mueller, K., Schaefer W., Chemistry and Physics of Carbon (ed. Walker Jr, P. L.). Marcell Dekker,
New York, Vol.7, p. 263 (1971).
References 319
Fleming, I. and Wildsmith, E., J. Chem. Soc. D: Chem Commun., 4, 223 (1970).
Fletcher, T. H., Kerstein, A. R., Pugmire, R. J., Grant, D. M., Energy Fuel, 3, 54 (1990a).
Fletcher, T. H., Kerstein, A. R., Pugmire, R. J. Grant, D. M., Energy Fuel, 6, 414 (1992a).
Flory, P. J., Principles of Polymer Chemistry, Cornell University Press, Ithaca, New York (1953).
Forrest, M. and Marsh, H., Fuel, 62, 612, Elsevier (1983).
Fossil Energy: Catalytic Inc., US Department of Energy, DOE/PC/50041-(79).
Fossil Energy: Ashland Synthetic Fuels Inc., US Department of Energy, DOE/ET 10143-19/37.
Franck, H. G. and Knop, A., Kohleverdlung- Chemie und Technologie, 14, Springer, Berlin (1979).
Franklin, H.D., Cosway, R. G., Peters, W. A., Howard, J. B., Ind. Eng. Chem. Process Des. Dev., 22, 39 (1983).
Franz, J. A., Fuel, 58, 405, Elsevier (1979).
Franz, J. A. and Camaioni, D. M., Fuel, 59, 803, Elsevier (1980a).
Franz, J. A. and Camaioni, D. M., J. Org. Chem., 45, 5247 (1980b).
Franz, J. A. and Camaioni, D. M., Proc., Int. Conf. Coal Science, p. 327 (1981a).
Franz, J. A., Camaioni, D. M., Skiens, W. E.,Adv. Chem. Ser., 192, 75 (1981b).
Franz, J. A. and Camaioni, D. M., Prepr.,Am. Chem. Soc., Div. Fuel Chem., 26, 106 (1981c).
Franz, J. A., Garcia, R., Linehan, J. C., Love, G. D., Snape, C. E., Energy Fuels, 5, 598 (1992).
Freriks, I. L. C., van Wechem, H. M. H., Stuvier, J. C. M., Bouwman, R., Fuel, 60, 463 (1981).
Fruhbeis, H., Klein, R., Wallmeir, H., Angew. Chem., Int. Ed. Engl., 26, 403 (1987).
Fuchs, W. and Scandohoff, A. G., Ind. Eng. Chem., 34, 567 (1942).
Fuchtenbusch, B. and Steinbuchel, A., Appl. Microbiol. Biotechnol., 52, 91 (1999).
Fujita, H., Uba, M., Nishida, S., J. Jpn. Fuel Soc. 61,301 (1982).
Fukuda, Y. and Sone, K., Bull. Chem. Soc. Jpn., 45, 465 (1972).
Fuller, E. L. Jr. and Smyrl, N., Fuel, 64, 1143, Elsevier (1984).
Fung, D. P. C. and Kim, S. D., Fuel, 63, 1197, Elsevier (1984).
Furlong, L. E., Effron, E., Vernon, L. W., Wilson, E. L., Chem. Eng. Progr., 72 (8), 69 (1976).
Fu, Y. C. and Blaustein, B. D., Chem. Ind., 29, 1257 (1967).
Fynes, G., Ladner, W. R., Nreman, J. O. H., Prog. Energy Combust, Sci., 6, 223 (1980).
G
Gaines, A. F. and Yurum, Y., Fuel, 55, 129, Elsevier (1976).
Gaines, A. F., New Trends in Coal Science (ed. Yurum, Y.), Kluwer, Dordrecht, The Netherlands. p. 197 (1988).
Garnett, J. L. and Hodges, R. J., J. Chem. Soc., Chem. Commun., 4546 (1967).
Garnett, J. L. and Kenyon, R. S., J. Chem. Soc., Chem. Commun., 1227 (1971).
Gavalas, G., AIChE Journal, 26, 577 (1980).
Gavalas, G. R., Coal Science and Technology 4, Elsevier (1982).
Gerstein, B. C., Chow, C., Pembleton, R. G., Wilson, R. C., J. Phys. Chem., 81,565 (1977).
Gibson, D. T., Roberts, R., L., Wells, M. C., Kobal, V. M.,Biochem. Biophys. Res. Commun., 50, 211 (1973).
Giray, E. S. V., Chen, C., Takanohashi, T., Iino, M., Fuel, 79, 1533, Elsevier (2000).
Given, P. H., Fuel, 39, 147, Elsevier (1960).
Given, P. H., Prog. Energy Combust. Sci., 10, 149 (1984a).
Given, P. H., Coal Science, Vol. 3 (eds. Gorbaty, M. L., Larsen, J. W., Wender, I.), pp. 63, Academic Press, New
York (1984b).
Given, P. H., Marzec, A. Barton, W. A., Lynch, L. J., Gerstein, B. C., Fuel, 65, 155, Elsevier (1986).
Godo, M., Ishihara, A., Kabe, T., Energy Fuels, 11, 724 (1997a).
Godo, M., Saito, M., Sasahara, J., Ishihara, A., Kabe, T., Energy Fuels, 11,470 (1997b).
Godo, M., Umemura, M., Ishihara, A., Kabe, T., AIChE J., 43 (11), 3105 (1997c).
Godo, M., Saito, M., Ishihara, A., Kabe, T., Fuel, 77, 947, Elsevier (1998c).
Godo, M., Qian, W., Otsuki, S., Ishihara, A., Kabe, T., Miwa, S., Katahira, H., Aramaki, T., J. Jpn. Energy Inst.,
77 (3), 119 (1998a).
Godo, M., Qian, W., Otsuki, S., Ishihara, A., Kabe, T., Miwa, S., Katahira, H., Aramaki, T., J. Jpn. Energy Inst.,
77 (3), 234 (1998b).
Godo, M., Ph. D, Dissertation, Tokyo University of Agriculuture and Techonology (1998).
Goldman, M. Shen, L., Phys. Rev., 144, 321 (1966).
Gonenc, Z. S., Gibbins, J. R., Katheklakis, I. E., Kandiyoti, R., Fuel, 69, 383, Elsevier (1990).
Gonikberg, M. G. and Nikitenkov, V. E., Izv. AN SSSR, OKhN, 936 (1954).
Gorbaty, M. L., Fuel, 57, 796, Elsevier (1978).
Gorbaty, M. L., Mraw, S. C., Gethner, J. S., Brenner, D., Fuel Process. Technol., 12, 31 (1986).
Gorbaty, M. L., George, G. N., Kelemen, S. R. Fuel, 69, 1065, Elsevier (1990).
Gorbaty, M. L., George, G. N., Kelemen, S. R., Sansone, M., Energy Fuels, 5, 93 (1991).
Gordinko, S. A. and Kunz, F., Sovi. J. Ecol., 15, 90 (1984).
Gotz, G. K. E. and Fakoussa, R. M., Appl. Microbiol. Biotechnol., 52, 41 (1999).
320 References
Goyal, A., and Gidaspow, D., Prepr.,Am. Chem. Soc., Div. Fuel Chem., 27 (1), 57 (1982).
Goyal, A., Zabransky, R. F., Rehmat, A., Ind. Eng. Chem. Res., 28, 1767 (1989).
Graebert, R. and Michel, D., Fuel, 69, 826, Elsevier (1990).
Graft, R. A., Brandes, S. D., Energy Fuels, 1, 84 (1987).
Grainger, L. and Gibson, J., Coal Utilization-Technology, Economics and Policy, Graham & Trotman, London, p.
7 (1981).
Granda, M., Bermejo, J., Moinelo, S. R., Menendez, R., Fuel, 69, 702, Elsevier (1990).
Grant, D. M., Pugmire, R. J., Fletcher, T. H., Kerstein, A. R., Energy Fuels, 3, 175 (1989).
Gray, D., Lytton, M., Neuworth, M., Tomlinson, G., The Impact of Developing Technology on Indirect
Liquefaction, MTR-80W326, Mitre Corp., November, (1980).
Gray, R. J., Proc., First Int. Meeting on Coal and Coke Applied to Ironmaking, Rio De Janeiro, Brazil, p. 551
(1987).
Gray, R. J. and Campagne, P. E., Proc. AIME Ironmaking Conf., Toronto, Canada, Vol. 47, p. 313 (1988).
Green, T., Kovac, J., Brenner, D., Larsen, J. W., Coal Structure (ed. Meyers, R. A.), p. 199, Academic Press, New
York (1982).
Green, T. K., Kovac, J., Larsen, J. W., Fuel, 63, 935, Elsevier (1984).
Green, T. K. and Selby, T. D., Energy Fuels, 8, 213 (1994).
Guillen, M. D., Dominguez, A., Iglesias, M. J., Blanco, C. G., Fuel, 74, 233, Elsevier (1995).
Guillen, M. D., Iglesias, M. J., Dominguez, A., Blanco, C. G., Fuel, 74, 1595, Elsevier (1995).
Guinier, A., X-ray Diffraction in Crystals, Imperfect Crystals, and Amorphous bodies, W. H. Freeman, San
Francisco (1963).
Guin, J. A., Tarrer, A. R., Lee, J. M., Lo, L., Curtis, C. W., Ind. Eng. Chem. Proc. Des. Dev., 18, 371 (1979).
Gupta, R. K., Deobald, L. A., Crawford, D. L., Appl. Biochem. Biotechnol., 24/25, 899 (1990).
Gutmann, V., The Doner-Acceptor Approach to Molecular Interactions; Plenum Press, New York (1978).
H
Hadzi, D. and Thompson, H. W. (eds.), Hydrogen Bonding, Pergamon Press Ltd., London (1959).
Haenel, M. W., Collin, G., Zander, M., Erdoel Erdgas Kohle, 105, 71 and 131 (1989).
Haenel, M. W., Fuel, 71, 1211, Elsevier (1992).
Haider, K. M. and Martin, J. P., Soil Biochem., 20, 425 (1988).
Haider, K. and Schulten, H-R., J. Anal. Appl. Pyrol., 8, 317 (1985).
Hager, G. T., Bi, X. X., Eklund, P. C., Derbyshire, F. J., Prepr., Am. Chem. Soc., Div. Fuel Chem., 38 (1), 34
(1993).
Halchuk, R. A., Russel, W. B., Saville, D. A., Proc. Int. Conf. Coal Science, Pittsburgh, 491 (1983).
