Coal and Coal-Related Compounds: Structures, Reactivity and Catalytic Reactions

Coal and Coal-Related Compounds Structures, Reactivity and Catalytic Reactions by Atsushi Ishihara, Eika Weihua Qian, I...

Author: Toshiaki Kabe | Atsushi Ishihara | Eika Weihua Qian | I. Putu Sutrisna | Yaeko Kabe

34 downloads 722 Views 21MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Coal and Coal-Related Compounds Structures, Reactivity and Catalytic Reactions by Atsushi Ishihara, Eika Weihua Qian, I. Putu Sutrisna, Yaeko Kabe

• ISBN: 0444517855 • Publisher: Elsevier Science & Technology Books • Pub. Date: October 2004

Preface

In the 20th century, coal played an important role in the development of economics and industry not only in Japan but throughout the world. As raw materials for industrial production such as fuel for the generation of electricity, cokes, tars, gases from coke oven, synthesis gas from coal gasification and liquid fuels from liquefaction, coal has been used extensively 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 society. On the other hand, shortage of petroleum and natural gas are anticipated in the second half of the 21st century. To compensate, the use of coal is expected to gradually increase during the 21st century. In the future, the development of the coal utilization technology will become more and more important to insure the supply of liquid fuels for transportation 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 established 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 scientific 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 technology of coal utilization with and without catalysts. The actual structure of coal, the chemistry included in the reactivity of coal, the methods to elucidate the structure of coal and reaction 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 engineering, 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 aspects of the science and technology for coal processing in the presence and absence of catalysts. The book consists of six chapters: The first chapter describes the classification and characterization methods of coal. The second chapter describes the chemical and macromolecular structure of coals. The third chapter describes the catalytic and noncatalytic pyrolysis of coal, coal tar, and coal tar pitch. The fourth chapter shows the catalytic and noncatalytic 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 preparation of the English manuscript. April 2004 Toshiaki Kabe Atsushi Ishihara Eika Weihua Qian I Putu Sutrisna Yaeko Kabe

Table of Contents

1

Methods of classification and characterization of coal

1

2

Chemical and macromolecular structure of coal

3

Pyrolysis

127

4

Liquefaction of coal

181

5

Gasification of coal

269

6

Microbial depolymerization of coal

303

81

Methods of Classification and Characterization of Coal

1.1 Classification of Coal Coal, a solid flammable rock, is formed through sedimentation of plants that have undergone peatification and subsequent coalification. Plant substances accumulated in prehistoric swamps to form peat deposits which were buried through the movement of the earth's crust, flood, etc. and were carbonized under the pressure and the heat of the earth over a long time. The degree of coalification changes depending on these factors of pressure, heat and time. As a result, various kinds of coal are generated. These coals have specific properties which are often related to the degree of coalification. Therefore coals are classified for their effective use by rank, a measure of the degree of coalification. According to the ASTM (American Society for Testing and Mateirial) ranking of coal in the United States, coal is classified into four classes, lignite, subbituminous, bituminous, and anthracite. In general, as coalification proceeds, coal rank increases in the order lignite, subbituminous, bituminous, and anthracite. Depending on the country, the names of the classes change. The rank of coal is specified by the proximate analysis (moisture, volatile matter, fixed carbon, and ash) and fuel ratio, heating value of a coal (JIS M8812, ASTM Designation D 388-84). The fuel ratio is defined by the following equation: Fuel ratio = Fixed carbon (%)/Volatile matter (%).

(1.1)

The proximate analysis is performed as follows. The content of moisture (%) is estimated by heating one gram of sample with constant moisture (exposed to saturated solution of NaC1) to 105-110 ~ for 60 min. The content of ash (%) is estimated by the combustion of one gram of sample at room temperature to 500 ~ for 60 min, 500-815 ~ for 30-60 min, 815 ___ 10 ~ until a constant value is obtained. The content of volatile matter (%) is estimated by heating the sample in a platinum crucible at 925 ~ for 7 min where the amount of moisture is subtracted from the amount decreased upon heating. The content of fixed carbon (%) is estimated by subtracting the contents of moisture (%), ash (%) and volatile matter (%) from 100%. In general, when the contents of moisture and ash are removed, coal is defined as dry ash free base, d.a.f. Further, when the contents of moisture and mineral matter (content of ash • 1.08) are removed, coal is defined as dry mineral matter free base, d.m.f. The elemental analysis data representative of the rank of coal have also been collected from published data on a moisture-ash-free basis and are presented in Table 1.1 (Hensel, 1981). Carbon content in elemental analysis is often used to classify rank of the coal. The four classes of coal can be explained as follows. Lignite is the lowest rank of coal. It contains the most moisture and volatile matter. Lignite has low heating value, brownish color and is generally referred to as brown coal in Asia, Europe and Australia. Subbituminous coal is black and has intermediate heating value. Bituminous coal is glossy black and has a

2

1 Methods of Classification and Characterization of Coal Table 1.1 Classification Profile Chart

Average analyses-moisture and ash-free Volatile matter (%)

Hydrogen (wt%)

Carbon (wt%)

Oxygen (wt%)

1.8 5.2 9.9

2.0 2.9 3.9

94.4 91.0 91.0

19.1 26.9 38.8 43.6 44.6

4.7 5.2 5.5 5.6 4.4

44.7 42.7 44.2 46.7

Anthracite Meta Anthracite Semi Bituminous Low-vol. Med-vol. High-vol.A High-vol.B High-vol.C Subbituminous Subbit. A Subbit. B Subbit. C Lignite Lignite A

(kJ/kg) a

C H"

C+ H O

2.0 2.3 2.8

34,425 35,000 35,725

46.0 33.6 23.4

50.8 42.4 31.3

89.9 88.4 83.0 80.7 77.7

2.6 4.2 7.3 10.8 13.5

36,260 35,925 34,655 33,330 31,910

19.2 16.9 15.0 14.4 14.2

37.5 25.1 13.8 8.1 6.2

5.3 5.2 5.1

76.0 76.1 73.9

16.4 16.6 19.2

30,680 30,400 29,050

14.3 14.7 14.6

5.0 5.0 4.2

4.9

71.2

21.9

28,305

14.5

3.6

Heating value

To convert kl/kg to Btu/lb, divide by 2.326. [Reproduced with permission from K. Lee Smith et al., The Structure and Reaction Processes of Coal, 10, Plenum Press (1994)]

a

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 typical standard coal samples, are classified using this method of classification. Their proximate 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 microlithotype 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 inTable 1.2 Proximate Analyses for the Argonne Premium Coal Samples a Proximate analysis

Moist

Coal Beulah-Zap (ligA) Wyodak (subC) Illinois #6 (hvCv) Blind Canyon (hvBb) Lewiston-Stockton (hvAb) Pittsburgh (hvAb) Upper Freeport (mvb) Pocahontas #3 (lvb)

Ash

Vol. mat.

F carbon

Cal.val.

rec (%)

rec (%)

dry (%)

rec (%)

dry (%)

rec (%)

dry (%)

rec (kJ/kg)

dry (kJ/kg)

32.24 28.09 7.97 4.63 2.42 1.65 1.13 0.65

6.59 6.31 14.25 4.49 19.36 9.10 13.03 4.74

9.72 8.77 15.48 4.71 19.84 9.25 13.18 4.77

30.45 32.17 36.86 43.72 29.44 37.20 27.14 18.48

44.94 44.73 40.05 45.84 30.17 37.82 27.45 18.60

30.72 33.43 40.92 47.16 48.78 52.05 58.70 76.13

45.34 46.50 44.47 49.45 49.99 52.93 59.37 76.63

17337 19598 25582 30887 26826 31176 30969 34716

25587 27252 27796 32387 27468 31699 31322 34944

Moist -- moisture content, ash -- ash content, vol. mat. = volatile-matter content, F carbon = fixed carbon content, cal. val. -- calorific value, rec = analysis on an as-received basis from the mine, dry = analysis on a moisture-free basis. [Reproduced with permission from Karl S. Vorres, Users Handbook for the Argonne Premium Coal Sample Program, 11 (1989)]

a

1.2 ProximateAnalysis and Elemental Analysis

3

Table 1.3 UltimateAnalyses for the Argonne Premium Coal Samplesa Ultimate analysis Carbon dry (%)

Coal

Beulah-Zap(ligA) 65.85 Wyodak(subC) 68.43 Illinois#6(hvCb) 65.65 Blind Canyon(hvBb) 76.89 Lewiston-Stockton(hvAb) 66.20 Pittsburgh(hvAb) 75.50 Upper Freeport(mvb) 74.23 Pochontas#3(lvb) 86.71

H y d r o g e n Nitrogen

Sulfur

Chlorine

Oxygen

Cal. val.

maf (%)

dry (%)

maf (%)

dry (%)

maf dry (%) (%)

maf dry (%) (%)

maf dry (%) (%)

maf (%)

maf (kJ/kg)

72.94 75.01 77.67 80.69 82.58 83.20 85.50 91.05

4.36 4.88 4.23 5.49 4.21 4.83 4.08 4.23

4.83 5.35 5.00 5.76 5.25 5.32 4.70 4.44

1.04 1.02 1.16 1.50 1.25 1.49 1.35 1.27

1.15 1.12 1.37 1.57 1.56 1.64 1.55 1.33

0.70 0.47 2.38 0.37 0.65 0.89 0.74 0.50

0.04 0.03 0.06 0.03 0.12 0.12 0.00 0.20

20.34 18.02 13.51 11.58 9.83 8.83 7.51 2.47

28340 29871 32887 33988 34267 34930 36076 36695

0.80 0.63 4.83 0.62 0.71 2.19 2.32 0.66

0.04 0.03 0.05 0.03 0.10 0.11 0.00 0.19

18.19 16.24 8.60 10.76 7.69 6.63 4.84 2.17

Dry -- analysis on a moisture-free basis, maf = analysis on a moisture to ash-free basis. [Reproducedwith permission from Karl S. Vorres, Users Handbook for the Argonne Premium Coal Sample Program, 11 (1989)]

a

Table 1.4 ClassificationStandard of Components of Japanese Coal (1958) Composition of coal (microlithotype)

Maceral Groundmass

Vitrite Clarite Durite

Exinite Durite

Main component Vitrinite

Vitrinite

Cutinite

Degradinite

Exinite

Inertinite Durite

Inertinite

Mineral Durite

Mineral matter

Fusinite

Semifusinite Fusinite

[Reproduced with permission from Kimura, H. et al., Chemistry and Industry of Coal, 28, Sankyo Pub. (1977)] 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.

1.2 Proximate Analysis and Elemental Analysis 1.2.1 Collection and Adjustment of Sample Coal has heterogeneity in quality and shape. In collection of samples, therefore, correct and appropriate amounts of coal samples representative of the coal m u s t be collected. Lot indicates the unit of coal for which rank is determined. S a m p l i n g methods, and a d j u s t m e n t methods, and m e a s u r e m e n t of moisture have b e e n established (JIS M8811).

1.2.2 Proximate Analysis Moisture: The m o i s t u r e content is the ratio of d e c r e a s e in w e i g h t that occurs w h e n one gram of a sample w e i g h e d under a constant moisture base is heated at 107 • 2 ~ for 60 min. The moisture content decreases with increasing coal rank. Ash: The ash content is m e a s u r e d by the decrease in weight w h e n one g r a m of a sample is burned c o m p l e t e l y at 815 ~

4

1 Methodsof Classification and Characterization of Coal

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 moisture, ash and volatile contents from 100. (JIS M 8812)

1.2.3 Ultimate Analysis Carbon, Hydrogen: Coal is burned in a combustion tube under an oxygen stream, and carbon and hydrogen are oxidized to carbon dioxide and water, respectively. Sulfur: There are two kinds of sulfur, combustible sulfur and non-combustible sulfur. Total sulfur indicates the sum of both and is measured by the Eschka method or the combustion volumetric technique. Nitrogen: The Kjeldahl method or semi-Kjeldahl method is used for nitrogen determination. Oxygen: Oxygen is determined from the following equation. 0 % = 100--(C% + H% + S% + N%)

(1.2)

(JIS M 8813)

1.2.4 Calorific Value (Heating Value) Calorific value is defined as the amount of calories generated when a unit amount of substance is completely oxidized and is determined using the bomb calorimeter. The calorific value of coal represents gross calorific value (He), which contains the latent heat of water vaporization. When the latent heat of water vaporization is not included, the calorific value is called net calorific value (HN). The relationship between gross calorific value and net calorific value is expressed by the following equation: HN -- HG -- 600 (w + 9h)

(1.3)

where w is the moisture content and h is the hydrogen content under a constant moisture base. (JIS M 8814)

1.3 Solvent Extraction Extraction of coal into an appropriate solvent is an important method for the characterization of coal because this method under mild conditions causes little chemical change to coals and can also be applied to the measurement of IR, NMR, etc. One early study is Wheeler' s method (Wheeler and Burgess, 1911; Jones and Wheeler, 1915, 1916). Fig. 1.1 shows a solvent fractionation diagram. Coal is extracted in order by pyridine, chloroform, petroleum ether, ethyl ether, acetone to form an a compound, a fl compound, and a ~' compound, which is further fractionated into ?q, ?'2, and ?'3 compounds. Fischer (Fischer and Gluud, 1916; Fischer et al., 1924, 1925) and Bone (Bone and Saijant, 1920, Bone et al., 1924, Bone and Tei, 1934) reported other methods. Coal has many aromatic rings, large molecules and a complicated network structure. Therefore very little coal is dissolved by simple solvents such as benzene, alcohol, etc. below 100 ~ (Oele et al., 1951; Marzec et al., 1979). Polar compounds containing nitrogen or oxygen such as amine, phenol, carbonyl, etc. are effective in extracting 20--40% of coal below 200 ~ Quinoline, pyridine, ethylenediamine, ethanolamine, etc. are available for coal extraction. As the Soxhlet extraction method is usually performed at temperatures

1.3 Solvent Extraction

5

Coal I Pyridine extraction

I

I

Extract Ichloroform

Residue

I

/..I Chloroform soluble I Petroleum ether

Chloroform insoluble

I

I

Petroleum ether soluble

Petroleum ether insoluble IEthyl ether

I

I

Ethyl ether insoluble

Ethyl ether soluble

Acetone

I

Acetone insoluble

I

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 bituminous 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 became 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 acetone and pyridine, respectively, according to the procedure shown in Fig. 1.2. The extraction yield was calculated based on Eq.(1.4): Extraction yiels (wt % daf)=

[1-(residue(g)/coal feed (g))] x 100 [1 -(ash (wt % d. b.)/100)]

(1.4)

From the results of the characterization of the raw coals, extracts and residues, it was suggested 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 soluble (AS)

Acetone insoluble (AI) Pyridine

I

I

Pardon soluble (PS)

Pyridine insoluble (PI)

Fig. 1.2 Extraction and fractionation procedures. [Reproduced with permission from Iino, M. et al., Fuel, 68, 1989, Elsevier (1989)]

lower extraction yields than the CS2-NMP mixed solvent. When extracts were further characterized, the quantity of pyridine insoluble-mixed solvent soluble fraction (MS), a heavier fraction than preasphaltene, was larger for extracts with higher extraction yields. MS or MS" (MS" is the lighter portion (about 30%) of the whole MS fraction) had higher values of % oxygen, fa (MS'), molecular weight and spin concentration, but a similar degree of aromatic condensation (MS'), compared with acetone insoluble-pyridine soluble (PS) fractions. Further, the quantity of VM in the residues is similar or slightly less than that in the extracts. Further, in the extraction of Upper Freeport coal, the addition of a small amount of tetracyanoethylene (TCNE) increased the extraction yield from 59 to 85 wt% by this mixed solvent (Liu et al., 1993). Coal can be extracted with decomposition above 200 ~ using solvents such as phenanthrene, fl-naphthol, pitch, etc. When a hydrogen donor solvent such as tetralin is used, hydrocracking and extraction of coal occurs simultaneously.