Hall, P. J., Marsh, H., Thomas, K. M., Prepr., Am. Chem. Soc., Div. Fuel Chem., 35 (2), 299 (1990).
Hall, P. J. and Larsen, J. W., Energy Fuels, 7, 47 (1993).
Halperin, W. P. and Jehng, J.-Y., Magn. Reson. Imagine, 12, 169 (1994).
Harris, L. A., Kennedy, C. R., Yust, C. S., Fuel, 58, 59, Elsevier (1979).
Harper, A. M., Meuzelaar, H. L. C., Given, P. H., Fuel, 63, 793, Elsevier (1984).
Hartmann, S. R. and Hahn, E. L., Phys. Rev., 128, 2042 (1962).
Harvey, P. J., Shoemaker, H. E., Palmer, J. M., FEBS Lett., 195, 242 (1986).
Hasegawa, T., Yokoyama, S., Sanada, Y., J. Jpn. Chem. Soc., 6, 885 (1980).
Hashimoto, K., Miura. K., Xu, J.-J., et al., Fuel, 65, 489, Elsevier (1986).
Hashimoto, K., Miura, K., Ueda, T., Fuel, 65, 1516, Elsevier (1986).
Hashimoto, K., Miura, K., Xu. J.-J., J. Jpn. Fuel Soc., 66, 418 (1987).
Hatcher, P. G., Energy Fuels, 2, 40 (1988).
Hayashi, J.-i., Nakagawa, K., Kusakabe, K., Morooka, S., Yumura, M., Fuel Process. Technol., 30, 237 (1992).
Hayashi, J.-i., Kawakami, T., Kusakabe, K., Morooka, S., Energy Fuels, 7, 118 (1993).
Hayashi, J.-i., Kawakami, T., Taniguchi, T., Kusakabe, K., Morooka, S., Yumura, M., Energy Fuels, 7, 57 (1993).
Hayashi, J.-i., Amamoto, S., Kusakabe, K., Morooka, S. Energy Fuels, 9, 1023 (1995).
Hayashi, J.-i., Mori, T., Amamoto, S., Kusakabe, K., Morooka, S., Energy Fuels, 10, 1099 (1996).
Hayashi, J.-i., Matsuo, Y., Kusakabe, K., Morooka, S., Energy Fuels, 11,227 (1997).
Hayashi, J.-i., Aizawa, S., Kumagai, H., Chiba, T., Energy Fuels, 13, 69 (1999a).
Hayashi, J.-i., Chiba, T., Energy Fuels, 13, 1230 (1999b).
Hayashi, J,-i., Norinaga, K., Kudo, N., Chiba, T., Energy Fuels., 15 (4), 903 (2001).
Heesing, A. and Mullers, W., Chem. Ber., 113, 9 (1980).
Hengel, T. D. and Walker, P. L., Fuel, 63, 1214, Elsevier (1984).
Hensel, R. P., Coal: Classification, Chemistry, and Combustion, Paper Presented at Coal-Fired Industrial Boilers
Workshop, Raleigh, NC, Fossil Power Systems, Combustion Emgineering, Inc., Windsor, CT (1981).
Hippo, E. J., Jenkins, R. G., Walker, P. L., Fuel, 58, 338, Elsevier (1979).
Hippo, E. and Walker, P. L., Fuel, 54, 245, Elsevier (1975).
References 321
Hirano, K., Nishibayashi, K., Kobayashi, M., Yoshida, Proc. the 5th China-Japan Symposium on Coal and C1
Chemistry, Huangshan, China, p. 243 (1996).
Hirano, K., Takatsu, S., Okata, T., Obayashi, M., Togawa, S., J. Jpn. Inst. Energy, 75, 909 (1997).
Hirschon, A. S. and Laine, R. M., Fuel, 64, 868, Elsevier (1984).
Hirschon, A. S., Wilson, Jr. R. B., Laine, R. M., Appl. Catal., 34, 311 (1987).
Hirschon, A. S. and Wilson, R. B., Prepr.,Am. Chem. Soc., Div. Fuel Chem., 36, 103 (1991a).
Hirschon, A. S. and Wilson, R, B., Coal Science II, Am. Chem. Soc. Washington, p. 273 (1991b).
Hirsch, P. B. Pro. Roy. Soc. Ser. A, 226, 143 (1954).
Hodek, W., Kraemer, M., Juentgen, H., Fuel Proc. Tech., 24, 417 (1990).
Hofrichter, M. and Fritsche, W., Appl. Microbiol. Biotechnol., 46, 220 (1996).
Hofrichter, M. and Fritsche, W., Appl. Microbiol. Biotechnol., 47, 419 (1997a).
Hofrichter, M. and Fritsche, W., Appl. Microbiol. Biotechnol., 47, 566 (1997b).
Hofrichter, M., Ziegenhagen, D., Sorge, S., Ullrich, R., Bublitz, F., Fritsche, W., Appl. Microbiol. Biotechnol., 52,
78 (1999).
Holker, U., Fakoussa, R. M., Hofer, M., Appl. Microbiol. Biotechnol., 44, 351 (1995).
Holker, U., Ludwig, S., Sheel, T., Hofer, M., Appl. Microbiol. Biotechnol., 52,57 (1999).
Holloway, P. H., and Nelson, G. C., Prepr.,Am. Chem. Soc., Div. Fuel Chem., 26, 1352 (1977).
Honda, H. and Ouchi, K., Fuel, 36, 159, Elsevier (1957).
Hooper, R. J., Battaerd, H. A., Evans, D. G., Fuel, 58, 132, Elsevier (1979).
Horocks, D. L., Application of Liquid Scintillation Counting, Academic Press, New York (1974).
Hu, H., Bai, J., Guo, S., Chen, G., Fuel, 81, 1521, Elsevier (2002).
Hu, W-G., Mao, J. Xing, B. Schmidt-Rohr, K., Environ. Sci. Technol., 34, 530 (2000).
Huai, H., Lo, R., Yun, Y., Meuzelaar, H. L. C., Prepr., Am. Chem. Soc., Div. Fuel Chem., 35 (3), 816 (1990).
Huang, H., Wang, K., Bodily, D. M., Hucka, V. J., Energy Fuels, 9, 20 (1995).
Hung, M. H. and Stock, L. M., Fuel, 61, 1161, Elsevier (1982).
Hung, M. H. and Stock, L. M., unpublished research.
Hurd, C. D., Macon, A. R., Simon, J. I., Levetan, R. V., J. Am. Chem. Soc., 84, 4509 (1962).
Hurst, H. M., Burges, N. A., Latter, P., Photochemistry, 1,227 (1962).
Hurt, R. H., Sarofim, A. F., Longwell, J. P., Fuel, 70, 1079, Elsevier (1991).
Htittinger, K. J. and Michenfelder, A. W., Fuel, 66, 1164, Elsevier (1987a).
Htittinger, K.J., Sperling, R. E., Proc. Int. Conf. Coal Science, Maastricht, p. 699 (1987b).
Htittinger K, J., Masling, B., Minges, R., Fuel, 65, 673, Elsevier (1986).
I
Idem, Molec. Cryst. Liq. Cryst., 63, 249 (1981).
Iijima, S, J. Crystal Growth, 50, 675 (1980).
Iimura, Y., Katayama, Y., Kawai, S., Morohoshi, N., Biosci. Biotech. Biochem., 59, 903 (1995).
Iino, M. and Matsuda, M., Fuel, 62, 744, Elsevier (1983).
Iino, M. and Matsuda, M., Bull. Chem. Soc. Jpn. 57, 3290 (1984).
Iino, M., Kumagai, J., Ito, O., J. Jpn. Fuel Soc, 64, 210 (1985).
Iino, M., Takanohashi, T., Ohsuga, H., Toda, K., Fuel, 67, 1639, Elsevier (1988).
Iino, M., Takanohashi, T., Obara, S., Tsueta, H., Sanokawa, Y., Fuel, 68, 1588, Elsevier (1989).
Iino, M., Liu, H., Hosaka, N., Kurose, H., Takanohashi, T., Prepr., Am. Chem. Soc., Div. Fuel Chem. 42 (1), 248
(1997).
Iino, M., Energy Fuels, 16, 1 (2002).
Ikenaga, N., Kan-nan, S., Sakoda, T., Suzuki, T., Catal. Today, 39, 99 (1997).
Ikoma, T., Ito, O., Tero-Kubota, S., Energy Fuels, 16, 40 (2002).
Imada, K, Kzumi, N., Mochizuki, M., Inokuchi, K., Nogami, Y., Sakawaki, K., in : Proc. 6th Japan-Australia
Joint Technical Meeting on Coal, Sapporo Japan, p. 235 (1996).
Ingold, C. K., Raisin, C. G., Wilson, C. L. J., Chem. Soc., 1637 (1936).
International Energy Outlook, DOE/ETA-0484 (98), Energy Information Administration, p. 73 (1998).
Ishihara, A., Wang, X., Shono, H., Kabe, T., Ind. Eng. Chem. Res., 32, 1723 (1993a).
Ishihara, A., Wang, X., Shono, H., Kabe, T., Energy Fuels, 7, 334 (1993b).
Ishihara, A., Takaoka, H., Nakajima, E., Imai, Y., Kabe, T., Energy Fuels, 7, 362 (1993c).
Ishihara, A., Morita, S., Kabe, T., Prepr., Am. Chem. Soc., Div. Petrol. Chem., August, 21 (1994).
Ishihara, A., Morita, S., Kabe, T., Fuel, 74, 63, Elsevier (1995).
Ishihara, A., Kawashima, K., Wang, X., Shono, H., Kabe, T., Energy Fuels, 10, 3, 726 (1996).
Ishihara, A., Sutrisna, I P., Saito, M., Qian, W., Kabe, T., Prepr., Am. Chem. Soc., Div. Fuel Chem., 44, (3), 657
(1999).
Ishihara, A., Nishigori, D., Saito, M., Qian, W., Kabe, T., Energy Fuels, 14 (3), 706 (2000).
Ishihara, A., Nishigori, D., Saito, M., Sutrisna, I P., Qian, W., Kabe, T., Energy Fuels, 15, 32 (2001).
322 References
Ishihara, A., Sutrisna, I P., Ifuku, M., Qian, E. W., Kabe, T., Energy Fuels, 16, 1483 (2002a).
Ishihara, A., Sutrisna, I P., Miura, T., Saito, M., Qian, E. W., Kabe, T., Energy Fuels, 16, 1490 (2002b).
Ishihara, A., Nishigori, D., Ohashi, Y., Kim, S., Qian, W., Kabe, T., Fuel, 81, 1409, Elsevier (2002c).
Ishikawa, Y. and Sato, S., Bull. Chem. Soc. Jpn., 52, 984 (1979).
Ishikiriyama, K., Todoki, M., Motomura, K., J. Colloid Interface Sci., 171, 92 (1995a).
Ishikiriyama, K. and Todoki, M., J. Colloid Interface Sci., 171, 103 (1995b).
Ishizuka, T., Takanohashi, T., Ito, O., Iino, M., Fuel, 72, 579, Elsevier (1993).
Ismail, I. M. K., Carbon, 25, (5), 653 (1987).
Iso, M., Chen, B. X., Eguchi, M., Kudo, T., Shrestha, S., J. Catalysis B, Enzymatic, 16, 53 (2001).
J
Jan, J. V., Antonius, P. G. K., Herman, B., Fuel, 63, 334, Elsevier (1984).
Japan International Cooperation Agency (JICA), The Final Report for the Feasibility Study on Effective Utilization
of Banko Coal in Republic of lndonesia, Volume I (1989).
Jenkins, R. G., Nandi, S., Walker., P. L., Fuel 52, 288, Elsevier (1973).
Jess, A., Fuel, 75, 1441, Elsevier (1996).
Joesten, M. D. and Drago, R. S.,J. Am. Chem. Soc., 84, 3817 (1962).
Johnson, C. A., Symposium Papers, Clean Fuels from Coal, Institute of Gas Technology, Chicago, p. 549 (1978).
Jones, D. T. and Wheeler, R. V., J. Chem. Soc., 1318 (1915).
Jones, D. T. and Wheeler, R. V., J. Chem. Soc., 707 (1916).
Joseph, J. T., Fuel, 70, 459, Elsevier (1991).
Jungten, H. and van Heek, K. H., Fuel Process. Technol., 2, 261 (1979).
Jurkiewicz, A., Marzec, A., Idziak., S., Fuel, 60, 1167, Elsevier (1981).
Jurkiewicz, A., Marzec, A., Pislewski, N., Fuel, 61,647, Elsevier (1982).
Jurkiewicz, A., Bronnimann, C. E., Maciel, G., Fuel, 68, 872, Elsevier (1989).
Jurkiewicz, A., Bronnimann, C. E., Maciel, G. E., Fuel, 69, 804, Elsevier (1990).
Jurkiewicz, A., Bronnimann, C. E., Maciel, G. E., Magnetic Resonance in Carbonaceous Solids, (eds. Botto, C. E.
and Sanada, Y.), Advances in Chemistry, ACS, Washington, DC, 2, 401 (1993).
K
Kabe, T., Nitoh, O., Kim, S., J. Jpn. Petrol. Inst., 26 (6), 424 (1983a).
Kabe, T., Nitoh, O., Kabe, Y., Ebuchi, S., J. Jpn. Petrol. Inst., 26 (6), 431 (1983b).
Kabe, T., Nitoh, O., Nagai, M., Ito, T., J. Jpn. Petrol. Inst., 26 (4), 286 (1983c).
Kabe, T., Nagai, M., Kabe, Y., J Jpn. Petrol. Inst., 26, 211 (1983d).
Kabe, T., Nitoh, O., Kim, S., Kawakami, A.,J. Jpn. Petrol. Inst., 27 (5), 392 (1984).
Kabe, T., Report on Special Project Research on Energy, SPEY 12, 41(1984).
Kabe, T., Nitoh, O., Kawakami, A., Funatsu, E., Proc. Int. Conf. Coal Science, 79 (1985a).
Kabe, T., Funatsu, E., Nitoh, O.,J. Jpn. Petrol. Inst., 28 (2), 136 (1985b).
Kabe, T., Iizuka, M., Nitoh, O., Onuma, H., Inoue, Y., J. Jpn. Energy Inst., 64 (10), 809 (1985c).
Kabe, T., Iizuka, M., Nitoh, O., Onuma, H., Inoue, Y., J. Jpn. Energy Inst., 64 (9), 741 (1985d).
Kabe, T., J. Jpn. Petrol. Inst., 29 (5), 345 (1986).
Kabe, T., Nitoh, O., Funatsu, E., Yamamoto, K., Fuel Process. Technol., 14, 91 (1986a).
Kabe, T., Nitoh, O., Kabe, Y., Kim, S., Sukie, Y., Takeuchi, M., J. Jpn. Petrol. Inst., 29 (2), 138 (1986b).
Kabe, T., Nitoh, O., Kawakami, A., Marumoto, M., Nakagawa, K., J Jpn. Inst. Energy, 653, 180 (1986c).
Kabe, T., Nitoh, O., Funatsu, E., Yamamoto, K., Fuel, 66, 1326, Elsevier (1987a).
Kabe, T., Nitoh, O., Marumoto, M., Kawakami, A., Yamamoto, Y., Fuel, 66, 1321, Elsevier (1987b).
Kabe, T., Onuma, H., Nitoh, O., Watanabe, S., Inoue, Y., J. Jpn. Inst. Energy, 66 (5), 348 (1987c).
Kabe, T., Report on Special Project Research on Energy, SPEY 16, 41(1988).
Kabe, T., Nitoh, O., Kawakami, A., Okuyama, S., Yamamoto, K., Fuel, 68, 178, Elsevier (1989a).
Kabe, T., Takaoka, H., Yamamoto, K., Ishihara, A., Horimatsu, T., Proc. 5th Int. Conf. Coal Science, p. 23
(1989b).
Kabe, T., Sawahiraki, N., Watanabe, S., Imai, Y., Yamamoto, K., Kameyama, H., J. Jpn. Inst. Energy, 69 (10),
940 (1990a).
Kabe, T., Takaoka, H., Ishiahra, A., Daita, Y., Chem. Lett., 1571 (1990b).
Kabe, T., Yamamoto, K., Ueda, K., Horimatsu, T., Fuel Process. Technol., 25, 45 (1990c).
Kabe, T., Kimura, K., Kameyama, H., Ishihara, A., Yamamoto, K., Energy Fuels, 4, 201 (1990d).
Kabe, T., Ishihara, A., Daita, Y., Nakajima, E., Imai, Y., Third China-Japan Symposium on Coal and C1
Chemistry, (1990e).
Kabe, T., Wang, X., Ishihara, A., Shono, H., Chem. Lett., 1235 (1990f).
Kabe, T., Ishihara, A., Daita, Y., Ind. Eng. Chem. Res., 30, 1755 (1991a).
Kabe, T., Horimatsu, T., Ishihara, A., Kameyama, H., Yamamoto, K., Energy Fuels, 5, 459 (1991b).
References 323
Kabe, T., Ishihara, A., Guo, S., Development of New Integrated Processes of Chinese Coal Utilization, p. 36
(1991c).
Kabe, T., Godo, M., Ishihara, A., Qian, W., Oki, H., J. Jpn. Petrol. Inst., 41 (2), 164 (1998).
Kabe, T., Godo, M., Ishihara, A., Qian, W., Otsuka, S., Mukai, K., Fuel, 77 (8), 815, Elsevier (1998a).
Kabe, T., Godo, M., Ishihara, A., Qian, W., Oki, H., J. Jpn. Petrol. Inst. 41 (2), 164 (1998b).
Kabe, T., Saito, M., Qian, W., Ishihara, A., Proc. The 1999 International Symposium on Fundamentals for
Innovative Coal Utilization, 2-4 February, p. 119 (1999).
Kabe, T., Saito, M., Qian, W., Ishihara, A., Fuel, 79 (3-4), 311, Elsevier (2000).
Kabe, Y., Bull. Tamagawagakuen Academy and Education Research Center, Shoho, 20, 20 (1992).
Kabe, Y., Research on The Environment o f The Earth, 27, 117 (1993).
Kabe, Y., Furuta, T., Takai, M., Higashi, K., Katoh, S., Kojima, T., Coal Science and Technology, 24, Coal
Science vol. II (8th ICCS), (eds. Pajares, J. A. and Tascon, J. M. D.), 1753, Elsevier (1995).
Kabe, Y., Takai, M., Furuta, T., Higashi, K., Tamagawa Univ. Research Review, 2, 127 (1996).
Kabe, Y., Ishihara, A., Iso, M., Kabe, T., Prospects for Coal Science in the 21 st Century, II (10th ICCS) (eds. Li, B.
Q. and Liu, Z. Y.), Shanxi Science & Technology Press, China, 1323 (1999).
Kaihara, M., Mametsuka, H., Gunji, N., Gohshi, Y., Fuel, 70, 931, Elsevier (1991).
Kakezawa, M., Miura, M., Nishida, T., Takahara, Y., J. Ferment. Bioeng., 75, 65 (1993).
Kamienski, B., l~'ruski, M., Gerstein, B. C., Given, P. H., Energy Fuels, 1, 45 (1987).
Kamiya, Y., Futamura, S., Mizuki, T., Kajioka, M., Koshi, K., Fuel Process. Technol., 14, 79 (1986).
Kamiya, Y., Nobukatsu, T., Futamura, S., Fuel Process. Technol., 18, 1 (1988).
Kaneko, R., Iimori, T., Miyawaki, S., Machida, M., Murakami, K., Biosci. Biotech. Biochem., 59, 1584 (1995).
Kaneko, T., Sugita, S., Tamura, M., Shimasaki, K., Makino, E., Silalahi, L. H., Fuel, 81, 1541, Elsevier (2002).
Kapteijn, F., Abbel. G., Moulijn, J. A., Fuel, 63, 1036, Elsevier (1984).
Kasaoka, S., Sakata, Y., Kayano, S., Kagaku-kogaku Ronbunshu, 8, 51 (1982).
Kasaoka, S., Sakata, Y., Shimida, M., Fuel, 66, 697, Elsevier (1987).
Katheklakis, I. E., Shin-Lin, L., Bartle, K. D., Kandiyoti, R., Fuel, 69, 172, Elsevier (1990).
Kawabata, T., Kozuru, H., Hashimoto, S., Izumiya, F., Proc. 5th China-Japan Symp. Coal C1 Chem., Huangshan
China, p. 237 (1996).
Keisch, B., Gibbon, G. A., Akhtar, S., Fuel Process. Technol., 78 (1), 269 (1978).
Keith, P. D. and Garnett, J. L., J. Am. Chem. Soc. Chem. Comm., 79 (1975).
Kelemen, S. R., George, S. N., Gorbaty, M. L., Fuel, 69, 939, Elsevier (1990).