1.4 Various Analytical Methods In recent years, analytical methods has greatly developed and been applied in studies on the structure and reactivity of coal. These include nuclear magnetic resonance (NMR), FTIR, mass spectrometry, gas chromatography, gel permeation chromatography, and X-ray diffraction. Some of these are introduced here.

1.4.1 Nuclear Magnetic Resonance (NMR) A nucleus precesses in a strong magnetic field. When an electromagnetic wave with the same energy as the frequency of the precessing nucleus is given to the nucleus from the direction vertical to the strong magnetic field, resonance (absorption of energy) occurs. NMR is the spectroscopy which observes this phenomenon. In coal chemistry, ~H and ~3C nuclei

1.4 VariousAnalyticalMethods

7

are investigated extensively. The classical Brown-Ladner method (Brown and Ladner, 1960) can be used to calculate fa (aromaticity), cr (degree of substitution), and p (degree of aromatic condensation) from data of elemental analysis and ~H-NMR according to the following modified equations.

A

~

o-=

Aromatic carbon _ C - H M x - H t J x - H r / 3 Total carbon C Substitution Aromatic site

H,~/x+O HMx+O+Har

Aromatic site H,~/xJr-O--[-Mar P= Aromatic carbon - C - H M x - H t J x - H r / 3

(1.5)

(1.6)

u

(1.7)

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 conditions 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 decoupling 83 kHz, and scan number 1000-2000. The spectrum shown in Fig. 1.3 could be divided 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 aromatic carbons (Car_H), and bridgehead carbons (Can). Tertiary aromatic carbons and bridgehead 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-

ppm

I

I

I

I

I

I

I

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)]

Table 1.5 Assignments of Carbon Functional Groups Chemical shift (ppm) C=O COOR/COOM O-bearing aromatic carbons alkylated aromatic carbons tertiary aromatic carbons and bridgehead carbons O-bearing aliphatic carbons methylene methylene methyl carbons connected with aromatics terminal methyl carbons in longer side chains

Half width (ppm)

187, > 190 178 167, 153 140 126, 113 93, 70, 56 40 31 20 13

12-15 12-15 15-16 16-17 17-18 16-18 16-17 11-13 10-12 10-12

[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-, 2 (2001)] Table 1.6 Carbon Distribution of the Sample Coals Carbon distribution (%) Carbon types C--O, C O 0 Ar-O Ar-C Bridgehead, Ar-H Aliphatic-O CH2 CH3

PC

UF

PT

ST

BC

IL

WY

BZ

0.7 5.0 17.7 62.7 2.8 6.1 5.0

0.4 5.4 18.0 56.1 3.1 10.0 7.0

0.9 7.8 16.4 49.9 4.9 12.6 7.4

0.9 6.7 16.0 52.7 3.2 13.2 7.3

1.2 10.2 13.5 41.5 3.8 19.4 10.3

2.2 10.2 17.7 41.9 4.2 16.3 7.5

4.5 9.5 12.9 42.2 5.2 17.1 8.6

4.1 9.9 14.2 43.4 6.9 14.4 7.2

[Reproduced with permission from Kidena, K. et al., J. Jpn. Inst. Energy, 78, 871 (1999)]


9

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 hydrocarbons. C a r - Car-o -]-- Car_R .71_Car-H -~- CBH

(1.8)

H X Har--C X Car-H

(1.9)

Car-H- (H X Har) / C

(1.10)

X b - CBH / C a r - ( C a r - Car-o -- Car-R -- Car-H) / Car

-- 1 -- (Car-o + Car-R) / Car--(H X Har) / (C X Car)

(1.11)

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 13CNMR 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

15

i

10

i

5

i

i

i

,,0 --5Ppm 15

i

10

i

i

5

0

i

--5ppm

Illinois #6

B

i

1'5

10

i

l

i

i

5

_0 --5Ppm 15

;

6--51ppm 15

i

10

i

i

5

0

i

;

0--;ppm

--5ppm

Wyodak ,r

1'5

l

10

I

l

10

l

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

Pocahontas #3 Upper Freeport Pittsburgh #8 Stockton Blind Canyon Illinois #6 Wyodak Beulah-Zap

0.40 0.36 0.29 0.36 0.24 0.27 0.33 0.37

Bridgehead carbon

Protonatedaromatic carbon

39.5 32.4 27.6 25.8 20.7 20.8 13.8 13.8

23.2 23.7 22.3 26.9 20.8 21.1 28.4 29.6

[Reproduced with permission from Kidena, K. et al., J. Jpn. Inst. Energy, 78, 869 (1999)]

(a)

1o Catenation C4n + 2H2n+4

@

(b)

Circular catenation C6n2H6n

@ Fig. 1.5 Two limiting cases of polyaromatic hydrocarbon condensation: (a) primary catenation: (b) circular catenation. [Reproduced with permission from Solum, M.S. et al., Energy Fuels, 3, 191 (1989)] 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 primary 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 bridgehead carbons Zb can be expressed as Eq. (1.12). ~b -- 1 -- ~p (1.12) Zp = H/C

(1.13)

In the linear catenation, Eq. (1.14) holds. H / C = (2n+4)/(4n+2)

(1.14)

1.4 Various Analytical Methods

11

0.8 0.7

-

0.60.5 Zb 0.4

0.3 0.2

-

0.1

-

0.8

"

0

i

,

10

i

,

i

,

i

20 30 40 Carbons per cluster

,

i

50

,

'

i

60

Fig. 1.6 Plot of the mole fraction of bridgehead carbons, Zb, vs. C where C is the number of carbon atoms per aromatic cluster. The solid curve is for the combined model, the upper dashed curve is for the circular catenation 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)] The linear series index (n) is expressed as in Eq. (1.15). 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). Zb " = 1 / 2 - 3/C

(1.16)

When Zb" is plotted against C, the lower dashed curve is drawn as shown in Fig. 1.6. In the circular catenation, Eq. (1.17) holds. H / C -- 6n/6n 2 = 1/n

(1.17)

The circular series index is expressed as in Eq. (1.18). n=~,/-C / ~/-6

(1.18)

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 tangent 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-'~~b 2 (1.20) Using this equation, the plot of Zb vs. C is expressed as the solid curve in Fig. 1.6.

12


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 calculated 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 determination of quarternary carbon (Kidena et al., 1999).

@

0.5 0.4 ,, 0.3 0.2

(a)

PC UF .Q) PT ~ ' " B2 ~t:4-) ~'~-S L

22

IL C,]ViK ~ ( ~O HV BZC, 'WY

YL /' O TH 0.1 9 - " G S B 0.0

(b)

18 ~

14

UF"e e'" O'PT BC ti LS ILeg HV Wy,,MKe

~

BZ e e , " e T H 10 YL..e.SB

. . . . . 8 65 70 75 80 85 90 95 65 70 75 80 85 9'0 95 Carbon content (wt%, daf)

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 determination 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 function 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 Aromatic hydrogen c Hydrogenb

Carbonyl d Carbon

Coals

Hal

HOH

Har

Htotal

1. Upper Freeport (mvb) 2. Wyodak (subC) 3. Illinois #6 (hvCb) 4. Pittsburgh (hvAb) 5. Pocahontas #3 (lvb) 6. Blind Canyon (hvBb) 7. Lewiston-Stockton (mvb) 8. Beulah-Zap (ligA)

3.43 3.03 3.41 3.60 1.97 4.79 3.48 2.02

0.11 0.33 0.23 0.16 0.06 0.16 0.23 0.34

2.08 1.73 2.07 2.07 2.19 1.90 2.12 1.58

5.62 5.09 5.71 5.83 4.22 6.85 5.83 3.94

Har/Htotal

Oxygen

1 Adj

2 Adj

3 or more

Cal

Units (Abs. • cm-~)

0.66 0.52 0.69 0.67 0.60 0.51 0.67 0.46

0.71 0.78 0.78 0.80 0.73 0.80 0.67 0.74

0.71 0.43 0.60 0.60 0.86 0.58 0.79 0.37

22.87 20.20 22.73 24.00 13.93 31.93 23.20 13.47

0.63 23.86 4.48 0.86 1.92 8.70 3.59 24.67

0.37 0.34 0.36 0.36 0.52 0.28 0.36 0.40

OOH

Oether

1.75 5.25 3.75 2.5 1.0 2.5 3.75 5.5

0.75 5.0 2.25 1.88 1.25 4.0 1.75 5.0

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)] a

Table 1.9 Data on Coal and Chars Produced at 0.5 ~

Heating Rate, 3-min Hold Time (wt % dmmf)

Hydrogen Sample coal 200 ~ 300 ~ 400 ~ 500 ~ 600 ~ 700 ~ 800 ~ a

char char char char char char char

Carbon

Hal

HOH

Har

ntotal

1.93 1.88 1.77 1.67 0.78 0.32 0.11 0.00

0.43 0.38 0.41 0.37 0.22 0.19 0.11 0.08

1.50 1.74 1.36 1.62 2.28 3.36 3.70 3.01

3.86 4.00 3.54 3.66 3.28 3.87 3.92 3.09

Har/Htotal 0.39 0.43 0.38 0.44 0.69 0.87 0.94 0.97

Peak height at 1700 cm-1 (arbitrary units), bND, not determined. [Reproduced with permission from Solomon, P.R. et al., Energy Fuels, 4, 52 (1990)]

Cal 12.89 12.55 11.80 11.10 5.20 2.10 0.73 0.00

Oxygen Carbonyl a 17 17 18 17 6 6 1 0

Oo.

aether

6.83 6.02 6.50 6.00 3.50 3.00 1.75 1.25

6.69 5.63 7.25 > 9.0 6.75 4.38 5.25 7.00

wt loss

Q

0.3 3.0 9.8 22.1 30.0 35.4 43.0

2.38 2.38 1.77 ND b 1.07 1.00 ND b ND b

14


An approximate correlation of oxygen or hydrogen content with coal rank is observed for oxygen functional groups and for aromatic hydrogen. In contrast, the content of aliphatic hydrogen is not necessarily correlated with the coal rank. Blind Canyon coal has a large amount of aliphatic hydrogen, 4.79 wt%, while the younger Beulah-Zap and Wyodak coals have relatively lower amounts of aliphatic hydrogen, as shown in Table 1.8. Illinois # 6 and Stockton coals have relatively larger amounts of hydroxy groups. Pittsburgh and Upper Freeport coals have fairly low amounts of carbonyl groups. Solomon and coworkers also investigated the effect of thermolysis on the hydrogen, carbon and oxygen content in coals (Solomon et al., 1990). One of the results is shown in Table 1.9, where Beulah-Zap coal was heated at 0.5 ~ and held at each temperature for 3 minutes. Significant changes in weight loss and aliphatic hydrogen and OH oxygen contents were not observed until the temperature reached 500 ~ As aromatic hydrogen content increased with increasing temperature, total hydrogen content did not change largely with the change in temperature. In IR measurement, there are several problems associated with the KBr pellet method; for example, a sloping base line appears, water adsorption by KBr occurs, and it is very difficult to subtract one spectrum from the other on a 1/1 basis to obtain a difference spectrum. In recent years, it was shown that the in situ diffuse reflectance IR Fourier transform (DRIFT) technique with neat, undiluted coal samples can be well utilized to trace in-situ reactions such as oxidation, dehydration, etc. (Smyrl and Fuller, 1982). Although there are several problems in quantifying the absorption bands near 3400 cm -~ using the KBr pellet method, DRIFT enables us to investigate the absorption bands ranging from 2400 to 3600 cm-~ which contain hydroxy groups associated with hydrogen bonds. A recent approach to the estimation of hydrogen bonds of coal using DRIFT is described in the following section. B. Estimation of Hydrogen Bond Distribution in Coal Using FTIR Hydrogen bonds, which are noncovalent associative interactions, are included in the coal structure and play a important role in keeping the macromolecular structure of low rank coals (Larsen et al., 1985; Larsen, 1990; Lucht and Peppas, 1987a, b; Suuberg et al., 1990; Aida and Squires, 1985; Miura et al., 1994b, 2001). Therefore, the presence of hydrogen bonds in coal has a critical influence in the coal conversion behavior of the lower rank coals. A large amount of water is held in lower rank coals with hydrogen bonds and the water content changes depending on the extent of drying. Carboxy and hydroxy groups associated with the hydrogen bond in the coal cross-link at an early stage of pyrolysis. These cross-linking reactions affect significantly the subsequent main pyrolysis reactions (Mae et al., 2000). The existence of hydrogen bonds and cross-linking reactions significantly affect the formation of volatiles, the interactions of coal surface and solvents in solvent extraction of coal and/or the liquefaction of coal, or physical properties such as the glass transition temperature. Therefore, it is very important to determine the amount and strength of the hydrogen bonds in coal from both practical and fundamental viewpoints. There are various methods for estimating the hydrogen bonds, e.g., solvent swelling (Larsen and Gurevich, 1996), DSC, NMR and FTIR (Larsen and Baskar, 1987; Painter et al., 1987). Larsen and Baskar related the strength of coal-solvent interactions with the changes in the hydroxy stretching frequency on hydrogen bond formation. Painter et al. attempted to define explicitly the types of hydrogen bonds present in their review and assigned the OH stretching frequency bands to the following hydrogen bonds:

1.4 VariousAnalyticalMethods 3611 cm -z" 3516 cm -1. 3400 cm -1. 3300 cm -1. 3200 cm-l" 2800-31 O0 cm - 1 .

15

free OH groups OH-zr hydrogen bonds self-associated n-mers (n > 3) OH-ether O hydrogen bonds tightly bound cyclic OH tetramers OH-N (acid/base structures)

Existence of the bands was determined by a curve resolving method based on the second 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 parameters 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 estimate the strength distribution of hydrogen bonds probably because of the difficulty in analyzing 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 alcohol-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 relation 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 different hydrogen bonds, were determined by the curve resolving method. First, the absorption band ranging from 2400 to 3750 cm-1 was regarded to consist of the following 10 absorption bands, 7 OH stretch bands and 3 CH stretch bands. 3640 3530 3400 3280 3150

cm-1; cm-1; cm-1; cm-~; cm-~;

free OH groups OH-re hydrogen bonds self-associated n-mers (n ~ 3) OH-ether O hydrogen bonds tightly bound cyclic OH tetramers

16


2940 cm-1; 2640 cm-1; 3050 cm-1; 2993, 2920 cm-~;

OH-N (acid/base structures) COOH dimers aromatic hydrogens aliphatic hydrogens

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 convert 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 intensity 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 model 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 hydrogen 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) obtained 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 heating. 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-fitting 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 OHOH at 3150 cm -1 HBS: OH-N at 2940 cm -~ and HB6" COOH-COOH at 2650 cm -~ 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 ,

9


o'

'

'o

. . . .