Kelemen, S. R., Gorgaty, M. L., George, G. N., Kwiatek, P. J., Sansone, M., Fuel, 70, 396, Elsevier (1991).
Kerskaw, J. R. and Barrass, G., Fuel, 56, 455, Elsevier (1977).
Ketum, P. A., Microbiology: Concepts and Applications, Willy International Edition. (1988).
Khan, M. RI, Fuel 66, 1626, Elsevier (1987).
Kidena, K, Murata, S., Artok, L., Nomura, M., J. Jpn. Inst. Energy, 78, 869 (1999).
Kikuchi, Y., Yasukochi, Y., Nagata, Y., Fukuda, M., Takagi, M., J. Bacteriol., 4269 (1994).
Kimura, H. and Fujii, O., Chemistry and Industry of Coal, 28, Sankyo Pub. (1977).
Kimura, H. and Fujii, O., Chemistry and Industry of Coal, 102, Sankyo Pub. (1977).
Kimura, N., Kato, H., Nishi, A., Furukawa, K., Biosci. Biotech. Biochem., 60, 220 (1996).
King, H. H., Stock, L. M., Fuel, 60, 748, Elsevier (1981).
King, H. H., Stock, L. M., Fuel, 61,257, Elsevier (1982).
King, H. H., Stock, L. M., Fuel, 63, 810, Elsevier (1984).
Kirk, T. K., Croan, S., Tien, M., Enzyme Microb. Technol., 8, 27 (1986).
Kirov, N. Y., O'Shea, J. M., Sergeant, G. D., Fuel, 47, 415, Elsevier (1968).
Kiyohara, H., Torigoe, S., Kaida, N., Asaki, T., Iida, T., Hayashi, H., Takizawa, N., J. Bacteriol., 2439 (1994).
Klotzkin, M. P., Fuel, 64, 1092, Elsevier (1985).
Kobayashi, Y. and Maudsley, D. V., Biological Applications of Liquid Scintillation Counting, Academic Press,
New York (1974).
Kontchou, C. Y. and Blondeau, R., Can. J. Soil Sci., 70, 512 (1990).
Korai, Y., Sone, Y., Mochida, I., Extended Abstracts of 16th American Carbon Conference, American Carbon
Society, San Diego, p. 96 (1983).
Korai, Y. and Mochida, I., Carbon, 23, 97 (1985).
Kotanigawa, T., Shimokawa, K., Yoshida, T., Tamamoto, M., J. Phys. Chem., 83, 3020 (1979).
Kovac, J., Macromolecules, 11,362 (1978).
Krichko, A. A. and Gagarin, S. G., Fuel, 69, 885, Elsevier (1990).
Kroto, H. W. and McKay, K., Nature, 331, 328 (1988).
Kuhlman, E., Boerminkle, E., Orchin, M., Fuel, 60, 1002, Elsevier (1981).
Kuhlmann, E. J., Jung, D. Y., Guptill, R. P., Dyke, C. A., Zang, H. K., Fuel, 64, 1552, Elsevier (1985).
Kumagai, H. and Tanabe, K., Proc. Int. Symp. on Primary and Higher Order Structure of Coal and Their Influence
on Coal Reactivity, Miyagi Zao, Miyagi, Japan, p 83 (2001).
324 References
Kuo, K., Marsh, H., Broughton, D., Fuel, 66, 1544, Elsevier (1987).
Kuo, K., and Marsh, H., Prepr.,Am. Chem. Soc., Div. Fuel Chem. 34, 153 (1989).
Kurogawa, M., Baba, A., Honda, H., Ouchi, K., Coal and Coal Chemistry, Nikkan Kogyo Shimbunsha (1966).
Kyotani, T., Kubota, K., Cao, J., Yamashita, H., Tomita, A., Fuel Process. Technol., 36, 209 (1993).
L
Laborda, F., Monistol, I. F., Fernadez, M., Appl. Microbiol. Biotechnol., 52, 49 (1999).
Lacroix, M., Boutarfa, N., Guillard, C., Vrinat, M., Breysse, M., J. Catal., 120, 473 (1989).
Laine, N. R., Vastola, F. J., and Walker, P. L., J. Phys. Chem., 67, 2030 (1963).
Lambert, J. M. Jr., Fuel, 61, 777, Elsevier (1982).
Langhoff, J. Wolowski, E., Funk, O., International Coal Conference, Pretoria, Aug. 1982.
Larsen, J. W., Green, T. K., Kovac, J., J. Org. Chem., 50, 4729 (1985).
Larsen, J. W., Lee, D., Shawver, S., E., Fuel Process. Technol., 12, 51 (1986).
Larsen, J. W. and Basker, A. J., Energy Fuels, 1, 230 (1987).
Larsen, J. W., New Trends in Coal Science (ed. Yiirtim, Y.), Kluwer, Dordrecht, The Netherlands p. 85, (1988a).
Larsen, J. W., Fuel Process. Technol., 20, 13 (1988b).
Larsen, J. W., Prepr., Am. Chem. Soc., Div. Fuel Chem., 35 (2), 376 (1990).
Larsen, J. W. and Shawver, S., Energy Fuels, 4, 74 (1990b).
Larsen, J. W. and Amui, J., Energy Fuels, 8, 513. (1994).
Larsen, J.W., Hall, P., Wernett, P. C., Energy Fuels, 9, 324 (1995).
Larsen, J. W. and Gurevich, I., Energy Fuels, 10, 1269 (1996).
Larsen, J. W., Gurevich, I., Energy Fuels, 10, 1269 (1996).
Larsen, J. W., Flowers, R. A., Hall, P. J., Carlson, G., Energy Fuels, 11,998 (1997).
Lawrey, J. D., Environ. Pollut., 14, 195 (1977).
Lazarov, L. and Marinov, S. P., Fuel Process. Technol., 15, 411 (1987).
Le Claire, C. D., J. Am. Chem. Soc., 63, 343 (1941).
Lee, T. V. and Beck, S. R., AIChE Journal., 30, 517 (1984).
Leidig, E., Prusse, J.,Vorlop, K-D., Winter, J., Bioprocess Eng., 21, 5 (1999).
Lewis, I. C., Fuel, 66, 1527, Elsevier (1987).
Lewis, I. C., Carbon, 18, 191 (1980).
Lewis, J. M. and Kydd, R.A., J. Catal., 132, 465 (1991 ).
Li, C-Z., Madrali, E. S., Wu, F., Xu, B., Cai H-Y., Gtiell A. J., Kandiyoti, R., Fuel, 73, 851, Elsevier (1994).
Linares-Solano. A., Mahajan, O. P. Walker, P. L., Fuel, 58, 327 (1979).
Linares-Solano, A., Salinas-Martinez, C., Alumela-Alarcon, M., Proc. Int. Conf. Coal Science, Elsevier,
Amsterdam, p. 559 (1987).
Lin, R., Davis, A., Bensley, D. F., Derbyshire, F. J., Int. J. Coal Geol., 6, 215 (1986).
Lin, R., Davis, A., Bensley, D. F., Derbyshire, F. J., Org. Geochem., 11,393 (1987).
Li, T. and Chen, T., J. Ferment. Bioeng., 78, 298 (1994).
Liu, H.-T., Ishizuka, T., Takanohashi, T., Iino, M., Energy Fuels, 7, 1108 (1993).
Lucht, L. M. and Peppas, N. A., New Approaches in Coal Chemistry, ACS Symp. Series No 169 (eds. Blaustein,
B. D., Bockrath, B. C., Friedman, S.,), Am. Chem. Soc., Washington, DC, p. 43 (1981).
Lucht, L. M. and Peppas, N. A., Fuel, 66, 803, Elsevier (1987).
Lucht, L. M. and Peppas, N. A., Fuel, 66, 815, Elsevier (1987a).
Lucht, L. M. and Peppas, N. A., J. Appl. Polym. Sci., 33, 2777 (1987b).
Lucht, L. M., Larsen, J. W., Peppas, N. A., Energy Fuels, 1, 56 (1987).
Luss, R. and Jaakola, T. H. I.,AIChE J., 19, 760 (1973).
Lynch, L. J. and Webster, D.S., Fuel, 58, 249, Elsevier (1979).
Lynch, L. J., Webster, D. S., Sakaurovs, R., Barton, W. A., Maher, T. P., Fuel, 67, 579, Elsevier (1988).
M
Maekawa, Y., J. Jpn. Petrol. Inst., 18, 746 (1975).
Maekawa, Y., Nakata, Y., Ueda, , S., Yoshida, T., Yoshida, Y., Coal Liquefaction Fundamentals, ACS Symp.
Ser., 139, 315 (1980).
Mae, K., Miura, K., Sakurada, K., Hashimoto, K., J. Jpn. Inst. Energy, 72, 787 (1993).
Mae, K., Maki, T., Okutsu, H., Miura, K, Fuel, 79, 417, Elsevier (2000).
Makino, M. and Ueda, S., Proc. 6th Japan-Australia Joint Technical Meeting on Coal, Sapporo, Japan, p. 209
(1996).
Malhotra, R. and McMillen, D. F., Energy Fuels, 4, 184 (1990).
Malhotra, R. and McMillen, D. F., Energy Fuels, 7, 227 (1993).
Marchon, B., Carrazza, J., Heinemann, H., Somorjai, G. A., Carbon, 26, 507 (1988).
Marsh, H. and Walker, P. L. Jr., Chemistry and Physics of Carbon, 15, 229 (1979).
References 325
Martinez-Alonso, A., Perrichon, V., Tascon. J. M. D., Coal Science, (eds. Pajares, J. A. and Tascon, J. M. D.)
Elsevier, p. 571 (1995).
Martin, K. A. and Chao, S. S., Prepr.,Am. Chem. Soc., Div. Fuel Chem., 33 (3), 17 (1988).
Martino, A., Wilcoxon, J. P., Sylwester, A. P., Kawola, J. S., Prepr., Am. Chem. Soc., Div. Fuel Chem., 38 (1), 21
(1993).
Marzec, A., Juzwa, M., Betlej, K., Sobkowiak, M., Fuel Process. Technol., 2, 35 (1979).
Marzec, A., J. Anal. Appl. Pyrol., 8, 241 (1985).
Marzec, A., Fuel Process. Technol., 14, 39 (1986).
Marzec, A. and Schulten, H-R., Prepr., Am. Chem. Soc., Div. Fuel Chem., 34 (3), 668 (1989).
Matson, D. W., Linehan, J. C., Darab, J. G., Prepr., Am. Chem. Soc., Div. Fuel Chem., 38 (1), 14 (1993).