30o'C

~"

70oc

\ ~

. . . .

,,',,-,

I

l~176 I

~'

I

,

17

I

i

30 ~

IL

o

< ,

i

i

,

,

i

i

UT

o

~.

PITT

,.Q

< vi

,

i

,

i

,

,~

i

-

_ i

i

i

i

i

i

UF

,

POC

~= 270 ~ 30-90 ~

3600

3200

30 2800

Wavenumbler (cm-1)

2400

4 ,o,- , v--,~7~70_230 3600 ' 3200 2800

.-__ 2400

Wavenumbler (cm-~)

Fig. 1.8 Changein in situ FTIR spectra with heating. [Reproducedwith permission from Miura, K. et al., Energy Fuels, 15, 607 (2001)] 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/molOH, 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 molOHNg 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 desorption 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 decreased. 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 proposed by Miura and co-workers were powerful methods for examining the change in hydrogen bonds in coal.

18


0 i

|

i

|

i

"~"

i

6/11

ND

50

DuB (kJ/mol-OH) 100 150

200

'

,

,

,~

,

,

,

--

30 oc ~

4

0| I

,

I

,

I

t

150 ~

~~

46 f

r~

< 0 ,

!

,

!

!

270 ~

150~

lit a

I

, I

t

I

i

I

6

I

270 ~

~ I 3600

3200

2800

W a v e n u m b e r (cm -~)

2400

o0

I,20 l l , I40

,I

60

80

--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)]

1.4.3

High-resolution Transmission Electron Microscope (HRTEM)

Over the years several studies (Hirsch, 1954; McCartney and Ergun, 1965; Evans et al., 1972; Millward, 1979) have been performed to gain insight into the coal structure using different techniques such as X-ray diffraction (XRD), optical microscopy (OM), and nuclear magnetic resonance (NMR). All these studies converge on one understanding: coal is heterogeneous in nature and made up of very small aromatic layers more or less randomly orientated. The layer size and stacking number increase with the rank of coal. Direct evidence concerning the arrangement of these aromatic layers, obtained by high-resolution transmission electron microscope (HRTEM) lattice imaging, is an attractive complementary technique. Some attempts (Millward, 1979; Millward and Jefferson, 1978; Oberlin, 1979, 1989) have been made but not much literature is available on HRTEM observation of raw coals as such, and most of the studies report on heat-treated coals (Bustin et al., 1995; Sharma et al., 2001). They found the coal samples to be completely amorphous and no evidence of any ordered layers was found. The reason given for the absence of fringes was not 9nly the amorphous nature of coals but also the change in coal structure due to the damage zaused by the intense radiation of the electron beam (Sharma et al., 2001; Ugarte, 1992). In recent years, Tomita and co-workers (Sharma et al., 1999a, b, 2000a, b, 2002) reported an 9bservation technique and the structure of different coals in detail, as observed by HRTEM. These works are presented here.

1.4 VariousAnalytical Methods

19

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 observing 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 damage.

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 magnification image is only of morphological interest. This approach gives more time to observe and image the fringes before irradiation starts damaging the fringe structure. The approach 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 observations, 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 orientated. Some layers can be seen as forming stacks. It is known that coals contain small condensed aromatic units that tend to stack parallel to each other, but the orientation and planarity 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


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 spheroidal carbon particles published by Iijima (1980). Kroto and Mckay (1988) studied the formation 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 agreement 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)]

1.4.4

Characterization of Coal Aggregate Structure by XRD

X-ray diffraction (XRD) has been applied to the characterization of carbonaceous materials including coals for a better understanding of the ordered packing of macromolecules (Cartz et al., 1956; Shiraishi and Kobayashi, 1973; Wertz and Bissell, 1994; Wertz and Quin, 1998; Wertz, 1999). Slow step scan XRD analyses have been reputed to give higher resolution of diffractograms, classifying the carbon-related peak around 20-26 ~ basically into two categories, one is derived from aromatic ring stacking around 26 ~ and the other is the ~'band around 20 ~ The latter band is believed to be derived from aliphatic chains, although the details were not known until recently. The ratio of the two peaks appears to reflect the coal rank (Mochida and Sakanishi, 2000). Slow step scan XRD profiles of various coals are illustrated in Fig. 1.12. The preheating treatment has been reputed to improve the fusibility and coking properties of non-fusible coals (Mochida et al., 1984; Mochida et al., 1983), especially at a rapid heating rate above 100 ~ Sakanishi et al. (2001) investigated five Argonne premium coals before and after heat treatment by step scan XRD profiles in order to clarify changes in secondary aggregated structure of the coals (Sakanishi et al., 2001; Watanabe et al., 2002). The effects of heating rates and temperatures on the aggregated coal structure were also investigated. Fig. 1.13 illustrates the XRD profiles of BZ, WY and IL coals before and after the heat treatment. The original BZ coal exhibited a broad peak around 20 ~ with a broad shoulder at 26 ~ The BZ coal heat-treated at 300 to 370 ~ gave the more intensified peak around 20 ~ compared with that of nontreated coal. Relative intensity at 26 ~ was reduced. The coal heat-treated


I

....

?

21

Scan speed : 0.2~

"-,~u~,-"

\[

.

.

.

.

~

Hongei

ul

m . 1. . _ C (wt%, daf)

^

ontas

~

~ J ~1 %

80.7 76.0 ~ ~' ~~...~

.

~A

~ ..... ~

H..~__,

~~

K-~, Upper Freeport

~~'~'--v,~, ~

Blind Canyon Illinois

~

Wyodak-anderson

~ I

I

I

0.0

I

;:~iUah-zap I

20.0

I

I

40.0

60.0

2 0 ( ~) Fig. 1.12 Slow step scan XRD profiles of variable coals. [Reproduced with permission from ed. 1no; M, Watanabe, 1. 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-, 167 (2001)] BZ

BZ

520 ~ 460 ~ 400 ~ 370 ~ 350 ~ 300 ~ Original

10

20

30 40 2 0 (deg)

50

60

4000

10

20

30 40 2 0 (deg)

50

400 ~ 350 ~ 300 ~ 250 ~ 200 ~ Original 60

4000 WY 100 ~

3000

IL 1O0 ~

~ 3000

i

.,-~

= 2000 1000

10

20

30 40 2 0 (deg)

50

2000 400 ~ 350 ~ ~ 300 oC 1000 250 ~ Original 0 60

10

20

30 40 2 0 (deg)

50

400 ~ 350 ~ 300 ~ 250 ~ Original 60

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)]

at higher temperatures above 400 ~ exhibited intensified diffraction at around 26 ~ with a decrease in diffraction at 20 ~ A similar tendency was observed with WY coal, although the extent of the change in the XRD was smaller compared to BZ coal. The XRD profile of IL coal was not changed so much by heat treatment up to 400 ~ Fig. 1.14 illustrates the XRD profiles of relatively higher rank coals of UF and POC before and after heat treatment. The heating of the coals above 300 ~ made the peak at 26 ~ sharper compared to the nontreated coals probably due to their partial fusibility, although the differences among them were relatively small compared to the lower ranking coals.

22 170 ~

100 ~

3000 '

'

UF

UF

300 ~

2000

1000

ll0

210

310 410 2 0 (deg)

3000

53

,

400 ~ 350 ~ 300 ~ 250 ~ Original 10 60

250 ~

Original 10 310 2 0 (deg)

40

, in

1

"~ 2000

1000

400 ~ C 350~ 300 ~ 250 ~ Original

~

~ 01

i 10

J 20

, , 30 40 2 0 (deg)

~ 50

60

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 ~

. 1400 . / ~ ~ . . .. 1200 ........................ I I ~ 1000 ..... 400 ~ 800 \ ~ , . ~ 3 5 0 ~ = 600 \ ~ ~ 3 0 0 ~ 400 \ /~------~~~ 250~ ~_~ ~ ~oo oc 2O0 . . . . . Kaw 0 10 20 30 40 50 60 2 0 (deg) ] 80 ~

] I26

,400

9

.

.~ 200 fi 100 0

15

........ 400 ~ 370 ~~ x...../~~~ 350 ~ / ~ 300 ~ -- Raw 3'0 4'0 5'0 6'0

\ \

2'0

2 0 (deg)

20 25 2 0 (deg)

I

30

I26

~ 400

I~o

........................p , . ~ 300

\ \

t

~300

I

1400 ~ 1200 1000 800 = 600 / 400 200 0 10

i~o

4

I

~ 200 '~ ~ 100 0

15

20

25

30

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)]


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-treatment 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 2000

~

~

180 ~

.

800

~'~ 1500 . . . . . . . . . . . . . .

~

~= 500

..------~ \\,v-7----~

300 oc

~

250 ~ 200 ~

0

\k,Z~ -x...._...~~ 20

30

2500 ~

.

Ii~

0

,

15

40

2 0 (deg)

[

I26

,

~ 400 ~0 E 200

Raw

10

,

~ 600

400 ~ 350 ~

1000 > " - k - " ~ ~ , \ \ k . _ ~

.,.~

]

20 25 2 0 (deg)

30

80 ~

.

1000

~ 2000 ............ C~, 1500 400~ .~ /--~/...l\\k,.~ 350 ~ ~ 1000 ~ \\'~t...--__ 300 ~ \\ ~ 250 ~ " 500 ~ ~ Raw 0 10 20 30 40 2 0 (deg)

I26

,

~. 800 i20

; 600 "~ ~ 400 ff~ 200

I

0

'

15

20 25 2 0 (deg)

30

Fig. 1.16 Smoothing and decombolution of XRD of heat-treated UF coal. [Reproduced with permission from Watanabe, I. et al., Energy Fuels, 16, 22 (2002)] Heating Rate (ave.K/min)

+ +

180 80 4

3.5 I

UF

3.5

?

3

~

2.5

.g

2

BZ

3

o 2.5 ~

1.5

2

1.5

Orig. 200

250

300

350

Holding temperature (~

400

Orig. 200

250

300

350

400

Holding temperature (~

Fig. 1.17 Effect of heat-treatment of coals on XRD intensity ratio: 126/120. [Reproduced with permission from Watanabe, I. et al., Energy Fuels, 16, 22 (2002)]

24


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-treatment, 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 calculated 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 represent 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 pyrolysis 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 distillation, desorption and thermal degradation. The amount and chemical nature of these products 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 related 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 techniques have the advantages of minimal sample preparation techniques, high sensitivity, specificity, analysis speed and good compatibility with computerized data reduction methods. The patterns in py-MS contain information concerning rank, depositional environment, maceral content and reactivity. Multivariate analysis techniques have correlated pyMS data to explain up to 80% of the total variance described by as many as 25 coal parameters obtained by conventional coal characterization techniques, including ultimate and proximate 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). A. Methods of Pyrolysis Mass Spectrometry In most computerized pyrolysis mass spectrometric methods, different py-MS systems can be distinguished by (1) pyrolysis technique (e.g., filament pyrolysis, direct probe pyrolysis, furnace pyrolysis and laser pyrolysis), (2) ionization method (e.g., electron ionization (EI),


25

chemical ionization (CI) and field ionization (FI)), (3) mass spectrometer type (e.g., integrated, 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-defined Curie-point temperature or to the transition temperature between the ferromagnetic and paramagnetic states. The samples are pyrolyzed directly in front of the electron ionization 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 alloys. Most analytical pyrolysis experiments are performed under vacuum py-MS conditions 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 discrimination effects of higher-mass ions in quadrupole mass spectrometers used in Curie-point pyMS, 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 temperature-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 released 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

IN~ 200

[..

100 0 270

b

200 100

32/0 42/0 5+0 670 Temperature (K)

7+0

8+0

0 270 370 470

570 670 Temperature (K)

770

870

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)]

To obtain reliable chemical information from py-MS, unsupervised multivariate factor analysis methods have proven highly useful since reference spectra are generally unavailable (Windig et al., 1986b). Multivariate analyses have been used to study many biopolymers, including experimental conditions (Windig et al., 1980, 1983, 1984; Windig, 1981). Multivariate factor analysis is used to remove redundant information in the data by replacing correlated mass variables with noncorrelated factors, each consisting of a linear combination of the original variables. The orthogonal factors represent correlated variance in the data set with the first factor explaining the largest amount of variance and so on. The original mass spectral data matrix is effectively decomposed into a factor score and a factor loading matrix. Whereas the former reveals the relationships between different spectra (objects), the loading matrix can be used to devonvelute complex mixture spectra into important components with component chemistry elucidated by mass patterns (Windig and Meuzalaar, 1984). The basis for multivariate analysis can be found in Windig and Meuzelaar (1987). Importance to the multivariate analysis is the graphical representation of the deconvoluted complex mass spectral data using factor analysis and the variance diagram (VARDIA) factor rotation techniques. The techniques are used to identify significant groupings of variables and to extract major component patterns from total ion current (TIC) curves such as total ion intensity as shown in Fig. 1.18 (b). Five factor or components can be numerically extracted from the lignite with variance contributions of 44%, 17%, 13%, 5% and 4% totaling 83% of the variance of the mass spectra series (Simmleit et al., 1992), as show in Fig. 1.19. The mass spectral data of each component are also obtained. This allows us to discuss the structure of coal in detail. Five convoluted and differentiated components found for the lignite are similarly found for other coals. Differences in the components are due to rank and depositional environment, including maceral content. Mass spectrometry shows structural changes according to rank with MS patterns for lignite dominated by a lignin-like dihydroxybenzene structure. High-volatile bituminous coal shows disappearance of oxygen with a prominence of alkyl naphthalenes and phenols, while low-volatile bituminous coal shows mostly alkyl-substituted aromatic products.


Ii

40

II i i i i

35 30 o

25

:

20

27

i

b m./z 424

'

i

i i

C

i i

i

i

m/z 110

i i i

/

/

a

m/z 544

/

[~

.,..~ r.~

.=

d

15

~=

-

./JFli 10

,

mZ 94

/

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)]

B. Results in the Studies by Py-MS Technique Significant results have been reported from py-MS coal structural studies. Meuzelaar et al. (1984a) and Harper et al. (1984) characterized and classified 100 coal samples from the Rocky Mountain Coal Province using Curie-point py-MS. The spectra were dominated by homologous ion series representing dihydroxybenzenes, phenols, naphthalenes, benzenes, alkenes, dienes, alkyl fragments and sulfur-containing compounds with varying degrees of alkyl substitution. The relative amounts of dihydroxbenzenes and naphthalenes correlated closely with the rank, while phenols, aliphatic hydrocarbons and sulfur compounds correlated closely with depositional environment (Meuzelaar et al., 1984a). Correlations were also made comparing the py-MS data obtained with 25 conventional coal characterization parameters. Strong correlations were found with up to 80% of common variance found in both conventional and py-MS data sets. With this large amount of common variance between both data sets, the calculations of parameters such as rank, calorific value and aromaticity can be made from py-MS data (Harper et al., 1984). Maceral concentrates have also been studied with Curie-point py-MS by Meuzelaar et al. (1984a). Vitrinites are characterized by prominent phenolic series, fusinites are dominated by aromatic hydrocarbon signals and sprinites by marked alkene peaks. Correlation of the data by means of multivariate analyses reveals subtle but systematic differences between macerals. Figure 1.20 shows the mass spectra of three coals with increasing rank, Beulah-Zap lignite, Pittsburgh high-volatile bituminous, and Pocahontas # 3 low-volatile bituminous obtained by Curie-point py-MS. The spectra show that pyrolysis products are rank dependent. The prominent pyrolysis products from the lignite are oxygen-containing compounds including phenols. With increasing rank, the abundance of these compounds decreases. For the Pittsburgh high-volatile bituminous coal, the dihydroxybezenes and methoxphenols almost disappeared with increasing alkyl naphthalene concentrations. The most prominent

28

1 Methods of Classification and Characterization of Coal (a) Lignite ""-"

~ Dihydroxybenzenes or methoxyphenols

n

i

0 , ~"

,

,

,

ti,. ,

,

,

,,, ,

,

,

,

Phenols

,

..................