Matsui, I., Kunji, D., Furuzawa, T., Ind. Eng. Chem., 26, 91 (1987).
McCabe, M. V. and Orchin, M., Fuel, 55, 266, Elsevier (1976).
McCartney, J. T. and Ergun, S., Nature, 205, 962 (1965).
McKee, D. W. and Chatterji, D., Carbon, 113, 381 (1975).
McKee, D. W. and Chatterji, D., Carbon, 16, 53 (1978).
McKee, D. W., Fuel, 62, 70, Elsevier (1983).
McMahon, P. and Snook, I., J. Chem. Phys., 105, 2223 (1996).
McMillen, D. F., Malhotra, R., Chang, S. J., Nigenda, S. E., Int. Conf. Coal Sci. Proc., Pergamon Press, Sydney, p.
91 (1985).
McMillen, D. F., Malhotra, R., Chang, S. J., Ogier, W. C., Nigenda, S. E., Fleming, R. H., Fuel, 66, 1611, Elsevier
(1987).
McMillen, D. F., Malhotra, R., Hum, G. P., Chang, S. J. E., Energy Fuels, 1, 193 (1987).
McMillen, D. F., Malhotra, R., Serio, M. A., Prepr., Am. Chem. Soc., Div. Fuel Chem., 35 (2), 430 (1990).
McMillen, D. F., Malhotra, R., Tse, D. S., Energy Fuels, 5, 179 (1991).
McPherson, W. P., Foster N. R., Hastings, D. W., Kalman, J. R., Gilbert, T. D., Fuel, 64, 457, Elsevier (1985).
Melander, L., Isotope Effect from Reaction Rates, Ronald Press, New York. (1960).
Mendez-Vivar, J., Campero, A., Livage, J., Sanchez, C., J. Non-Crysttal. Solids, 121, 26 (1990).
Meuzelaar, H. L. C., Haverkamp, J., Hileman, F. D., Curie-point Pyrolysis Mass Spectrometry of Recent and
Fussil Biomaterials, Compendium and Atlas, Elsevier, Amsterdam (1982).
Meuzelaar, H. L. C., Harper, A. M., Hill, G. R., Given, P. H., Fuel, 63, 640, Elsevier (1984a).
Meuzelaar, H. L. C., Harper, A. M., Pugmire, R. J., Karas, J., Int. J. Coal Geol., 4, 143 (1984b).
Meuzelaar, H. L. C., Windig, W., Harper, A. M., Huff, S. M., McClennen, W. H., Richards, J. M., Science, 226,
268 (1984c).
Meuzelaar, H. L. C., Hoesterey, B. L., Windig, W., Hill, G. R., Fuel Process. Technol., 15, 59 (1987).
Meuzelaar, H. L. C., Yun, Y., Chakravarty, T., Metcalf, G. S., Advances in Coal Spectroscopy (ed. Meuzelaar, H.
L. C.), Plenum, New York, p. 275 (1992).
Metcalf, G. S., Windig, W., Hill, G. R., Meuzelaar, H. L. C., Int. J. Coal Geol., 7, 245 (1987).
Millward, G. R. and Jefersson, D. A. Chemistry and Physics of Carbon, Marcel Dekker, New York, Vol. 14, p. 1
(1978).
Millward, G. R. in : Coal and Modern Coal Processing, An introduction. Academic Press, New York; pp. 87
(1979).
Mims, C. A., Rose, K. D., Melchior, M. T., Pabst, J. K., J. Am. Chem. Soc., 104, 6886 (1982).
Mims, C. A. and Pabst, J. K., Fuel, 62, 176, Elsevier (1983).
Miura, K., Hashimoto, K., Xu, J.-J., J. Jpn. Fuel Soc, 66, 264 (1987).
Miura, K., Hashimoto, K., Silveston, P., Fuel, 68, 1461, Elsevier (1989).
Miura, K., Mae, K., Asaoka, S., Yoshimura, T., Hashimoto, K., Energy Fuels, 5, 340 (1991).
Miura, K., Mae, K., Yoshimura, T., Masuda, K., Hashiuioto, K., Energy Fuels, 5, 803 (1991).
Miura, K., Mae, K., Murata, A., Sato, A., Sakurada, K., Hashimoto, K., Energy Fuels, 6, 179 (1992a).
Miura, K., Mae, K., Sakurada, K., Hashimoto, K., Energy Fuels, 6, 16 (1992b).
Miura, K., Mae, K., Monkawa, H., Hashimoto, K., Fuel, 73, 443, Elsevier (1994a).
Miura, K., Mae, K., Takebe, S., Wakiyasu, H., Hashimoto, K., Energy Fuels, 8, 874' (1994b).
Miura, K., Mae, K., Wakiyasu, H., Hashimoto, K., Kagaku Kogaku Ronbunshu, 20, 926 (1994c).
Miura, K., Mae, K., Morozumi, F, Prepr., Am. Chem. Soc., Div. Fuel Chem., 42 (1), 209 (1997).
Miura, K., Mae, K., Li, W., Kusakawa, T., Kumano, A., Prepr., Am. Chem. Soc., Div. Fuel Chem., 44 (3), 642
(1999).
Miura, K., Fuel Process. Technol., 62 (2-3), 119 (2000).
Miura, K., Mae, K., Li, W., Kusakawa, T., Morozumi, F., Kumano, A., Energy Fuels, 15 (3), 599 (2001).
Miura, K., Mae, K., Hasegawa, I., Chen, H., Kumano, A., Tamura, K., Energy Fuels, 16, 23 (2002a).
Miura, K., Mae, K., Ashida, R., Tamura, T., Ihara, T., Fuel, 81, 1417, Elsevier (2002b).
Mizumoto, M., Yamashita, H., Matsuda, S., Ind. Eng. Chem. Prod. Res. Dev., 24, 394 (1985).
Mobley, D. P. and Bell, A. T., J. Catal., 64, 494 (1980).
326 References
Mochida, I., Kudo, K., Takeshita, K., Takahashi, R., Suetsugu, Y., Furumi, J., Fuel, 53, 253, Elsevier (1974).
Mochida, I., Kudo, K., Fukuda, M., Takeshita, K., Carbon, 13, 135 (1975).
Mochida, I. and Marsh, H., Fuel, 58, 797, Elsevier (1980).
Mochida, I., Matsuoka, H., Fujitsu, H., Korai, Y., Takeshita, K., Carbon, 19, 213 (1981).
Mochida, I., Korai, Y., Fujitsu, H., Takeshita, K., Komatsubara, Y., Koba, K., Fuel, 60, 1083, Elsevier (1981b).
Mochida, I., Tamaru, K., Korai, Y., Fujitsu, H., Takeshita, K., Carbon, 20, 231 (1982).
Mochida, I., Oishi, T., Korai, Y., Fujitsu, H., Ind. Eng. Chem. Prod. Res. Dev., 23, 203 (1984).
Mochida, I. and Korai, Y., J. Jpn. Fuel Soc., 64, 796 (1985).
Mochida, I., Yufu, A., Sakanishi, K., Korai, Y., Shimohara, T., J. Jpn. Fuel Soc., 65, 1020 (1986).
Mochida, I., Sone, Y., Korai, Y., Fujitsu, H., J. Jpn. Petrol. Inst., 26, 101 (1987).
Mochida, I., Ueno, I., Korai, Y., Fujitsu, H., Sakanishi, K., J. Jpn. Petrol. Inst., 30, 31 (1987).
Mochida, I., Zhao, X. Z., Sakanishi, K., Fuel, 67, 1101, Elsevier (1988a).
Mochida, I., Yufu, A., Sakanishi, K., Korai, Y., Fuel, 67, 114, Elsevier (1988b).
Mochida, I., "Chemistry and Chemical Engineering of Carbon Materials". Asakura Inc., p. 162 (1990).
Mochida, I., Zhao, X. Z., Sakanishi, K., Ind. Eng. Chem. Res., 29, 234 (1990a).
Mochida, I., Takayama, A., Sakata, R., Sakanishi, K., Energy Fuels, 4, 398 (1990b).
Mochida, I. and Sakanishi, K., Fuel, 79, 221, Elsevier (2000).
Mochida, I., Priyanto, U., Murti, S., Sakanishi, K., Okuma, O., Aramaki, T., Onozaki, M., J Jpn. Inst. Energy,
80 (4), 237 (2001).
Mojelsky, T. W., Ignasiak, T. M., Frakman, Z., Mclntyre, D. D., Lown, E. M., Montgomery, D. S., Strausz, O. P.,
Energy Fuels, 6, 83 (1992).
Monib, M., Hosny, I., Zohdy, L., Khalafallah, M., Central. Bakteriol. 136, 189 (1981).
Montano, P. A. and Granoff, B., Fuel, 59, 214, Elsevier (1980).
Montano, P. A., Bommannavar, A. S., Shah, V., Fuel, 60, 703, Elsevier (1981).
Montano, P. A., Stenberg, V. I., Sweeny, P.,J. Phys. Chem., 90, 156 (1986).
Montgomery, R. S., Holly, E. D., Gohlke, R. S., Fuel, 35, 49, Elsevier (1956).
Morales. F., Garzon, F. J. L., Peinado, A. L. et al., Fuel, 64, 666, Elsevier (1985).
Morales, A., Ramirez, M., de Agudelo, M., Hemandez, F.,Appl. Catal., 41,261 (1988).
Morii, H., Nakamiya, K., Kinoshita, S., J. Ferment. Bioeng., 80, 296 (1995).
Moritomi, H., Nagaishi, H., Haruse, M., Sanada, Y., Chiba, T., J. Jpn. Fuel Soc. 62, 254 (1979).
Moulijn, J. A. and Kapteijn, F., Carbon, 33, 1155 (1995).
Moulijn, J. A., Cerfontain, M. B., Kapteijn, P., Fuel, 3, 1043, Elsevier (1984).
Mraw, S. C. and Naas-O'Rourke, D. F., Science, 205, 901 (1979).
Muheim, A., Waldner, R., Leisola, M. S. A., Fiechter, A., Enzyme Microb. Technol., 12, 204 (1990).
Mukherjee, D. K., Mitra, J. R., Fuel, 63, 722, Elsevier (1984).
Murakami, H., J. Jpn. Fuel Soc., 66, 448 (1987).
Murakata, T., Saito, Y., Yoshikawa, T., Suzuki, T., Sato, S., Fuel, 34, 1436, Elsevier (1993).
Murata, S., Uesaka, K., Inoue, H., Nomura, M., Energy Fuels, 8, 1379 (1994).