,

,

,

,

,

|

,

(b) High volatile bituminous coal

/I I\

nes

I.I .

,

.

.... .......

Benzenes

5

(c) Low volatile bituminous coal 4 3 2 k

1

0 40

Phenanthrenesor anthracenes .

60

-

~

~

l

l

...............

8'0 ' 100' 120' l h.O l~iO' 180' 200' 289 2a,O m/z

Fig. 1.20 Curie-pointpyrolysis low-voltage EIMS of (a) Beulah-Zap lignite, (b) Pittsburgh and (c) Pocahontas #3 coals. [Reproduced with permission from Huai, H. et al., Prepr. ACS., Div. Fuel Chem, 35, 821 (1990)] 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 possible. The number and nature of pyrolysis products are characteristic of the coal's rank, particularly 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 150

29

TII (103 counts) ,.,,..

120

90

(a)

60

~,

30 o

o 0

0 100 260 300 400 560 660 760 860

2500 m/z 401-900

2000

(b)

1500 o

1000

m/z 2 0 1 - 4 ~ ] t/~

500 50-200__ 0

100 200 300 400 500 600 700 800 Tamperature (~

[.., 1.0 ~,

[.., 1.0

0.8

~,

r~

r~

= 0.6 9~D =- 0.4

o

0.8 0.6

99=- 0.4

>

>

9= 0.2 9

122

9"~ 0.2

100 200 300 400 500 600 700 800 m/z

~,

o

100 200 300 400 500 600 700 800 m/z

Fig. 1.21 Time-resolved py-FIMS analysis of Pittsburgh coal during low-temperature, low-heating-rate devolatilization. (a) Temperature profile of total ion intensity. (b) Temperature profiles of integrated ion intensities representing simulated differental thermogravimetric curves for the molecular ranges indicated. (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 existence 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-temperature 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 (a) Curie-point Py-MS

1"01

Phenols

\

~

Benzenes

~~e~henanthreneSn~

!

.0 1.0

(b) TG/MS

.~.

~ .5 "\

.0 1.0-

._--.-.

60

80

100

L

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 furnace 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 patterns 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


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 systems only record 10-40% of the pyrolysis products as compared to py-FIMS. Thus pyFIMS 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 relatively 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.

1.4.6

Ruthenium Ion-catalyzed Oxidation of Coal

Ruthenium tetroxide, RuO4, has the unique characteristic of preferentially attacking unsaturated carbons in organic compounds (e.g., olefins, alkynes and aromatics). Using this feature, a reagent was applied to the oxidation of some aromatic or olefinic compounds to carboxylic acids in the area of organic synthesis. In the field of fuel science, this oxidation system was first introduced by Stock and Tse (1983) to analyze the aliphatic structure of coal. Typical coal model compounds such as alkylated aromatics, alkanes with two or more aromatic substitutents, partially hydrogenated aromatics and condensed aromatics are oxidized to aliphatic mono- and polycarboxylic acids, and aromatic polycarboxylic acids (Fig. 1.23). Therefore, analysis of the acids produced gives us valuable information concerning the chemical structure of coal, especially its aliphatic portion.

~ ~

R

/R

~

RuO4

RuO4

CO RuO4

R-COOH

HOOC" R "COOH

HOOC ~

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 chromatograph for the analysis of lower carboxylic acids (Murata et a1.,1994), because of its advantage 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 ~ Extraction with 5% NaOH aq

1st run

co2J

24-48 h, N2

2nd run CH2C12, H20

Lower carboxylic acids (C~-C5) Analysis with ion chromatography

I

Filtration residue Elemental analysis

1

Organic phase

Aqueous phase Removal of water

GC, GC-MS Elemental analysis 13C_NMR

Solid esterification with CH2N2 followed by extraction with ether

....

Esterified water soluble fractions GC, GC-MS Elemental analysis ~3C_NMR

L

Residne Elemental analysis

F i g . 1 . 2 4 Product workup for ruthenium ion catalyzed oxidation of 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-, 5 (2001)]

HOOC (CH2)(.-2) COOH

C ( n - i ) H ( 2 n - i) C O O H

10

1

o e,.) .,..~

=o -~9 x

o

~~

----+--- YL . . . . . SB . . . . . WB = GN

J

.1

.1 .01

e _

.01

o o

.001

~

.0001

.001 n~ o o

g

Iv" "~o

90 0 0 0 1

90 0 0 1

90 0 0 0 1

0

~

10

15

20

Carbon number, (n)

25

30

0

'

10

20

Carbon number, (n)

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)]


9

33

= 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

300

I ]'q.De

400

(b)

500 9

600

700

m/z

o CH3(CH2)n COOCH3 n--23-37

o o

| @o

o

it [iL~ ,,li,ldllllil~ ~,

J

I

1i i i

I

I

I

300

400

500

/ ,t

n--37

I

600

,,ta "

"

,,.,,., ,,., ,,..

,,, ,l ,, I

700

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 organic-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 (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 detected being aliphatic di- and tricarboxylic acids with 6-8 carbons and benzene polycar-

34


boxylic acids, as shown in Fig. 1.26. As for the solvent insoluble fraction, it was difficult to obtain distinct structural features, because this fraction was contaminated by a large amount of inorganic salts; its carbon content was only 3%. FT-IR analysis of this portion from RICO of SB coal was conducted. Major absorbances observed were O-H stretching (3500 cm-1), aromatic and aliphatic C-H stretching (3050 and 2920 cm-1), C=C and C=O stretching (1650 and 1600 cm-1), and aromatic C-H out-of-plane bending (around 800 cm-~). The residual fraction from RICO of WB coal also showed a similar IR spectrum. These results indicate that the residue is not unreacted coal but aromatic carboxylic acids with less solubility in organic solvents or water. Nomura and coworkers (Murata et al., 2001) also analyzed both water- and organicsoluble fractions by FD-MS (Field Ionization Mass Spectrometry). FD-MS spectra of these fractions from RICO of SB coal are shown in Fig. 1.27. Fig. 1.27 (a) showed a very complicated profile, but very careful observation taught them that major components could be assigned as benzene polycarboxylic acids and biphenyl polycarboxylic acids with a longer alkyl side chain (Fig. 1.28). m.n-- 1-6

(COOH) n

m (HOOC)

(COOH) n

Fig. 1.28 Benzene polycarboxylic acids and biphenyl polycarboxylic 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-, 7 (2001)]

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 components were methyl esters of aliphatic monocarboxylic acids.

1.5 Tritium Tracer Methods for Coal Characterization Tritium described as 3H or T is a/3-emitter with a half-life of 12.33 years and maximum energy 0.0186 MeV. Among practical/3-emitters, the energy of/3-ray is lowest. A liquid scintillation counter is suitable for accurate determination while a gas flow counter is used for only approximate analysis. Tritium is naturally made by the nuclear reaction in the upper layer of the atmosphere and is miscible with hydrogen and rainwater in the atmosphere. Artificially tritium can be made by the reaction in Eq. (1.25) 6Li +

In ---) 3H -t--4He

(1.25)

Tritium is also generated by nuclear explosion, retreating factory, etc. and is monitored for environmental pollution by radioactivity. Tritium is used for X-ray fluorescence analysis and the radiography of a light alloy and a thin iron plate as a source of bremsstrahlung in the form of 3H/Zr. Tritium-labeled compounds can be extensively used as a tracer to elucidate the mechanism of complicated reactions. When one expects to estimate a reliable reaction rate, the isotope effect should be considered. Although isotopes have the same atomic number and their chemical properties can be equal to each other, differences in physical and chemical features are often recognized because their mass numbers are different. This phenomenon is known as the isotope effect. When elements heavier than atomic number 6, carbon, are used, the isotope effect is very

1.5 TritiumTracer Methods for Coal Characterization

35

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 absolute 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 (disintegration per minute) are used. 1 Ci indicates that 3.7 • 10 l~ of a radioactive nucleus disintegrate per second. 1 Ci is equal to 3.7 • 10 l~ Bq or 3.7 • 10 l~ dps. The number of disintegration 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 follows: AN= 0.693 x -W-Wx NA T M

(1.26)

where W is the mass of the radioactive nucleus, M is atomic weight, and NA is Avogadro's constant. Now we calculate the weight of tritium with 106 dpm in 1 g of water. 106 dpm =

0.693 W x x 6.02 x 1023 12.33x365x24x60 3.02

(1.27)

W = 4.69 • 10-11 g

(1.28)

N = 9.35 • 1012

(1.29)

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 hydrogen, 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 behavior 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 compounds 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 liquefaction 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 determine hydrogen accurately. In this section, recent approaches to elucidate the coal structure, 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

1.5.1


H y d r o g e n E x c h a n g e Reaction of Coal with Tritiated Water

The heteroatom functionality in coal, e.g., hydroxy group, thiol, and amino group, plays a critical role in the processing of coal because they constitute the more polar fraction of the coal and stabilize free radicals (Attar and Hendrickson, 1982; Shinn, 1984a). Consequently, it is very important to know the forms in which they appear in coal and their accurate content present in coal to construct the very complex structure model of the coal and to develop coal conversion techniques. There are in principle two kinds of methods to investigate the chemical structure of coal. One is to attempt breaking down the coal macromolecules into representative fragments and then to deduce the initial structure of the coal from the structure identified from such fragment. The other is the direct non-destructive characterization of coal in its original form in the solid state spectroscopic methods (Haenel, 1992). Generally, the oxygen group is primarily present in the form of hydroxy group and ether group and a little in carbonyl group and carboxylic acid. Fourier transform infrared (FTIR) was available for the measurement of oxygen functional groups (Solomon, et al., 1990; Martin and Chao, 1988). The heteroatom nitrogen is mainly present in the form of pyrrolic and pyridinic nitrogen, and X-ray photoelectron spectroscopy (XPS) has been recently used to quantify their content (Berkowits, 1985; Wallace et al., 1989; Stock et al., 1989; Derbyshire, 1991). The organic sulfur in coal appears mainly in thiophenic heterocycles and aliphatic sulfides and it appears in smaller quantities in thiol, thiophenols, and diaryl sulfides. XPS and X-ray absorption near-edge structure (XANES) spectroscopy were appropriate methods for the determination of the proportions of thiophenic and aliphaticsulfidic sulfur in coal (Gorgaty et al., 1990; Gorgaty et al., 1991; Kelemen et a1.,1991). However, discrepancies between two sets obtained from different researchers need to be resolved before the data can be used with great confidence. Also, the available results are somewhat limited in scope, not being an exhaustive analyses of functional groups found in coal. Hence, many studies are only qualitative or semi-quantitative. Recently, Kabe and coworkers reported that tritium tracer methods are effective to quantify the mobility of hydrogen in coal under coal liquefaction conditions (Kabe et al., 1991a, b; Ishihara et al., 1995). In these works, the reactions of coals were performed with tritiated molecular hydrogen where the hydrogen exchange as well as the hydrogen addition is estimated quantitatively. In particular, it was found that in the reaction of Wandoan coal (subbituminous coal) with tritiated water, water can be regarded as a proton donor rather than a hydrogen atom donor (Ishihara et al., 1993c). Werstiuk and Ju reported that in the protium-deuterium exchange of heteroaromatics with deuterium oxide in the neutral condition and at low temperature, hydrogen exchange between the aromatic hydrogen and water scarcely occurred (Werstiuk and Ju, 1989). Therefore, it can be assumed that the relatively acidic hydrogen such as the hydrogen in the hydroxy group is preferentially exchanged through protium-tritium with tritiated water at lower temperature. This means that the behavior of hydrogen in the functional groups of coal can be determined by the isotope tracer method. Thus, the correlation of behavior of hydrogen at lower temperature and the content of functional group in coal may be determined using the isotope tracer method. Here the results about hydrogen exchange reaction of various Argonne coals with tritiated water are presented. The hydrogen in the functional group of the coals was determined and a comparison with Wandoan coal was also carried out using a glass batch reactor. Further, a method using pulse tritium tracer was also reported to determine the amount of tritium more easily than the conventional method using a batch reactor (Qian et al., 1997). Four kinds of Argonne Premium Coal Samples were obtained in 5 g ampoules ( < 100

1.5 Tritium Tracer Methods for Coal Characterization

37

Table 1.10 Ultimate Analysis of Coals Used (% daf) a Coal

C

H

N

S

O

ND WA IL UF POC

72.94 76.9 77.67 85.50 91.05

4.83 6.7 5.00 4.70 4.44

1.15 1.1 1.37 1.55 1.33

0.70 0.3 2.38 0.74 0.50

20.38 15.0 13.58 7.51 2.68

(L) (SB) (HVB) (MVB) (LVB)

Abbreviations: ND: Beulah-Zap, WA: Wandoan, IL: Illinois #6, UF: Upper Freeport, POC: Pocahontas #3; L: lignite, SB: subbituminous coal, HVB: high-volatile bituminous coal, MVB: medium-volatile bituminous coal, LVB: low-volatile bituminous coal. Except for WA, samples are coals of the Argonne Premium Coal Sample Program. [From Qian, W. et. al., Energy Fuels, 11, 1289 (1997)] a

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 diluted with water to 106 dpm/ml. Phenol, 1-naphthol, toluidine, indole, and phenanthrene (guaranteed reagent) were used without further purification. Commercially available deuterium oxide was used (D ~ 99.99%). All scintillator solvents for the measurement of radioactivity 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 temperatures were 50 ~ and 100 ~ and the reaction times were 1-24 h. After the reaction, the reaction 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 radioactivity. 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 distillation. The radioactivity of the recovered water was measured in a way similar to that described 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 hydrogen 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 bubbling 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 "

GC~ column

Coal 0.5 g

chamber

Gas chromatography

Monophase S trap

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)] Table 1.11 Radioactivities and Weights of Recovered Illinois # 6 Coal (Dried) and Tritiated Water after Reaction at 100 ~ time (h) 0 1 2 4 6 12

Rcoa~(dpm)

Wcoa~(g)

Rwater(dpm)

Ww. . . . (g)

0 13611 21426 24400 23500 26387

0.9223 0.9352 0.9498 0.9047 0.9827

1018500 1004889 997073 994097 994997 992112

1.0410 1.0667 1.0663 1.0472 1.0597

[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 scintillation 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)


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)

Rcoal/Hex = Rwater/Hwater

or

(1.31)

nex----HwaterRcoal/Rwater

In Eq. (1.31). it was assumed that the hydrogen exchange reaction between water and coal reached equilibrium. Thus, after the reaction, the ratio of the radioactivity in coal to the amount of the exchangeable hydrogen in coal (RcoadHex) is equal to the ratio of the radioactivity in water to the amount of hydrogen in water (Rwater/Hwater). The isotope effect was regarded as small and ignored in these calculations. 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 8

100 ~

k

Q

9

~. 6

~z 4 2

0

o As received

0

1

2

3

4

5 6 7 8 Reaction time (h)

9

10

11

12

13

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 diffusion rate of the water into coal. When the hydrogen exchange with tritiated water was carried out using other coals,

40


=

4

2

9

100~

II 50 ~ 0

,

0

I

1

,

I

2

,

t

3

,

~

4

,

~

,

t

,

~

,

i

5 6 7 8 Reaction time (h)

,

i

,

9

I

,

10

I

,

1~2

,

11

13

Fig. 1.31 Effect of temperature on hydrogen exchange. (Illionis # 6 coal, as received). [From Qian, W. et. al., Energy Fuels, 11, 1290 (1997)]

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 ~ Coal HER with tritiated water HER with deuterium oxide

Batch Reactor)

ND

WA

IL

UF

POC

19.2 21.5

7.87 --

7.77 8.90

2.84 2.90

1.60 2.94

[From Qian, W. et. al., Energy Fuels, 11, 1290 (1997)]

In order to investigate the accuracy of the tritium tracer method in the hydrogen exchange 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 water (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 between 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. B. H y d r o g e n Exchange of Model Co m p o u n d s with Tritiated Water Further, a series of heteroatom compounds such as phenol, naphthol, toluidine and indole are regarded as model compounds of the functional groups present in coal, and the hydrogen exchange of these compounds with tritiated water was conducted at 100 ~ to identify the position of exchanged hydrogen. In addition, phenanthrene was also used as a model compound of non-substituted aromatic ring. The results are presented in Table 1.13. Since


41

Table 1.13 Hydrogen Exchange Ratio of Model Compounds with Tritiated Water (%, 100 ~ 6h, Batch Reactor) Compound Phenol Naphthol Toluidine Indole Phenanthrene a

Ratio A a

Ratio B b

17.0 13.1 21.8 15.0 0.13

16.7 12.5 22.2 14.3 0.00

Hydrogen exchange ratio obtained from exchange reaction with tritiated water, oRatio of hydrogen in functional group to total hydrogen in each compound. [From Qian, W. et. al., Energy Fuels, 11, 1290 (1997)]

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 heteroatom 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 basecatalyzed 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) reported that the hydrogen in the aromatic ring of phenol was scarcely exchanged by deuterium 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 ionexchange 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. C. Hydrogen Exchange Reaction in a Pulse Flow Reactor The pulse flow reactor was used to investigate the hydrogen exchange reaction of coal with water at higher temperatures. Fig. 1.32. shows the change in radioactivity of tritiated water in a recovered pulse with the number of pulse introduced when a pulse of tritiated water (8 ml) with a constant radioactivity (9100 dpm/pulse) was introduced into Illinois #6 coal at 200 ~ every 30 min. After the first pulse was introduced, the radioactivity of the recovered pulse was only 580 dpm. This indicates that some tritium in tritiated water was incorporated into coal. Further, the radioactivity in the recovered pulse increased with the number of pulses introduced and approached a constant value (9100 dpm) at the seventh pulse. In contrast to this, the amount of recovered water, which was monitored by a TCD at the outlet of coal-packed column, remained approximately constant for every pulse introduced. This indicates that the decrease in the radioactivity of the introduced pulse cannot be attributed to the adsorption/desorption of water in the coal but attributed to the hydrogen exchange between the tritiated water and the coal. It can be assumed that hydrogen exchange between the tritiated water and the coal reached equilibrium after the introduction of the seventh pulse. According to Eqs. (1.30) and (1.31), the amount of exchanged hydrogen and

42 10000

ooo 6ooo ~

4000

~/,

000

0

I

1

i

2

i

3

i

4

i

5

i

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)] Table 1.14 Hydrogen Exchange Ratio of Coal with Tritiated Water in a Pulse Flow Reactor (%) Coal

Temperature (~

ND WA IL UF POC

100

200

300

16.9 6.4 7.0 2.0 0.8

17.7 7.6 7.9 2.6 1.50

18.7 12.0 11.9 4.9 2.6

[From Qian, W. et. al., Energy Fuels, 11, 1291 (1997) 30 100 ~ 25 20 #. ud

15 10

5

-

0 70

i

75

80

85 Carbon (%)

90

i

i

I

95

i

i

i

i

100

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)]


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 exchange. 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 negligible in the present study. D. Effect of Coal Rank and Temperature on H E R As noted above, the hydrogen exchange reaction of coal with water at low temperature primarily proceeds through the proton exchange between coal and water. Reviewing much of the literature on distribution of oxygen functional groups in coals, Attar and Hendrickson developed (1982) an empirical correlation between the distribution of oxygen functional groups and the ultimate analysis of coals. According to this correlation, the contents of the hydroxy group in each coal were estimated. Further, Solomon et al. (1990, 1991) determined the contents of hydroxy groups in Argonne premium coals using FTIR. The maximum content of hydroxy group can also be calculated from the analytical data listed in Table 1.10 assuming that all oxygen in coals is present in the form of hydroxy groups. In Fig. 1.34 HERs of coals with tritiated water at 100 ~ were compared with the ratios of hydrogen in the hydroxy group to total hydrogen in coal calculated by the several methods described above. It is observed that the HER decreases with increase in rank of coals. The HER for a high-rank coal (Pocahontas # 3) is very close to the HER calculated from the ratio of hydrogen in the hydroxy group whereas the HER for a low-rank coal (Beulah-Zap) is higher than the calculated one. The results shows that, in the high-rank coal, the hydrogen exchanged with water may be hydrogen of the hydroxy group because the content of the other functional groups is very low. On the other hand, coal includes nitrogen and sulfur functional groups such as amino and thiophenyl groups, and the hydrogen in these groups as well as the hydrogen in the hydroxy group can readily be exchanged through ion exchange. This can be verified by the results shown in Table 1.17. It indicates the presence of COOH, SH, NH and other groups in the low-rank coals. In addition, dihydric phenols, aminophenols, and hydroxythiophenols are more abundant in the low-rank coals, especially in lignite, because of its poor coalification. The hydrogen in the aromatic ring of these functional groups may be more mobile and may be exchanged with water under relatively mild conditions. This may also be a cause of the difference in the HER and the hydroxy content for Beulah-Zap coal. All HERs obtained in the batch reactor or the pulse flow reactor are summarized in Fig. 1.35. This figure shows the effect of temperature on the ratio of hydrogen exchange of coal. HERs for all coals hardly changed up to 200 ~ however, they increased slightly at 300 ~ This indicates that other hydrogen in coal rather than the hydrogen in the function-

44


20

10

I

ND

I

i

WA

i

IL

POC

UF

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

-

15

~

II

~

t.

B

10 _

0

i

0

i

50

i

I:1:1

100

t,

~

I

150

i

I

200

i

I

250

i

I

,

300

350

Temperature (~ Fig. 1.35 Effect of temperature on hydrogen exchange ratio. Batch reactor: [] ND A W A (3IL ~ U F + P O C Pulse reactor: 9 9 O I L O U F []POC [From Qian, W. et. al., Energy Fuels, 11, 1292 (1997)]

al groups was exchanged with hydrogen in water. It was proposed that, in the hydrogen exchange 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-naphthol became exchangeable at 300 ~ Therefore, in this study using the pulse flow reactor,


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 ~ 1.5.2

Determination of Aromatic Hydrogen around Functional Groups of C o a l s in R e a c t i o n o f C o a l w i t h T r i t i a t e d W a t e r

In order to develop coal utilization technology, it is important to elucidate the structure of coal. Since the mobility of hydrogen in coal reflects the coal structure, determining the amount of mobile hydrogen provides significant information regarding the structure. In the previous section, it was shown that the tritium tracer method is very effective to estimate the reactivity of hydrogen in coal and coal related-compounds. Further, the amount of hydrogen of functional groups of coal was determined and a pulse flow system as well as a batch system developed. In 1993, Ishihara and Kabe reported that 1-naphthol reacted with deuterium oxide or tritiated water at 300 ~ for 120 min in a batch type reactor to give deuterated or tritiated naphthol where not only hydrogen in a functional group but also hydrogen at ortho and para positions were quantitatively exchanged, (Ishihara et al., 1993). In contrast, at 100 ~ only hydrogen in the functional group was exchanged. When the amount of hydrogen in functional groups at 100 ~ is subtracted from the total amount of hydrogen exchanged at 300 ~ the amount of aromatic hydrogen at the ortho and para positions of the functional group can be determined. Using these reactions, in this section, the determination of the amount of aromatic hydrogen around a functional group in coal is described (Ishihara et al., 2002). The reaction of model compounds other than naphthol with deuterium oxide is also performed to predict the exchangeable positions in coal. OH

OD (T) O Oor O 300 ~

(1.32)

120 min D (T)

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 materials. Coal samples were dried under vacuum at 120 ~ for 1 h. When the reaction was performed 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 furnace 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 liquid scintillation counter. In order to determine the hydrogen in functional groups, the hydrogen 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 13CNMR.

46


A. Determination of Aromatic Hydrogen Around Functional Groups in Coal In order to determine the amount of aromatic hydrogen around functional groups in coal, reactions of coal with tritiated water were performed in a stainless reactor at elevated temperatures. Fig. 1.36 shows the change in the hydrogen exchange ratio (HER) of coal with reaction time at 250 ~ Total HER increased with reaction time and reached constant value at 90 min. At that time, it can be considered that the hydrogen exchange reaction has reached the equilibrium state. As shown in this figure, the HER of functional groups (HER-OH) immediately reached the constant value at the beginning of the reaction. The difference between total HER and HER-OH reveals the amount of aromatic hydrogen exchanged. These results show that the reaction rate is very fast, especially for the functional group and that ionic exchange with protons may occur. All the data are shown in Table 1.15. For all coals, total HER increased with increasing temperature while the HERs of the functional groups remained almost the same with change in temperature, except for ND coal. ND coal initially has 19% exchangeable hydrogen in functional groups, which decreased to about half at 300~ This shows that all carboxyl groups included in ND coal by about 50% of HER of functional group decomposed at the higher temperature under high water pressure. Both total HER and the HER-OH decreased with increasing coal rank. Further, the difference between total HER and HER-OH represents the amount of aromatic hydrogen, which also decreased with increasing coal rank. This result suggests that the hydrogen exchange reaction between aromatic hydrogen and tritiated water is closely related to the presence of functional groups on the aromatic 2O

O O 15

O---

O ........................

5

00

100

I 200 Reaction time (min)

0 " HER-Total

I 300

400

O 9 HER-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

100 ~

250 ~

200 ~

300 ~

Coal

HER-OH (%)

HER-Total (%)

HER-OH (%)

HER-Total (%)

HER-OH (%)

HER-Total (%)

HER-OH (%)

AR-H/OH (mol/mol)

ND IL UF POC

19.2 7.1 2.8 1.6

24.2 10.0 3.3 1.6

13.7 6.1 1.7 0.8

36.5 16.2 4.6 2.2

14.1 6.4 2.1 0.9

45.2 23.4 5.2 4.3

8.5 5.8 2.4 0.9

2.7 2.3 0.9 1.7

[From Ishihara, A. et. al., Energy Fuels, 16, 34 (2002)]


47

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). (1.33)

AR-H/OH -- (HER-Total300 oc -- HER-OH100 oc)/HER-OH100 oc

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 functional group (phenolic OH group) were 2.7, 2.3, 0.9 and 1.7, for ND, IL, UF, and POC, respectively. 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 system 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 reactions 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 T

T

ND

II

I T ~

~

IL (POC)

T IL (POC) III

~

T

UF POC

IV

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 oxide and the reaction almost reached the equilibrium state. No aromatic hydrogen exchanged with deuterium oxide in anisole. Although a trace amount of deuterium seemed to

48

1 Methods of Classification and Characterization of Coal 6.247 OH 3.891 4.086 @ 4 . 1 4 3 4.027 4.109

5.229 NHE 3.962 4.093@ 4.091 4.034 4.034

4.096

4.086 3.785 6.286 O-CH3 ~A,,3.867 4.134 ('(~"] 4.085 4.031 k,,,',,-J~ 4.022 4.094

Fig. 1.38 Electron densities in various organic compounds calculated with WinMOPAC. [From Ishihara, A. et. al., Energy Fuels, 16, 34 (2002)]

OH

pKa (25 ~

O+

OH

O-

O

O+

~2

H§

9.14

H+

9.82

H+

4.65

HN~

Fig. 1.39 Dissociation of naphthol, phenol and aniline. [From Ishihara, A. et. al., Energy Fuels, 16, 34 (2002)] OH

O-

0-O+ O-

.+ 0

0

0 .0- -0 0 -T~o.O T .0" o

o

o

o

0 T~

iT

i

Oi

OT

-0

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


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 values 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 mechanisms (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 important 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 determination 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 hydrogen 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 hydrogen 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 effective 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. A. Procedure of Hydrogen E x c h a n g e Reaction between Coal and Tritiated Molecular Hydrogen Four kinds of Argonne premium coal samples (Beulah-Zap (ND), Illinois #6 (IL), Upper Freeport (UF), and Pocahontas #3 (POC)) were obtained in 5g ampules ( < 100 mesh). The samples of Wandoan coal (WA) were ground to -150-mesh particles and dried for 3 h at 100 ~ under 10 -4 Torr. The analytical data for the coals are shown in Table 1.10. Coal rank increases in the order ND < WA < IL < UF < POC and the oxygen content decreases in the same order. Tritiated gaseous hydrogen was obtained by electrolysis of tritiated water purchased from Japan Isotope Association (185 MBq/ml) and was diluted with water to 106 dpm using a hydrogen generator. Coal was packed into a reactor and dehydrated by nitrogen gas at the desired reaction temperature for 3 h before reaction with gaseous hydrogen. The Pt/A1203 catalyst used in this study was prepared by an incipient wetness impregnation process using aqueous solutions of H2PtC16 and ?'-alumina, followed by drying at 120 ~ for 3 h and calcination at 450 ~ for 20 h. The prepared catalyst was denoted 1 wt%-Pt/A1203, crushed and screened to under 150-mesh particles before use. To estimate the behavior of hydrogen in coal under the milder conditions than liquefaction conditions, a tritium pulse tracer apparatus equipped a thermal conductivity detector as shown in Fig. 1.41 was developed. After about 0.4 g of coal sample and 0.05 g of Pt/A1203 catalyst were mixed in argon gas, the mixture was packed into a reactor (i.d. 4 mm, stainless steel) and fixed with glass wool and quartz sand at both ends of the reactor under a pressure of 15 kg/cm 2. The flow rate of nitrogen as the carrier gas was 5 ml/min. After the

[3H]H2 Qi ~ ~

Recovery

I

--1

"

Scintillator1 ] N cell r-'--

] Recoder ~

I

t

, ~ _

. . . . . . . . .