Murata, S., Artok, L., Kidena, K, Nomura, M., Primary and Higher Order Structures of Coal and Their Influence
on Coal Reactivity-Final Report on "Research for the Future" Coal Research Project-, (ed. lino, M.) p. 1
(2001)
Mu, R. and Malhotra, V. M., Fuel, 70, 1233, Elsevier (1991).
Myers, R. A., Coal Structure, Academic Press, New York (1982).
N
Nakagawa, T., Tajibara, H., Yamashita, B., Zhang, D., Okura, A., Proc. of 32nd Conf. Japan Petrol. Inst., p. 55
(1985).
Nakamura, K., Takanohashi, T., Iino, M., Kumagai, H., Sato, M. Yokoyama, S., Sanada, Y., Energy Fuels, 9,
1003 (1995).
Naumann, A. W., Behan, A. S., Thorsteinson, E. M., in : Proc. Int. Conf. Chem. and Uses of Molybdenum (eds.
Bany, H. E. and Mitchel, C. N.), p. 313 (1982).
Neavel, R. C., Phil. Trans. R. Soc. London, A300, 1961 ( 1981 ).
Neavel, R. C., Fuel, 55, 237, Elsevier (1976).
Neavel, R. C., Coal Structure (eds. Gorbaty, M. L., Larsen, J. W., Wender, I.), Academic Press, New York p. 1
(1982).
Nelson, J. R. Fuel, 62, 112, Elsevier (1983).
Nelson, J. R. Fuel, 33, 381, Elsevier (1954).
Neuworth, M. B. and Moroni, E. C., Proc. Int. Conf. Coal Science, DiJsseldorf, p. 542 (1981).
Nimz, H.,Angew. Chem., 77, 336 (1974).
Nip, M., de Leeuw, J. W., Crelling, J. C., Energy Fuels, 6, 125 (1992).
Nishioka, K. and Yoshida, S., Transactions oflSIJ, 23, 475 (1983).
References 327
Nishioka, M., Prepr.,Am. Chem. Soc., Div. Fuel Chem., 34 (3), 685 (1989).
Nishioka, M. and Larsen, J. W., Energy Fuels, 4, 100 (1990a).
Nishioka, M. and Gorbaty, M. L., Energy Fuels, 4 (1), 70 (1990b).
Nishioka, M. and Larsen, J. W., Prepr., Am. Chem. Soc., Div. Fuel Chem., 35 (2), 319 (1990c).
Nishioka, M., Fuel, 71, 941. Elsevier (1992).
Nishioka, M., Fuel, 72, 1719. Elsevier (1993).
Nishizawa, Y., Nakabayashi, K., Shinagawa, E., J. Ferment. Bioeng., 80, 91 (1995).
Nitoh, O., Kabe, T., Kabe, Y., Bunsekikagaku, 34 (5), 290 (1985).
Norinaga, K., Hayashi, J.-i., Chiba, T., Cody, G. D. Energy Fuels, 12, 1239 (1999a).
Norinaga, K., Hayashi, J.-i., Kudo, N., Chiba, T., Energy Fuels, 13, 1058 (1999b).
Norinaga, K., Kumagai, H., Hayashi, J.-i., Chiba, T., Energy Fuels, 12, 574 (1998a).
Norinaga, K., Kumagai, H., Hayashi, J.-i., Chiba, T., Energy Fuels, 12, 1013 (1998b).
Norinaga, K., Kuniya, M., Iino, M., Energy Fuels, 16, 62 (2002).
Norinaga, K,, Iino, M., Gody, G, F., Energy Fuels, 14 (6), 1245 (2000).
O
Obara, T., Yokono, T., Miyazawa, K., Sanada. Y., Carbon, 19, 263 (1980).
Obara, T., Yokono, T., Sanada, Y., Marsh, H., Fuel, 64, 995, Elsevier (1985).
Oberlin, A. A., Chemistry and Physics of Carbon, Marcel Dekker, New York, Vol. 22, p. 1 (1989).
Oberlin, A. Carbon, 1% 7 (1979).
O'Brien, R. J., Gibbons, J. R., Kandiyoti, R., Fuel Process. Technol., 15, 71 (1987).
Oda, Y., J. Jpn. Inst. Energy, 75, (7), 512 (1996).
Oele, A. P., Waterman, H. I., Goedkoop, M. L., van Krevelen, D. W., Fuel, 30, 169, Elsevier (1951).
Office of New Sunshine Project, New Sunshine Program., AIST, Ministry of International Trade and Industry,
Japan (1995).
Ofosu-Asante, K., Stock, L. M., Zabransky, R. F., Fuel, 68, 567, Elsevier (1989).
Ogata, E., Nomi, T., Tanaka, T., Watanabe, S., Kamiya, Y., J. Jpn. Fuel Soc., 62, 131 (1983).
Ogawa, T., Stenberg, V. I., Montano, P. A., Fuel, 63, 1660, Elsevier (1984).
Ogo, Y., Kuranuki, K., Tachibana, E., J. Jpn. Chem. Soc., 9, 1721 (1985).
Ohe, S., Itoh, H., Makabe, M., Ouchi, K., Fuel, 64, 1108, Elsevier (1981).
O'Neill, H. E., Benson, S. W., Vol. II, Chap. 17., John Wiley & Sons, Inc. New York (1973).
Orendt, A. M., Solum, M. S., Sethi, N. K., Pugmire, R. J., Grant, D. M., Advances in Coal Spectroscopy (ed.
Meuzelaar, H. L. C.), Plenum Press, New York, p. 215 (1992).
Osawa, Y., Seki, E., Kakuda, T., Fujishiro, S., Koukashi, Kogyokagaku Zasshi, 73, 2212 (1979).
Oshima, S., Yumura, M., Kuriki, Y., Uchida, K., Kamaya, K., Igazaki, F., J. Jpn. Inst. Energy, 76, 897 (1997).
Otake, Y. and Suuberg, E. M., Energy Fuels, 11, 1155 (1997).
Otake, Y. and Suuberg, E. M., Fuel, 77, 901, Elsevier (1998).
Otani, S., Carbon, 3, 31 (1965).
Otani, S., Okuda, K., Matsuda, H. S., Carbon Fiber, 231 (1983).
Otani, S., Oya, A., Carbon Fiber Nyumon, 139 (1993).
Ouchi, K., Makabe, M., Fuel, 67, 1536, Elsevier (1988).
Ouchi, K., Itoh, S., Makabe, M., Itoh, H., Fuel, 69, 735, Elsevier (1989).
Oya, A., Tanso (Carbon), 102, 118 (1980).
P
Painter, P., Energy Fuels, 6, 863 (1992).
Painter, P. C., Synder, R. W., Starsinic, M., Coleman, M. M., Kuehn, D. W., Davis, A., in: "Coal and Coal
Products: Analytical Characterization Technique", (ed. Fuller, E. L.), ACS, Washinton, D. C., Vol. 205,
Chap. 3, p. 47 (1982).
Painter, P. C, Starsinic, M., Coleman, M. M., in : Fourier Transform Infrared Spectroscopy, Vol. 4, (eds. Ferraro,
J. R. and Basile, L. J.), Academic Press, Orlando (1985).
Painter, P. C., Park, Y., Coleman, M., Energy Fuels, 2, 693 (1988).
Painter, P. C., Sobkowiak, M., Youtcheff, J., Fuel, 66, 973, Elsevier (1987).
Painter, P. C., Park, Y., Graf, J., Sobkowiak, M., Coleman, M., Prepr., Am. Chem. Soc., Div. Fuel Chem. 35 (2),
307 (1990).
Pal, N., Lewandowski, G., Armenante, P. M., Biotechnol. Bioeng. 46, 599 (1995).
Papirer, E., Li, S. and Donnet, J.-B., Carbon, 25, 243 (1987).
Paradhan, V. R., Herrick, D. E., Tierney, J. W., Wender, I., Energy Fuels, 5, 712 (1991).
Park, Y. D., Korai, Y., Mochida, I., J. Mater. Sci., 21,424 (1986).
Patrick, J. W., Reynold, M. J., Walker, A., Fuel, 62, 129, Elsevier (1983).
Pelofsky, A. H. ed. Coal Conversion Technology, A CS Symp. Ser., No. 110, Am. Chem. Soc., Washington DC,
328 References
(1979).
Penninger, J. L. M., Int. J. Chem. Kinet., 14 (7), 761 (1982).
Peppas, N. A. and Lucht, L. M., Chem. Eng. Commun., 30, 2910 (1984).
Pestryakov, B. V., Solid Fuel Chem., 20 (6), 3 (1986).
Peters, B. C., Dow Chemical Co., US/ERDA/FE1534-48 (1977).
Peters, W., Fourth International Conference on Coal Research, Vancouver (1978).
Petrakis, L. and Grandy, D. W., Fuel, 60, 1017, Elsevier (1981).
Pfeifer, F., Schacht, S., Klein, J., Truper, G.,Arch. Microbiol., 152, 515 (1989).
Pfeifer, F., Truper, G., Klein, J., Schacht, S., Arch. Microbiol. 159, 323 (1993).
Pier, M., DE 542992 (1929).
Pine, A., Gibby, M. G., Waugh, J. S., J. Chem. Phys., 59, 569 (1973).
Pitt, C. J. and Millward, G.R., Coal and Modern Coal Processing: An Introduction, Academic Press, New York
(1979).
Poutsma, M. L., "Assessment of Advanced Process Concepts for Liquefaction of Low H2:CO Ratio Synthesis Gas
Based on the Kolbel Slurry Reactor and the Mobil-Gasoline Process," ORNL-5365, February, 1980.
Poutsma, M. L. and Dyer, D. W., J. Org. Chem., 47, 4903 (1982a).
Poutsma, M. L., Youngblood, E. L., Oswald, G. E., Cochran, H. D., Fuel, 61,314, Elsevier (1982b).
Poutsma, M. L., Energy Fuels, 4, 113 (1990).
Powles, J. G. and Mansfield, P., Phys. Lett., 2, 58 (1962).
Pradhan, V. R., Herrick, D. E., Tierney, J. W., Wender, I., Energy Fuels, 5, 712 (1991).
Pradhan,V. R., Hu, J., Tierney, J. W., Wender, I., Prepr., Am. Chem. Soc., Div. Fuel Chem., 38 (1), 8 (1993).
Purcell, K. F. and Drago, R.S., J. Am. Chem. Soc., 89, 2874 (1967).