,

Radioanalyzer Recorder

~ al N2or H2

Reactor

Recovery

Fig. 1.41 Schematicdiagramof experimentalapparatus. [Reproducedwith permissionfrom Kabe,T. et al., Fuel, 79, 312, Elsevier(2000)]


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 approached 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 reactions 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 combustion 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). HER = Hex/Hcoal

(1.34)

Hcoal was calculated using the analytical data in Table 1.10. The amount of hydrogen exchanged between gaseous hydrogen and coal (Hex) was calculated using Eq. (1.35).

Rcoal/Hex =

Rgas]Hgas ;

Hex = Hgas 9Rcoal/Rgas

(1.35)

Hgas is the amount of hydrogen contained in a pulse of introduced gas and Rgas is the radioactivity of tritium contained in a pulse of introduced gas. In Eq. (1.35), it is assumed that the hydrogen exchange reaction between gaseous hydrogen and coal is at equilibrium. Thus after the reaction, the ratio of the radioactivity in coal to the amount of the hydrogen exchanged in coal (Rcoal/Hex) is equal to the ratio of the radioactivity in gaseous hydrogen to the amount of hydrogen in gaseous hydrogen. B. Hydrogen Exchange Reaction between Coal and Tritiated Gaseous H y d r o g e n Figure 1.42 shows the change in radioactivity of tritiated gaseous hydrogen in a recovered pulse with the number of introduced pulse. A pulse of tritiated gaseous hydrogen (9.62 mL) with a constant radioactivity (8792 counts/pulse) was introduced into POC coal with 50 mg of 1%-Pt/A1203 at 250 ~ every 30 min. After the first pulse was introduced, the radioactivity of the recovered pulse was only 1899 counts. This indicates that some tritium in tritiated gaseous hydrogen was incorporated into coal. Further, the radioactivity in the recovered pulse increased with the number of introduced pulses and approached a constant value (8792 counts) at the fifth pulse. In contrast, the amount of recovered gaseous hydrogen, which was monitored by a TCD installed at the outlet of the reactor, remained approximately constant for every introduced pulse. 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 between the tritiated gaseous hydrogen

52

1 Methods of Classification and Characterization of Coal 3000 2500 2000

_

0

9~

1500

.= O

1000 500 0

i 1

2

i 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 temperatures, 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-

0 150

i 200

t 250 Reaction temperature (~

i 300

350

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)]


53

301 25 20 15 10

0

2OO

250 Reaction temperature (~

3OO

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 scintillation counter. It has been already reported that hydrogen in the hydroxy group present in coal exchanges 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 hydrogen is very similar to the HER with water, indicating that most of the former HER corresponds 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 catalyst. C. Effect of Coal Rank on H E R Hydrogen exchange reactions in various coals with tritiated gaseous hydrogen were performed in the absence and presence of a catalyst at 250 ~ The results are shown in Fig. 1.45. In the absence of a catalyst, HER is only 1% even in the case of ND coal and the hydrogen exchange reactions hardly occur in the other coals. In contrast, the hydrogen exchange reactions proceed remarkably in the presence of a catalyst and HERs are lower for higher coal ranks. HERs of ND, WA, IL UF and POC were 45, 16, 16, 5 and 6%, respectively. The result is similar at all temperatures, as shown in Fig. 1.46. Hydrogen exchange reactions between various tritiated coals and water were also performed in a batch reaction system. In Fig. 1.47, HERs between tritiated coal and water (HER removed with water) were compared with the original HERs of tritiated coal and the HERs between coals and tritiated water described above (Qian et al., 1997). In all cases,

54

1 Methods of Classification and Characterization of Coal 50 40 ~" 30 ~Z 20

I I

10

ND

ND No catalyst

WA

IL

II

II

UF

POC

Coal rank

Low

High

Fig. 1.45 Hydrogen exchange ratio of coal with tritiated gaseous hydrogen at 250 ~ permission from Kabe, T. et al., Fuel, 79, 314, Elsevier (2000)]

[Reproduced with

60 ND

WAIL

UF

POC

50 40 30 20100

70

I

I

I

I

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 tritiated 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 exchange 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-


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 liquefaction conditions by assuming that the reactivity toward hydrogen exchange is related to the hydrogen shuttling. They reported that methyl-substituted aromatics such as methylnaphthalene and toluene undergo extensive exchange reactions while nonsubstituted aromatics 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 location 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 hydrogen 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, functional 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 (a) Hydrogen exchange in an alkyl group when toluene is used as a model compound CH3

CH2 ~

CH2 9

CH2T

(b) Hydrogen exchange in a functional group with heteroatom and an aromatic ring with the functional group when phenol is used as a model compound Oo

OH +

~

T9

+

HT

OT

O9 +

~

T2 OH

O9

+

T~ OH

'or G

OH

OH or

OH +

9

T2

OH or

+

T9

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)]

1.5.4

Effects o f Particle Size of C o a l on Catalytic H y d r o g e n E x c h a n g e R e a c t i o n o f Coal with Tritiated G a s e o u s H y d r o g e n and W a t e r

The reactions in coal liquefaction include hydrocracking and hydrogenation by gaseous hydrogen and donor solvents while the reactions in pyrolysis include thermal cracking and dehydrocondensation. In these reactions, the hydrogen transfer, a key reaction, occurs among gas, liquid and solid phases. In order to estimate the mechanism of these reactions, therefore, it is necessary to estimate the hydrogen transfer quantitatively (Ishihara et al., 1995). On the other hand, to obtain useful fuels and chemicals from coal by liquefaction and pyrolysis, it is important to know the coal structure (Solomon et al., 1991; Kelemem et al., 1990). The heteroatom functionalities in coal such as the hydroxy group, thiol, and amino group, etc., especially, play a critical role in the processing of coal because they constitute the more polar fraction of the coal and stabilize free radicals (Attar and Hendrickson, 1982; Shinn, 1984a). Therefore, it is very important to know the forms in which they appear in coal and an accurate determination of their presence in the coal to construct the very complex structure model of the coal and to develop the coal conversion techniques (Solomon et al., 1991; Kelemem et al., 1990). Further, the reactivities of hydrogen in coal reflect the


57

coal structure since each hydrogen in the aromatics, aliphatics, functional groups with heteroatoms, etc. has a different nature. Thus, the determination of the reactivities of hydrogen in coal will provide significant information regarding the coal structure. Useful methods to measure hydrogen in coal utilize isotopes such as deuterium and tritium 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 quantitatively 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 hydrogen 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 hydrogen radicals (Kabe et al., 1999, 2000). In these studies, the direct nondestructive determination 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 particle 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 hydrogen exchange. In the experiment, a pulse of tritiated hydrogen was introduced into a reactor containing coal and the catalyst. Initially, to determine the amount of the Pt/A1203 catalyst, 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 hydrogen, 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


A. Hydrogen Exchange Reactions between Coal and Tritiated Gaseous Hydrogen Three kinds of Argonne premium coal samples (Beulah-Zap (ND), Illinois #6 (IL) and Pocahontas #3 (POC)) were obtained in 5g ampules ( < 100 mesh and < 20 mesh). The particle size distribution of Argonne Premium Coal samples used is shown in Fig. 1.49. The sample of Wandoan coal (WA) was ground to 20-22, 150-250 and < 250 mesh particles and dried for 3 h at 100 ~ under vacuum. The analytical data of coals are shown in Table 1.10. The experimental details and the calculation of hydrogen exchange ratio (HER) appearing in this section have already described in the preceding sections. (a) 5O 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 E

15

25

45

75

106 150 212 300 425 600 850 1000

6( c ) 50 40

9~

400

o 200 ~

r -:a

800 600

0910

"....

9 250 ~

O _

O0

%0 D O ,-D_x3 CCr' CroO

,oo

m% ~

'

200 ,

D

Or. 0

, o

o

-~Y.IP,,P..~

I

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


7

O 200 ~

6

9 250 ~

_= 5 4

2 0

J 1000

I 2000 Time (s)

I 3000

4000

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 reaction 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 reactive 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 ~

IL UF POC

4.2 X 10-2 1.9 X 10-~ 1.9 X 10 -~

3.0 X 10-2 1.3 X 10-I 1.4 X 10-!

300 ~ m 1.5 X 10-~ 8.4 X 10-2

[Reproduced with permission from Ishihara, A. et al., Fuel, 81, 1413, Elsevier (2002)] Table 1.20 Apparent Rate of Hydrogen Exchange Reaction in Coal (g-H/g-coal/min) Coal

200 ~

250 ~

IL UF POC

4.5 X 10-5 1.3 X 10 -4 1.6 X 10 -4

1.1 X 1 0 - 4

1.1 X 10-4 1.5 X 10-4

300 ~ w

2.4 X 10 -4 8.9 X 10-s

[Reproduced with permission from Ishihara, A. et al., Fuel, 81, 1413, Elsevier (2002)]


71

700 i i

'

600 |

,-, 500 =

! i

i

~0

"~ 300

9

o

250 C

o: ~ ' \ : 0

'

-.,,e.

400

J-- 200 ~

,P-~o . . . . . .

r.,~ 9

O

o o

~ )

'

-

0

oo

0

No

O

200 100

0

100

200

300

380

Temperature (~ 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, slowing 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 exchanged for UF coal increased the apparent rate. The above results indicate that less reactive hydrogen clearly participated in the exchange reaction under a reductive atmosphere where H2 is used as a carrier gas. To examine this in detail, the reaction of tritiated coal with gaseous hydrogen (carrier gas) was performed during heat treatment at a heating rate of 3 ~ and the change in the radioactivity 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 temperature 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


periment, as temperature was raised to a maximum of 380 ~ partial hydrogen addition may have occurred. In the preceding sections, the authors have shown that since the generation of a large amount of hydrogen (tritium) atoms and the spillover of those hydrogen atoms to the coal surface can occur on the Pt/A1203 catalyst, those radicals may travel all around the coal structure (Kabe et al., 2000; Ishihara et al., 2000). Further, it was suggested that most of exchangeable hydrogen in coal at least existed on the coal surface where hydrogen atoms or hydrogen molecule could be accessible (Ishihara et al., 2000). In this situation, the hydrogen exchange seems to proceed through direct abstraction of exchangeable hydrogen in functional groups and a-carbon of aromatic tings by radicals. Hydrogen radicals may access exchangeable hydrogen atoms by passing over aromatic ring planes or aromatic hydrogen atoms, which may play a role like a railway carrying the radicals. Although aromatic hydrogen atoms seem to be difficult to exchange at lower temperature (less than 300 ~ (Buchanan et al., 1996; Ishihara et al., 2000), it can be assumed that they begin to react when the temperature increases and the concentration of hydrogen radicals generated on the catalyst becomes very high in the H2 carrier gas atmosphere. 1.5.6

H y d r o g e n Transfer between Coal and Tritiated Organic Solvent

In the preceding sections, hydrogen exchange reactions of coal with tritiated water and tritiated gaseous hydrogen have been described. In these investigations, the amounts of hydrogen of functional groups in coal have been determined using a pulse flow system as well as a batch system. Hydrogen exchange with tritiated water includes hydrogen exchange with proton while hydrogen exchange with tritiated gaseous hydrogen includes hydrogen exchange with a hydrogen atom generated on the catalyst. These tritiated reagents, protons and hydrogen atoms, are inorganic. This section focuses on the elucidation of the reactivity of hydrogen in coals with simple organic compounds because coal is an organic solid mass having a complex structure. The methods used were the hydrogen transfer and exchange reactions of coal with hydrogen donor solvent tritiated tetralin and non-donor solvent tritiated toluene. Tetralin and toluene were tritiated with tritiated water. After the reaction of coal with tritiated tetralin and toluene, the reaction of the tritiated coal and water was carried out to remove the tritium in the functional groups and to obtain the desired information on the position of the hydrogen exchanged in coal. Further, to clarify the position of the hydrogen exchange, deuterated toluene and gaseous deuterium were also used. To avoid significant destruction of the coal structure, the reaction was performed in the range 200-300 ~ below usual coal liquefaction conditions. The results were compared with the reactions of coal with tritiated water and gaseous hydrogen. A. Reaction of Coal with Tritiated Tetralin Four kinds of Argonne premium coal samples (Beulah-Zap (ND), Illinois #6 (IL), Upper Freeport (UF), and Pocahontas #3 (POC)) and Wandoan coal (WA, under 150-mesh parti-' cles) were used. Coal samples were dried for 2 h at 110 ~ under 10- ~Torr. Tritiated tetralin was prepared by modifying a reported method (Yao and Evilia, 1994). Tritiated water (2 g, 108 dpm/g), tetralin (1.2 g), and sodium carbonate (0.005 g) were added into a stainless tube reactor. After argon purge, the reactor was kept for 1 h at 420 ~ under supercritical condition of water. After separation of tritiated water, tritiated tetralin was diluted to about 106 dpm/g. Coal (0.5 g) and tritiated tetralin (0.5 g, 106 dpm/g) were packed into a stainless tube reactor (6 ml). After the reactor was purged with argon, the re-


73

actions were performed under the conditions 200-300 ~ and 5-360 min. After the reaction, 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 location of exchangeable hydrogen in coal, the hydrogen exchange reactions of coal that reacted with tritiated tetralin, i.e., tritiated coal, with water were performed. Calculation of Hydrogen Transfer Ratio: Hydrogen transfer includes both hydrogen addition 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 original coal (Hcoa0. HTR between coal and tetralin was calculated using Eq. (1.41):

(1.41)

HTR -- ntr/Hcoaz

Hcoa~ was calculated with the analytical data presented in Table 1.10. The amount of hydrogen transferred from tetralin into coal (Htr) was calculated using Eq. (1.42)" (1.42)

Rcoal/ntr- Rtet/ntet ; n t r - ntet- Rcoal/etet

ntet is the amount of hydrogen contained in tetralin; and Rtet is the radioactivity of tritium contained in tetralin after the reaction. In Eq. (1.42), the hydrogen transfer reaction between tetralin and coal was assumed to be at equilibrium. Thus after the reaction, the ratio of the radioactivity in coal to the amount of the hydrogen transferred in coal (ecoJHtr) is equal to the ratio of the radioactivity in tetralin to the amount of hydrogen in tetralin after the reaction. Figure 1.62 shows the change in hydrogen transfer ratio (HTR) of coal with reaction time at 300 ~ Total HTR increased with reaction time and reached a constant value at 180 min. At this time, the hydrogen transfer reaction can be considered to have reached the equilibrium state. After the hydrogen transfer reaction, the hydrogen exchange between the tritiated coal sample and water was performed to remove the tritium in the functional group of coal and to learn the extent of hydrogen exchange of the functional group in the reaction of coal with tritiated tetralin. As shown in Fig. 1.62, the HTR corresponding to the hydro20 9 TotalHTR II HTR with-OH [Coal] 15

lO

m

5

,

0

0

60

120

I

I

180 240 300 Reaction time (min)

I

360

420

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


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 reaction 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 reaction 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 re20 9 Total HTR III HTR with-OH [Coal]

15

[.., 10 =:

Q

5

m

00

I

I

60

120

I

I

180 240 Reaction time (min)

I

I

300

360

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

-

420

I

150

I

I

200 250 300 Reaction temperature (~

350

Fig. 1.64 Effectof temperature on HTR of coal with [3H]tetralin for 3h. [From Ishihara, A. et al., Prepr. ACS. Div. Fuel Chem. 44 (3) 661 (1999)]


75

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 hydrogen 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 ~ hydrogen 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 remarkably 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 suggests 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 radTable 1.21 HTRs of Coals with Tritiated Tetralin Coal

ND WA IL UF POC

200 ~

250 ~

300 ~

Total HTR

HTR of OH a

Total HTR

HTR of OH a

Total HTR

HTR of OH a

4.4 2.5 4.5 2.8 2.3

1.7 0.3 1.6 1.4 0.1

4.6 3.3 5.7 3.5 2.9

-1.0 0.9 ---

8.1 9.0 14.4 11.7 4.2

3.8 4.6 6.2 2.0 0.5

Hydrogen transfer ratio of functional groups such as hydroxy group which was determined by the reaction of tritiated coal with water at 100 ~ [From Ishihara, A. et al., Prepr. ACS. Div. Fuel Chem. 44 (3), 660 (1999)]

a

10 [] Beulah Zap 8 -

/

9 Wandoan

~,

A Illinois # 6 "~ o

6 -

o

Ld~~

//

9 Upper Freeport

/ /

J/ w

O Po

._.= 4 [2-

0 150

T

200

250

300

350

Reaction temperature (~ 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_.~ WA

ND

I UF

IL Coal rank

POC

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)] 4 e~0

3

I

i~!~i

>, 2 d:: Q =1

o 1 < E

ND

WA

IL Coal rank

UF

POC

Total hydrogen transfer II

H-released (Tet.- > Napht.)