Q
Qian, S. A. and Ling, L. C., Fuel, 69, 377, Elsevier (1990).
Qian, W., Ishihara, A., Fujimura, H., Saito, M., Godo, M., Kabe, T., Energy Fuels, 11 (6), 1288 (1997).
Quinga, E. M. Y. and Larsen, J. W., New Trends in Coal Science (ed. Ytiriim, Y.), Kluwer, Dordrecht, The
Netherlands, p. 85 (1988).
R
Radke, M., Lwythaeuser, D., Teichmueller, M., Org. Geochem. 6, 423 (1984).
Radke, M., Willsch, H., Leythaeuser, D., Teichmueller, M., Geochim. Cosmochim. Acta, 46, 1831 (1982).
Radovic, L. R., Walker, P. L., Jenkins, R. G., Fuel, 62, 209, Elsevier (1983a).
Radovic, L. R., Walker, P. L., Jenkins, R. G., Fuel, 62, 849, Elsevier (1983b).
Radovic, L. R., Walker, P. L., Jenkins, R. G., J. Catal., 82, 382 (1983c).
Radovic, L. R. and Walker, P. L., Fuel Process. Technol., 8, 149 (1984a).
Radovic, L. R., Walker, P. L., Jenkins, R. G., Fuel, 63, 1028, Elsevier (1984b).
Radovic, L. R., Steczko, K., Walker, P. L., Jenkins, R. G., Fuel Process. Technol., 10, 311 (1985).
Ralph, J. P. and Catcheside, D. E. A.,Appl. Microbiol. Biotechnol., 42, 536 (1994).
Ralph, J. P., Graham, L. A., Catcheside, D. E. A., Appl. Microbiol. Biotechnol., 46, 226 (1996).
Ralph, J. P., Catcheside, D. E. A., Appl. Microbiol. Biotechnol., 52, 70 (1999).
Ramachandran, P. A. and Doraiswamy, L. K., AIChE J., 28, 881 (1983).
Ratcliffe, C. T., and Vaughn, S. N., Prepr., Am. Chem. Soc., Div. Fuel Chem., 30, 304 (1985).
Reddy, G. V. B., Joshi, D. K., Gold, M. H., Microbiology, 143, 2353 (1997).
Reidelbach, H. and Summerfield, M., Prepr.,Am. Chem. Soc., Div. Fuel Chem. 20 (1), 161 (1975).
Reidelbach, H. and Algermisser, J., Proc. 13 'h Intersoc. Energy Conversion Conf., San Diego, 1,469 (1978).
Ren, R., Itoh, H., Makabe, M., Ouchi, K., Fuel, 66, 643, Elsevier (1987).
Rhoads, T. M., Mikell, A. T. Jr., Eley, M. H., Can. J. Microbiol., 41,592 (1995).
Robinson, R. C., Frumkin, H. A., Sullivan, R. F., Energy Progress, 3 (3), 163. (1983).
Rochdi, A. and Landais, P., Fuel, 70, 367, Elsevier (1991).
Rogoff, M. H., Wender, I., J. Bacteriol., 73, 264 (1957).
Romey, I., Weigand, Technical and Economic Prospects of Coal Liquefaction in the Federal Republic of
Germany, Essen, 1981.
Rosa, L.D., Pruski, M., Gerstein, B. C.,Advances in Chemistry Series 229, Am. Chem. Soc., 359 (1993).
Rottendorf, H. and Wilson, M. A., Fuel, 59, 175, Elsevier ( 1981 ).
Ruchardt, C., Gerst, M., Ebenhoch, J., Angew. Chem. Intl. Ed. Engl., 36, 1406 (1997).
Ruckenstein, E. and Wang, X-B., Biotechnol. Bioeng. 44, 79 (1994).
Runnion, K., Combie, J. D., Appl. Biochem. Biotechnol., 24/25, 817 (1990).
Ryan, T. A. and Stacey, M. H., Fuel, 63, 1101, Elsevier (1984).
References 329
S
Saber, J. M., Falconer, J. L., Brown, L. F., Fuel, 65, 1356, Elsevier (1986).
Saber, J. M., Kester, K. B., Falconer, J. L., Brown, L. F., J. Catal., 109, 329 (1988).
Saiki, H., Kimura, K., Ohmura, N., 1st Conference on Coal Utilization Technology, Abstract, Center for Coal
Utilization Japan, p. 342 (1991).
Sakabe, T., Shigishihoukoku, 49 (1961).
Sakanishi, K., Hasuo, H., Kishino, M., Mochida, I., Energy Fuels, 10, 216 (1996).
Sakanishi, K., Hasuo, H., Nagamatsu, T., Mochida, I., Proc. Int. Conf. Coal Science. Essen, Germany, p. 1421
(1997).
Sakanishi, K., Watanabe, I., Ishom, F., Mochida, I., Proceedings, Int. Symp. on "Primary and Higher Order
Structures of Coal and Their Influence on Coal Reactivity", p. 166 (2001).
Sakanishi, K., Saito, I., Ishom, F., Watanabe, I., Mochida, I., Okuyama, N., Deguchi, T., Simazaki, K., Fuel, 81,
1471, Elsevier (2002).
Sakata, R., Takayama, A., Sakanishi, K., Mochida, I., Energy Fuels, 4, 585 (1990).
Sakaurovs, R. Lynch, L. J., Barton, W. A., Prepr., Am. Chem. Soc., Div. Fuel Chem., 34 (3), 702 (1989).
Samaras, P., Diamadopoulos, E., Sakellaropoulos, G. P., Fuel, 75 (9), 1108, Elsevier (1996).
Sams, D. A., and Shadman, F.,,4IChE J., 32, 1132 (1986).
Sanada, Y., Sasaki, M., Ohara, T., Chiba, Y., Nagaishi, H., Yokoyama, S., Sato, M., Transfer Technology of Coal
- Liquefaction of coal, Composition Structure and Properties Estimation of Coal-derived Oil-, p. 21, IPC,
Tokyo (1994).
Sanada, Y. and Honda, H., Fuel, 45, 295, Elsevier (1966).
Sanada, Y., Energy Fuels, 16, 3 (2002).
Sanada, Y., Sasaki, M., Kumagai, H., Aizawa, S., Nishizawa, T., Mineo, T., Chiba, T., Fuel, 81, 1397, Elsevier
(2002).
Sasaki, M. and Sanada, Y., J. Jpn. Petrol. Inst., 34, 218 (1991).
Sato, H. and Akamatsu, H., Fuel, 33, 195, Elsevier (1954).
Sato, Y., Yamamoto, Y., Miki, K., J. Jpn. Petrol. Inst., 35, 274 (1992).
Schaefer, J. and Stejskal, E. 0., J. Am. Chem. Soc., 98, 1031 (1976).
Schaub, G., Peters, W. A., Howard, J. B., Proc. Int. Conf. Coal Science, Dusseldorf, 229 (1981).
Schmal, M., Monteiro, J. L. F., Toscani, H., Ind. Eng. Chem. Proc. Des. Dev., 22, 563 (1983).
Schmid, B. K., Jackson, D. M., Philos. Trans. R. Soc. London, Ser. ,4 300, 129 (1981).
Schobert, H., 6th Japan-China Syrup Coal and Cl Chem., Zao, p. 358 (1998).
Schopf, J. M., Econ. Geol., 51, 521 (1956).
Schroeder, K., Bockrath, B., Miller, R., Davis, H., Energy Fuels, 11, 221 (1997).
Schulten, H-R., Simmleit, N., Mtiller, R.,Anal. Chem., 59, 2903 (1987).
Schulten, H-R., Simmleit, N., Marzec, A., Fuel, 67, 619, Elsevier (1988).
Schulten, H-R., Marzec, A., Dyla, P., Simmleit, N., Mtiller, R., Energy Fuels, 3, 481 (1989a).
Schulten, H-R., Simmleit, N., Mtiller, R.,Anal. Chem., 61, 221 (1989b).
Schumacher, J. D., Ph. D. Thesis, Univ. Bonn, Germany (1997).
Schumacher, J. D., Fakoussa, R. M.,,4ppl. Microbiol. Biotechnol., 52, 85 (1999).
Schumacher, J. P., van Vucht, H. A., Groenewege, M. P., Blom, L., van Krevelen, D. W., Fuel, 35, 281, Elsevier
(1956).
Schuster, P., Zundel, G., Sandorfy, C. eds., The Hydrogen Bond, Recent Developments in Theory and Experiments,
North Holland Publishing Co., N.Y., Oxford, (1976).
Schweighardt, F. K., Bockrath, B. C., Friedel, R. A., Retcofsky, H. C.,Anal. Chem, 48, 1254 (1976).
Serio, M. A., Hamblen, D. G., Markham, J. R., Solomon, P. R., Energy Fuels, 1, 138 (1987).
Serio, M. A., Peters, W. A., Howard, J. B., Ind. Eng. Chem. Res., 26, 1831 (1987).
Serio, M. A., Solomon, P. R., Suuberg, E. M., Proc. Int. Conf. on Coal Science, Elsevier, Amsterdam, p. 597
(1987).
Serio, M. A., Kjoo, E., Charpenay, S., Solomon, P. R., Prepr., Am. Chem. Soc., Div. Fuel Chem., 37 (4) 1681
(1992).
Seshadri, K. S., Bacha. J. D., Albaugh. E. W., Fuel, 61, 1095, Elsevier (1980).
Sharma, A., Kyotani, T., Tomita, A., Fuel, 78, 1203, Elsevier (1999a).
Sharma, A., Kyotani, T., Tomita, A., Prepr.,,4m. Chem. Soc., Div. Fuel Chem., New Orleans, 43 (4), 960 (1999b).
Sharma, A., Kyotani, T., Tomita, A., Energy Fuels, 14 (2), 515, Elsevier (2000a).
Sharma, A., Kyotani, T., Tomita, A., Energy Fuels, 14 (6), 1219, Elsevier (2000b).
Sharma, A., Kyotani, T., Tomita, A., Fuel, 80, 1467, Elsevier (2001).
Sharma, A., Kadooka, H., Kyotani, T., Tomita, A., Energy Fuels, 16, 54 (2002).
Sheel, T., Ludwig, S., Hofer, M., Holker, U., ,4ppl. Microbiol. Biotechnol., 52, 66 (1999).
Shen, J., Stiegel, G., Bose, A. C., Fuel Sci. Tech. Int'l., 14 (4), 559 (1996).
Shimada, Y., Koga, C., Sugihara, A., Nagao, T., Takada, N., Tsunasawa, S., Tominaga, Y., J. Ferment. Bioeng.,
330 References
T
Takagi, H., Isoda, T., Kusakabe, K., Morooka, S., Energy Fuels, 16, 12 (2002).
Takanohashi, T. and Iino, M., Energy Fuels, 4, 452 (1990).
Takanohashi, T. and Iino, M. Energy Fuels, 9, 1003 (1995).
Takanohashi, T., Iino, M., Nishioka, M., Energy Fuels, 9, 788 (1995).
Takanohashi, T., Yoshida, T., Iino, M., Katoh, K., Fukada, K., Energy Fuels, 12, 913 (1998).
Takanohashi, T., Nakamura, K., Iino, M., Energy Fuels, 13, 922. (1999).
Takanohashi, T., Terao, Y., Iino, M., Yun, Y., Suuberg, E. M., Energy Fuels, 13, 506 (1999).
Takanohashi, T., Terao, Y., Yoshida, T., Iino, M., Energy Fuels, 14, 915 (2000a).
Takanohashi, T., Nakano, K., Yamada, O., Kaiho, M., Ishizuka, A., Mashimo, K., Energy Fuels, 14, 720 (2000b).
Takanohashi, T. and Kawashima, H., Prepr., Am. Chem. Soc., Div. Fuel Chem., 45, 238 (2000c).
Takanohashi, T., Terao, Y., Iino, M., Fuel, 79, 349, Elsevier (2000d).
Takanohashi, T., Kawashima, H., Yoshida, T., Iino, M., Energy Fuels, 16, 6 (2002).
Takanohashi, T., Xiao, F., Yoshida, T., Saito, I., Energy Fuels, 17, 255 (2003).
Takarada, T., Ohtsuka, Y., Tomita. A., Tamai, Y.,J. Jpn. Fuel Soc. 62, 414 (1983).
Takarada, T., Tamai, Y., Tomita, A., Fuel, 64, 1438, Elsevier (1985a).
Takarada, T., Tamai, Y., Tomita, A., Fuel, 65, 679, Elsevier (1985b).
Takarada, T., Tonishi, T., Takezawa, H., Kato, K., Fuel, 71, 1087, Elsevier (1992).
Takemura, Y., Satio, Y., Okada, K., Koinuma, Y., Energy Fuels, 3, 342 (1989).
Takeuchi, M., Shono, H., Wang, X., Ishihara, A, Kabe, T.,J. Jpn. Petrol. Inst. 37, (2), 136 (1994).
Tanaka, H. and Nishi, T., Phys. Rev. B, 33, 32 (1986).
Tanaka, H., Maruyama, K., Yasuda, E., Kimura, S., Tanso (Carbon), 86, 107 (1986).
Tanaka, S., Uemura, T., Ishizaki, K., Nagayoshi, K., Ikenaga, N., Ohme, H., Suzuki, T., Energy Fuels, 9, 45
(1995).
Temp, U., Meyrahn, H., Eggert, C., Biotechnol. Lett., 21, 281 (1999).
Thiyagarajan, P., Epperson, J. E., Crawford, R. K., Carpenter, J. M., Hjelm, R. P., Proceedings of International
Seminar on Structural Investigation on Pulsed Neutron Sources Dubna, p. 194 (1993).
Thomas, A. F., Deuterium Labeling in Organic Chemistry, Appleton-Century-Crofts, New York, p. 204 (1971).
Tian, D., Sharma, R. K., Stiller, A. H., Stinespring, C. D., Dadyburjor, D. B., Fuel, 75, 751, Elsevier (1996).
332 References
U
Ueda, S., Nakata, Y., Yoshida, T., Maekawa, Y., Proc. Int. Conf. on Coal Science, Dusseldorf, p. 380 (1981).
Ugarte, D., Nature, 359, 707 (1992).
Unger, P. E. and Suuberg, E. M., Proc. Int. Conf. Coal Science, Pittsburgh, 475 (1983a).
Unger, P. E. and Suuberg, E. M., Prepr., Am. Chem. Soc., Div. Fuel Chem. 28 (4), 278 (1983b).
V
Vahrman, M., Chem. Br., 8, 16 (1972).
Van Heek, K. H. and Jungten, H., Prepr.,Am. Chem. Soc., Div. Fuel Chem., 29 (2), 2 (1984).
Van Heek, K. H. and Mtihlen, H. J., Fuel, 64, 1405, Elsevier (1985).
Van Heek, K. and Miihlen, H. J., Fuel Process. Technol., 15, 113 (1987).
Van Krevelen, D. W., Coal : Typology, Chemistry, Physics, and Constitution, Elsevier, New York (1981).
Van Krevelen, D. W., Fuel, 44, 229, Elsevier (1965).
Vastola, F. J. and Walker, P. L., J. Chem. Phys., 58, 20 (1960).
Veraa, M. J. and Bell, A. T., Fuel, 57, 194, Elsevier (1978).
Vinogradov, S. N. and Linnell, R. H., (eds.), Hydrogen Bonding, Van Nostrand Reinhold Co., New York (1971).
Virk, P. S., Fuel, 58, 149, Elsevier (1979).
Vlieger, J. J., Kieboom. A. P. G., Bekhum, H., Fuel, 63, 334, Elsevier (1984).
Voll, M. and Boehm, H. P., Carbon, 9, 481 ( 1971 ).
Von, W., Doering, E., Rosenthal, J. W., J. Am. Chem. Soc., 89, 4534 (1967).
Vorres, K. S., User's Handbook for the Argonne Premium Coal Sample Program, 11, Argonne National
Laboratory, Argonne, Illinois (1989).
W
Waldner, R., Leisola, M. S. A., Fieter, A., Appl. Microbiol. Biotechnol., 29, 400 (1988).
Walker, P. L., Fuel, 60, 801 ( 1981 ).
Walker, P. L. Jr., Taylor, R. L., Ranish. J. R., Carbon, 29, 411 (1991).
Wallace, S., Bartle, K. D., Perry, D. L. Fuel, 68, 1450 (1989).
Wang, X., Matsumoto, M., Shono, H., Ishihara, A., Kabe, T., J. Jpn. Petrol. Inst., 34, 314 (1991).
Wang, X., Ishihara, A., Kabe, T., Shono, H.,Abstracts ofPaper, 35rd Jpn. Petrol. Inst. Symposium, p. 26 (1991).
Wang, X., Ishihara, A., Satou. T., Shono, T., Kabe, T., J. Jpn. Petrol. Inst., 35, 451 (1992).
Wang, X., Ph.D. Dissertation, Tokyo University of Agriculture and Technology (1993).
Wang, X., Ishihara, A., Shono, H., Kabe, T., Fuel Process. Technol., 38, 45 (1994).
Wang, X., Ishihara, A., Shono, H., Matsumoto, M., Kabe, T., Fuel Process. Technol., 38, 69 (1994).
Wargadalam, V. J., Norinaga, K., lino, M., Fuel, 81, 1403, Elsevier (2002).
Wariishi, H., Valli, K., Gold, M. H., J. Biol. Chem., 267, 23688 (1992).
Wasaka, S., J. Jpn. Petrol. Inst., 78 (10), 802 (1999).
Wasaka, S., Ibaragi, S., Hashimoto, T., Tsukui, Y., Katsuyama, T., Shidong, S., Fuel, 81, 1551, Elsevier (2002).
Wasaka, S., Ibaragi, S., Itonaga, M., Sakawaki, K., Inokuchi, K., Mochizuki, M., Oda, H., Suzuki, T., Energy
Fuels, 17, 172 (2003).
Watanabe, I., Sakanishi, K., Moshida, I., Energy Fuels, 16, 18, (2002).
Watanabe, Y., Yamada, O., Fujita, K., Takegami, Y., Suzuki, T., Fuel, 63, 752, Elsevier (1984).
Waters, B. J., Squires R. G., Laurendeau N. M., Carbon, 24, 217 (1986).
Weber, A., Tesch, S., Thomas, B., Schmiers,H., Appl. Microbiol. Biotechnol., 54, 681 (2000).
Wei, X. Y., Shen, J. L., Takanohashi, T., lino, M., Energy Fuels, 3, 575 (1989).
References 333
X
Xu, W. C. and Tomita, A., Fuel, 66, 632, Elsevier (1987).
Xu, W. C. and Tomita, A., Fuel Process. Technol., 21, 25 (1989).
Y
Yamada, O., Matsumoto, R., Fukuda, K., Honda, E., Tanso (Carbon), 1, 44 (1981).
Yamada, K., Fujii, H., Takayanagi, M., J. Mater. Sci., 21, 4067 (1986).
Yamada, Y., Shiraishi, M., Furuta, T., Yamakawa, T., Sanada, Y., Bull Chem. Soc. Jpn., 57, 3027 (1984).
Yamada, Y. Sakawaki, K., Ida, S., J. Jpn. Chem. Soc., 85 (1987).
Yamamoto, K., Kimura, K., Ueda, K., Horimatsu, T., Nitoh, O., Kabe, T., Proc., Int. Conf. Coal Science.,
Maastricht, Netherlands, 211 (1987).
Yang, X., Garcia, A. R., Larsen, J. W., Silbemagel, B. G., Energy Fuels, 6, 651 (1992).
Yang, X., Larsen, J. W., Silbernagel, B. G., Energy Fuels, 7, 439 (1993).
Yang, X., Silbernagel, B. G., Larsen, J. W., Energy Fuels, 8, 266 (1994).
Yang, Y. and Watkinson, A. P., Fuel, 73, 1786, Elsevier (1994).
Yang, Y., Chen, R., Shiaris, M. P., J. Bacteriol. 2158 (1994).
Yao, J. and Evilia, R. F., J. Am. Chem. Soc., 116, 11229 (1994).
Yarzab, R. F., Fuel, 59, 81, Elsevier (1980).
334 References
Z
Zander, M., Fuel, 66, 1538 (1987).
Zhang, L., Calo, J. M., Lu, W., Prepr, Am. Chem. Soc., Div. Fuel Chem 41, 98. (1996).
Zhang, S.-F., Herod, A. A., Kandiyoti, R., Fuel, 76, 39, Elsevier (1997).
Zhang., T., Jacobs, P. D., Haynes, Jr., H. W., Catal. Today, 19, 353 (1994).
Zoheidi, H. and Miller, D. J., Carbon, 25, 809 (1985).
Index
Advisory Editors:
B. Delmon, Universit~ Catholique de Louvain, Louvain-la-Neuve, Belgium
J.T. Yates, University of Pittsburgh, Pittsburgh, PA, U.S.A.