/

Net H-exchanged

Fig. 1.67 Amount of hydrogen transfer versus 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 hydrogen 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 corresponding 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


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 order 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 naphthalene. 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 functional 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 relatively 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, because 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. B. Reaction of Coal with Tritiated Toluene Tritiated toluene was similarly prepared and the reaction of coal (0.5 g) with tritiated toluene (0.5 g, 106 dpm/g) was performed in a stainless tube reactor (6 ml). Fig. 1.68 shows the variation in HER of coal with reaction time in the range 200-300 ~ Total HERs increased with reaction time and reached constant values at 250 ~ after 90 min and 300 ~ after 180 min, respectively. At this time, the hydrogen exchange reaction can be assumed to have reached the equilibrium state. About 0.8% of HER was observed even at 200 ~ and this value changed very little with reaction time. This may be due to the sorption of the toluene molecule into coal rather than hydrogen exchange. Even at 250 and 300 ~ there may be sorption of toluene into coal corresponding to 0.8% of HER. The HER increased significantly with increasing temperature from 250 ~ to 300 ~ Further, when recovered toluene was analyzed by GC, toluene conversion was less than 1% even at 300 ~ 360 min, and decomposition of toluene hardly occurred. After the hydrogen exchange reaction, hydrogen exchange between the tritiated coal sample and water was performed to remove the tritium in the functional group of coal and discover the extent of hydrogen exchange of the functional group in the reaction of coal with tritiated toluene. As shown in Table 1.22, the HER corresponding to the hydrogen exchange between hydrogen in functional group of coal and toluene was only about 1% even at 300 ~ and 180 min, indicating that the hydrogen in the functional group of coal hardly exchange with hydrogen in toluene at this temperature. These results were very similar to the result from the reaction of coal with tritiated tetralin at lower temperatures 200 and 250 ~ but 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

78 10

I I . 200 ~

~,

~'

250 ~

A " 300 ~

6

4

I

9 m m

m 0

I

I

I

I

90

180

270

360

450

Reaction time (min) Fig. 1.68 HERs of Illinois # 6 coal with tritiated toluene. Table 1.22 HER of IL # 6 Coal with Tritiated Toluene 200 ~

250 ~

300 ~

Reaction time Total HER

HER of OH a

Total HER

HER of OH"

Total HER

HER of OHa

0 0.5 --

2.6 3.4 3.6

1.1

90

0.8

0

1.8

180

1.2

0

1.6

360

1.1

0

1.6

0

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

6 f..rd 4

3.4

ND Low

WA

IL Coal rank

UF

POC 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 tritiated 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 ~ decreasing the hydrogen exchange in the cases of toluene and tetralin. C. Reaction of Coal with Deuterated Organic Reagents To obtain detailed information on the position of the hydrogen exchange, experiments were performed using deuterium. Toluene was reacted with gaseous deuterium, D2, in the presence of IL coal. 0.25 g of toluene and 0.05 g of coal were packed into a batch reactor and pressurized by D2. The reaction was performed at 350 and 400 ~ for 10 h. After the reaction, the amount of hydrogen and the position exchanged in toluene were measured by 1HNMR and 13C-NMR. The hydrogen exchange reaction between toluene and D2 gas did not occur at 350 ~ However, at 400 ~ 28% of the deuterium was introduced into the methyl group of toluene. This shows that approximately one of three hydrogens in the methyl group was exchanged with D2 in the presence of coal at 400 ~ No deuterium was detected in other positions. From this result, a possible route was suggested. D2 was activated by radicals generated in coal at 400 ~ to form the deuterium radical. This deuterium radical reacted with toluene to form a benzyl radical which reacted with the gaseous deuterium molecule to form toluene deuterated at the methyl position. Thus, it was assumed that in the reaction with hydrogen radical, hydrogen in the a carbon of the aromatic ring was activated, and that the hydrogen in the functional group may also be exchangeable under the conditions with hydrogen radical.

2 Chemical and Macromolecular Structure of Coal

2.1 Introduction Coal is an aggregate of heterogeneous substances composed of organic and inorganic materials. 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, ultimately affecting the petrographic composition of the coal. Coals forming in a peat swamp environment 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 autochthonous 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 containing organic material amounting > 50 wt% and 70 vol% as coal. Its genesis alone suggests that coal is a nonhomogeneous material. Coal, being an organic sedimentary rock, is composed of fossilized plant remains, which, in analogy to minerals, are called macerals, and of mineral occlusions. Macerals show distinctive areas under 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 material (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 geochemical phase, which extended over several hundred million years, the peat underwent coalification under the influence of pressure and temperature caused by overlying sediment, forming

82


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

Peat 15 m

Lignite 3 m

Bituminous coal 1.5 m

Antharcite

9 Plant---Peat--~Lignite--~Bituminous coal--~Antharcite-~Graphite Coalification (rank) -Fig. 2.2 Genesis of coal. [Reproduced with permission from Grainger, L. and Gibson, J., Coal UtilizationTechnology, Economics and Policy, 7, Graham, & Trotman Ltd. (1981)]

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 carbon content, amounting to roughly 55% in peat, increases to more than 92 wt% in anthracite, 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 Table 2.1

Plant C (%) H(%) 0(%)

Car/Ctot H/C

~

Peat 55 10 35

--

83

Elementary Composition of Peat and Coals

Lignite 70 8-5 25 --0.5 1

Coalification ~ Bituminous coal 8 0 - 90 6-4 10-5 0.6

--

Anthracite 92 3 2 ----- 0.95 ~0.5

..... ,,-

Graphite

[Reproduced with permission from Haenel, M.W., Fuel, 71, 1212, Elsevier (1992)]

is approximately 1 for lignite and.decreases to less than 0.5 for anthracite. In this chapter, firstly, the advance in chemical structures of coal is summarized. Then, the advance in macromolecular structures of coal is described. Finally, the newest advance in the study on the aggregated structure of coal, i.e., macromolecular structure of coal is introduced.

2.2 Chemical Structure of Coal Many properties can be tested and calculated to characterize coal. Several basic analytic methods of coal, separated into approximate types or groups as follows, are described below. These include (a) petrographic analysis and (b) chemical analyses, including proximate and ultimate analysis, atomic ratios, and elemental analysis. Other methods focusing on the properties of coal such as (c) physical properties, including density, porosity and pore structure and surface area; (d) mechanical properties, including hardness and abrasiveness, elasticity, strength, friability, and grindability; (e) thermal properties, including calorific value, heat capacity, free swelling, plastic and agglomerating properties, and thermal conductivity; (f) electrical properties, including resistivity, electrical conductivity and dielectric constants; (g) ash analyses, including ash elemental and mineralogical analyses and ash fusion temperatures; etc., have been reviewed as well (Smith et al., 1994). The details on the analysis methods about the chemical composition of coal are described in Chapter 1.

2.2.1 Basic Structure Unit and Polymer-like Properties of Coal It is readily known that coal does not consist of unique material of homogeneous texture when one take a broad view of coal. In other words, the material which composed coal was classified in many groups, and the efforts to proceed with the research from the viewpoint of (for example, solvent fractionation) analysis was comparatively made with the reason that coal is considered to be a mixture of the different material in the early research. The result obtained from the fractionation by the solvent extraction is only on a little part of the coal, and information about the solvent-insoluble fraction which occupies the most part of the coal cannot be got. Thus, it is unacceptable to make the chemical constitution of the coal clear with a technique like pure organic chemistry. Therefore, it is necessary to use an average and statistical method is arising for the research of the chemical constitution of the coal. This method would lose its meaning if coal is a mixture of the variously miscellaneous innumerable compound. Fortunately, many experiment facts show that coal is formed from the similar structure. For example, it was reported by Brown (1959) that the infrared spectra of extracts were alike to that of the original coal when the coal was extracted with a series of ketones one after another. Moreover, a hydrocracking using an Adkins catalyst at 350 ~ subsequently hydrogenation at 200 ~ by using the Raney nickel cata-

84

2 Chemicaland Macromolecular Structure of Coal

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 distribution 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 aggregate 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 particular chemical constitution in the viewpoint of organic chemistry, but is a polymer of a certain 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 estimate the amount of aromatic ring in coal as an index, and is determined by various methods, 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 carbon 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 aromatic 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 (%)

Statistical and physical methods X-ray X-ray Density Refractive index Combustion heat Combustion heat IR NMR Magnetic coefficient Chemical methods Hydrocracking Hydrocracking Hydrocracking Oxidative decomposition Oxidative decomposition

80

85

4- 5 >4 15 6 4-

4- 5 >4 17 9 9 peri 3.3 - 4.3 cata 3.8 - 5.6 6.8 3.2 - 4.5 3

-

2

Reference 90

95

730>4 24 26 16 31 10 18 peri 5.3 - 5.8 cata 9.5 6.8 4.3 - 5.5 5 -

Hirsch, 1954 Nelson,1954 Dryden,1958, 1962 van Krevelen and Chermin, 1954 Dryden,1958, 1962 Dryden,1958, 1962 Dryden, 1958, 1962 Dryden,1958, 1962 Dryden,1958, 1962 Hondaand Ouchi, 1957

1-3 1-3

Schuhmacher, et al., 1956 Le Claire, 1941 Sakabe, 1961 Entel, 1954 Montgomery et al., 1956

6 a

2-6 1 - 5b(Ave.3)

Sample 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)]

a

2.2 Chemical Structure of Coal Anthracite

1O0

85

o

90 80

o=/S o

Bituminous

60 50

Fuming acid o

40

~

o Subbituminous coal

30 20 45

o Wood . . . . . . 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)]

2.2.2

Molecular Models of Coal Structure

The structure of coal is inherently complex and varies widely depending on the origin, history, age and rank of the particular coal examined. Nonetheless, because of the relationship between the structure of coal and its reactivity in combustion, pyrolysis and liquefaction processes, there have been many studies to define its molecular (chemical) and conformational (physical) structure and properties. At the early time, it was considered that the coal is of the same structure as to graphite and/or black carbon. Based on this consideration, one model of coal structure was proposed by Fuchs and Sandohoff (1942), as shown in Fig. 2.4. In this model, the main constituent of coal is an enormous aromatic condensed ring, and naphthene ring, alkyl side chain, endocyclic carbon combination and carbonyl substituents etc. surround around in circumference of the ring, which is estimated by the elemental analysis value and properties of product in pyrolysis.

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


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-dihydrophenanthrene type, as shown in Fig. 2.5. The feature of this model is that aromatic nuclear 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 strucHO

9

H O/ 0 ~ / ~ C H ,

H2 H2

,,j(i i ~ C H ~ H H H 2

H2

OH

H2 ./~ ~

CH2/~/,~

_.cn.

H 2 ~ H E ~ N ~ ~ c

2C'CHE~~H

H 3 C / ~

2

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

~///~""'~9~ .?, ~.,-c- c.~ %c.,c ~ / 7 ~~c '

H

c~. H--

~C. /CH~ ]

O

~, II i CH2 "V/~ C' c ~, '.c~S.). CH~..~.... dH

CH 1,C~ /C~ "~" c

/ [ ~

H .C.

\ CH

Small molecules

~

/ H

-o.

k~\\W ~

COHIcgC ~.,

OH CH2 / H C[__ CH_ CH_ .Cr162 I Ha'///Za..a'~/Z~H OH CH2 ~ ~ "~C~" / ~ H H I

/

'

~

H~~C'e/~H

HCW@a

\ ~ ]

H

H~~~H

~

/c I,cH,.H.

H C~2 H2

'

"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 reaH

H

H" ~ "

~1

6

~

6

~"'r H

OH NH2

I

~" Y~

I

H

t

H C H

I

I

H-C -H H H I H-

H

I

H

~- Tt

H

Y,--,H

I

I

~

.~/

-I-

~ "

:~ I

,

II

__"&'O

,~",.~CH3 I

3L /.3

"'o

"

"iN" ~..-.

i

u__ H O

i

\ d/ ~ ~

I

....t.~

H

I

I

r-r

T

T

~

II

II

~

__"_C"__'ta

~

\

/

/

\

H

H

\

/

/

\

H

~

Hz H2

H-(~- H

H2 H-C- H CtS H-(~- H ll3"O'~

~

!

.-c-.

11 , n_l

~'

~

17 II

"~_

17

-on

II "l

,

CH3 H - o . ' H

H-C- H

OH

H - C - CH3 H2H-(~- H H2 H2 ~ " Q , . ~ O*~'~.. IH2

O

H.N ~ H c H ~ H 2 H2

from Wiser, W.H.,

i

.

g4,

"l /3, "H

H

[Reproduced w i t h p e r m i s s i o n

i H-C-H

H

~

..

"d

~ , , .H H ' ' H

Fig. 2.7 Structure model of a h i g h - v o l a t i l e b i t u m i n o u s coal. Prepr. ACS, Div. Fuel Chem., 20, 122 (1975)]

H

r~.,......"~.~tJ ' -""

rr ~

I ," " -~"r-~-r~

C" I

S

H-C-H

H I

.1.

c

_c...

1-12 112 C

]

,_.,,.J~/1,,~/J.~ I ~

T

II

3-

H

~r

cr=o

H ,

" T~V " ( Y TT

H~S'~H

.-c-.

CH3

o ----k.~~,. ~~ I

H H-C-H

6

N

c~

.

H- C-H

H

H

H

I

H\O ~] "C~t-/

CR3

(583 82

rl 2

SH L. iL..1 2. U H2

~

"IN" ",if

H2

Fig. 2.8 Structure model of a P i t t s b u r g h h i g h - v o l a t i l e b i t u m i n o u s coal. [Reproduced w i t h p e r m i s s i o n from Solomon, P.R., New Approaches in Coal Chemistry, ACS Symp. Series No.169, 4, 63 (1981)]

88

2

C h e m i c a l and M a c r o m o l e c u l a r Structure o f Coal

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 chromatography, 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 predicted from this model are close to experimental results. Shinn (1984) proposes a more complex reactive model of coal structure. Shinn's model 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 liquefaction. 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 proHO OH

A~.,,~

OH

HBC

OH

HO HO OH CH3 z,,~_

H3C "-ff

~

,-, ~

~

A

~

x~

H2

OH

OH

HO OH m

.o

ou

. t . Fig. 2.9

HO

.L .t.

Structure m o d e l o f a b i t u m i n o u s coal structure. 1190 Elsevier (1984)]

Fuel, 63,

OH OH [ R e p r o d u c e d with p e r m i s s i o n f r o m Shinn, J.H.,

2.2

Chemical Structure of Coal

89

posed by Huttinger and Michenfelder (1987) is shown in Fig. 2.10. This model was developed from results of elemental analyses, pyrolysis experiments, titration studies, and extrapolation of literature data for similar coals. One, two, and three aromatic tings and longchain 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 encountered in constructing low rank coal models. Polymeric models for low rank coals related to conversion are reviewed by McMillen et al. (1990). 0 H3C ~ 0

0,, OH OH KO "C" H

O

J .

~

O~ 0 ~

?

OH OH ONa 0,,_ OH ~ A ~'~ -C.~ OH OH C HOT('- ~ "]"('--"~"]"""~0 .L .,k . . ~ ~__...~,,,,/ ~ (~""~y g OH 0

I I,, 0.,.~ ...J (

~"

" R - o o " ""

R"

-'Al "

~

. g ~

.

. V

C

~ -

~o9,.L.99..9,~.. " ~ C 2 H ,

"C15H3, O

H

o o" 6 X

o

,.O-C

0

c~

o

n3.0 R.

.O

"o.

"C S'-

:ao0

o 0

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 glassto-rubber transition that coals go through when heated, and the nature of coal-solvent interactions 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 decomposition 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 dimensions. Spiro and Kosky (1982) propose space-filling models for low-, intermediate- and highrank coal molecules shown in Fig. 2.11. The molecules are designed to conform to experimentally 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

(
4.5/k) observed 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 indicates 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). G. Chemical Degradation of M a c r o m o l e c u l e s in Coal For so-called condensation polymers, the coordination number is generally identical to the number of condensable functional groups per monomer, but for macromolecules in coal, it is impossible to determine the number experimentally by nondestructive analyses. The number is a "fictitious" structural parameter, which is introduced in order to describe the degradation characteristics of a particular type of coal. Hence, an effective means to estimate the coordination number is to degrade a coal and measure the extent of degradation (e.g. decrease in the bridge density) and the corresponding increase in the fraction of low molecular mass components and their molecular mass distribution during the course of the reaction. If they are proved to have a quantitative relationship, the coordination number can be estimated. In order to determine such relationships, the following conditions are required for degradation: (a) bond cleavage must take place below temperatures at which pyrolysis cormnences, (b) a considerable proportion of the coal must be solubilized, and (c) the extent of degradation must be measured. In recent years, Hayashi et al. (1997, 1999a, 1999b) proposed an oxidation in weakly alkaline aqueous solution using molecular oxygen and applied it to analytical degradation of some low-rank coals to clarify the mechanism of the oxidation and examine the applicability of general lattice statistics to a quantitative de-

120


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 solution of Na2CO3. By a series of separations including extraction, washing and drying followed 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 solvent-extractables from the oxidized coal increased from 0.15 to 0.96 with progress of oxidation. 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 sampies. 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 lower 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 solubility parameter. Based on these results a reaction mechanism, which oxidation converts aromatic carbons 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 solventextractable 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

9

I

i

I

i

I

i

I

i

I

i

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) THFmethanol (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

O OH Oxidation ..-"-

O HO

O ~..~

O

Aromatic cluster

~

Bridge

121

CO2 (COOH)2 HCOOH CH3COOH etc.

9 Aromatic carbon bonded to bridge

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 macromolecular 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. H. Sorbed Water as M o l e c u l a r Probe for Moist Coal As-received coals hold more or less residual water. In particular, low rank coals such as lignite and brown coals contain a number of oxygen-containing functional groups, resulting in material with hydrophilicity, and this is the primary reason for water content as high as 30-60 wt% on a wet basis. It is a widely recognized fact that coal has gel-like structures that shrink and swell in response to loss and uptake of water, respectively (Evance, 1973; Gorbaty, 1978; Deevi and Suuberg, 1987; Suuberg et al., 1993). Since the solubility parameter, a factor determining the extent of solvent swelling and extraction of coals, of water is very different from that of coal, hydrogen bonds between water and the coal matrix should contribute to the three-dimensional structure or the macromolecular network therein. Thus, the properties of water sorbed in coal will provide information on the intermolecular association and spatial arrangement of water molecules and coal macromolecules interacting via hydrogen bonds. In this section, some studies on the properties of water sorbed in coals are reported. Generally, water sorbed in pores of solid materials of diameter smaller than 10 nm has properties which differs from those of bulk water in its normal thermodynamic state. That is, water sorbed in such pores freezes at temperatures lower than 273 K evolving latent heat smaller than 334 J/g. Fig. 2.36 illustrates the DSC curves for pure water and eight different coals under cooling. The positive peaks appearing in the thermograms are the result of exothermic processes, i.e., congelation as the water sorbed in the coals turns to ice (Mraw

122


c5 Pure water O o

9

i

YL o. tr

LY .t= O X

MW

SB BZ

I

100

WY IL BL

I

I

200 Temperature (K)

300

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 exclusively. No exothermic peaks are detected at temperatures lower than 213 K. Fig. 2.37 illustrates 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 water". 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 congelation as a primary phase transition; hence, this type of water is referred to as "non-freezable water". The DSC analysis also revealed that the desorption of free, bound and non-

2.3 Macromolecular Structure of Coal 300

'

I

'

I

'

I

'

I

'

I

123

'

250 ~)YL~B~

/

~

o 200 150 100 50 0

n

I

d1 l. i i [ I -

i

I

I

I

,

I

,

I

,

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 Water content (g/g-m.f. coal) 50

'

I

'

I

'

I

'

(b)

I

'

I

'

I

,

A BZ [] SB 9 IL

40 "~ 30 | exl)

20 "~ 10 0

, ~ ,

0.0

i

,

_

,

,

0.1 0.2 0.3 0.4 0.5 0.6 Water content (g/g-m.f. coal)

Fig. 2.37 Quantityof heat generated by congelation as a function of water content: (a) YL; (b) BZ, SB, and IL. [Reproduced with permission from Norinaga. K., Energy Fuels, 12, 577 (1998)]

Table 2.6 Contentsof Different Types of Water Determined by DSC and 1H-NMR Method

DSC Free

Coal YL SB BZ WY

0.70 0 0 0

NMR

Bound Non-freezable (g-H20/g-mf coal) 0.35 0.26 0.17 0.15

0.30 0.20 0.31 0.24

Free 0.74 0 0 0

Bound Non-freezable (g-H20/g-mf coal) 0.34 0.24 0.17 0.15

0.27 0.22 0.31 0.24

[Reproduced with permission from Norinaga. K. et al., Energy Fuels, 13, 1058 (1999)]

freezable water occurs successively upon drying at ambient temperature. For the determination of the content of the three different types of water by DSC analysis, drying of coals is inevitable and it may cause shrinkage or collapse of pores, thereby inducing the transition of one type of water into another. Thus, 1H-NMR relaxation measurements were made in order to e x a m i n e the validity of the above classification m e t h o d (Gerstein et al., 1977; Cutmore et al., 1986; Graebert and Michel, 1990; Yang et al., 1992; Rosa et al., 1993; Lynch and Webster, 1979). The congelation of bound water as well as that of free water can be directly observed as the conversion of mobile water proton undergoing slow and exponential magnetization decay into immobile or rigid proton experienc-

124

2 Chemicaland Macromolecular Structure of Coal

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 relaxation 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 content 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-freezable 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 water, it swells but often does not regain its original volume. This irreversible change in volume 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 effect of predrying on the porous structure of water-swollen coal. Using the GibbsThompson 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 10 8

6

YL Boundwater 9 ,._.~ I ~,~

" \'-i-"

Freewater------~

:

:,

,

0

Non-freezable water '

=6 .o

'

I

'

I

'

I'

'

I

'

I

9 BZ

_~-Non-freezabl~Bound--~ water ' water

.

i

dpc

~

~D

e

9~ 4 Q_o

O O

~2

~-, 2 0

,

0.0

o:

o

,

|

I

,

,

,

I

,

,

,

i

0.4 0.8 1.2 w (kg-H20/kg-mfcoal)

,

,

,

O

1.6

0.0

|

0.1

0.2 0.3 0.4 0.5 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)]

0.6

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 cylindrical 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 water. 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 coal

w (g/g of daf coal)

Dpc (nm)

Dps (nm)

YL

1.46 0.33 0.25 0.13 0 0.48 0

4.0 4.0 3.2 3.0 3.2 3.8 3.4

2.4 2.4 2.0 1.9 2.0 2.3 2.1

BZ

[Reproduced with permission from Hayashi, J.-i. et al., Energy Fuels, 15, 908 (2001)]

In recent years, porous structures of solid materials sorbing water have been investigated by 'H-NMR employing Carr-Purcell-Meiboom-Gill (CPMG) pulse sequence. The CPMG method can realize transverse relaxation of water protons with relatively long relaxation 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 estimated 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 Table 2.8 Pore Dimensions for YL (w -- 0.64) and BZ (w -- 0.53) Sample

dpc

Dpc

dps

Dps

2.8 2.2

2.6 2.1

(nm) YL BZ

5.6 4.3

4.4 3.4

[Reproduced with permission from Hayashi, J.-i. et al., Energy Fuels 15, 908 (2001)]

./-

Mobile hydroxyls

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 hydroxylic 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

3.1 Pyrolysis of Coal Coal is a typical conventional solid fuel that has been exploited as an important source of fuel by humankind for thousands of years. It constitutes approximately 75% of the total world resources of fossil fuels. Known international reserves of coal are greater than any other fossil fuel, including oil and gas. A formal report made in 1998 states that at least 1150 billion tons of coal reserves are recoverable, that is, enough to last about 250 years at current consumption levels (Miura, 2000). For comparison, oil and natural gas reserves will last about 43 and 67 years, respectively (Linday, 1993). The main use of coal is as fuel for electric power plants, for which more than 50% of the coal produced in the world is used (Elliot, 1981; Schobert, 1987). Even in industrialized countries such as the United States and Japan, coal is the fuel of choice for electric power generation. Other uses of coal that may be increasingly important in the future are in the production of liquid fuels by direct or indirect liquefaction to replace fuels made from petroleum; production of methanol, a possible substitute for gasoline; and production of synthetic gases. Although the use of coal as energy (fuel) may continue and will be indispensable, "nonfuel" use of coal is another aspect of coal utilization. The main "nonfuel" uses of coal are the production of metallurgical coke and coal tars formed as its byproducts. Coal tars are still an important source of aromatic chemicals. They account for about 15-25% of benzene, toluene and xylene production, and 95% of the larger aromatics (Murakami, 1987). Furthermore, the pitch fraction of the tar is an important raw material of carbon materials such as graphite, carbon fiber, activated carbon fiber, etc. The aromatic coal structure in coals are utilized for these materials, and naphthalene derivatives obtained from coals will be an important raw material for the next generation of polymers such as engineering plastics and new carbon materials (Schobert, 1998). The aromatic carbons from coal are produced from metallurgical coke production byproducts, tars, as stated above. Another method is the coal liquefaction which mainly produces benzene derivatives. Since the tar yield from metallurgical coke production is very small (less than 7% to 8%) and coal liquefaction has not been commercially realized for economical reasons, the two processes will not supply enough aromatic carbons required in the future. It will thus be very important to seek other economical conversion routes to produce aromatic carbons from coal (Schobert, 1998). This section discusses pyrolysis of coal which yields, through destructive distillation, condensable tar, oil and water vapor, and condensable gases. In this process, coal is transformed at elevated temperatures in an inert atmosphere or a vacuum to produce gases, condensable liquids (tar) and char. Closely related to pyrolysis is "hydropyrolysis", the heating of coal in a stream of hydrogen. The processes involve cleavage of chemical bonds in the

128

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, hydrogen 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 (pyrolysis 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 reaction 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 pyrolysis". Fig. 3.1 shows schematically how flash pyrolysis of coal proceeds (Miura, 2000). Extended exposure to .2 high temperature co

~ c

Volatile products

Formation of radicals

.~8"'-. HcH r " ' . 6 n +o,"~" ~a 2

"":.~oe @. 2 . \. o %co

%.L~ c J ~ 97. ~'~ ' t ' ~ .'O, ,.~ / \.Q~ n2ov-s- ~O'T.~ co

",,

.0" [,oJ

/

~

,._o,/2a

, H2S

cna ~'

XX

o~, o;~ t,.Y."I

82

"

/ofra

~o

,ca, s ~

/5 9, ,,,J.~ + ~ o ~ ' "~ o4"6~-c]" ~ L~176 " < ~ ~ ~ ........~ ~ ~ '

~

,'

|

Polymerization and \ condensation reactions

/

,,/ /

~

/Stabilization

,

Coal particle

cH.

@.c.,

\

/

~)

,~ ,

~

~

~

.

~

o

2

H2 CO

,'" 4t~H2

/

~

~

"

Char particle

Fig. 3.1 Reactions and processes which occur upon flash pyrolysis of coal [Reproduced with permission from Miura, K., Fuel Process. Technol., 62, 121, Elsevier (2000)]

3.1 Pyrolysis of Coal 1.5

I

/

I

/

/

,r t

J

I

I

,

129

[

1

1.0

0.5

0

800

1000 Temperature (K)

1200

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 subsequent secondary gas phase reactions. The former reactions are very rapid reactions which consist of radical formation reactions, polymerization-condensation reactions, radical recombination reactions, hydrogen addition reactions, etc., and the latter reactions are decomposition 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 primary reactions must be controlled to further increase TVM. The secondary gas phase reactions 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 socalled flash hydropyrolysis, under a high hydrogen pressure of over 20 bars at high temperatures 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 pretreatment. However, such a significant effect has not always been detected in spite of reexaminations 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 increased 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 recent advances have been reviewed (Gavals, 1982; Fynes et al., 1980). Pyrolysis disproportionates coal in a technically simple manner (although this involves very complex physicochemical 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 socalled 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 mechanisms 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 pyrolysis 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 summarized (Heek and Jungten, 1984). Reidelbach and Summaerfield (1975) developed a general model of pyrolysis in various reactors and expanded the model to cover a wide range of conditions. A model for entrained-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 extractable residue. This led to a model based on simultaneous tar formation, evaporation and polymerization which was tested by correlating reaction conditions with changes in molecular mass. Simons (1983) stressed the importance of fluid transport in mathematical models, 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 consisted 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 optimizing 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 develop 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 concentrates 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 procedure 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.

3.2 Pyrolysis of Coal Tar Use of the coke oven is the largest and simplest process in coal pyrolysis technology, producing blast furnace coke as the primary product. Usually, the coke oven comprises several batches containing about 15 tons of coal for each batch that is charged into single slot ovens grouped together in batteries. The ovens are heated by burning product gas in flues between each oven. Coking takes about 12-20 hours for blast furnace coke. (Gray et al., 1987, 1988; Nishioka and Yoshida, 1983). The coke-making industry is dealt with in terms of coke as the primary product. Coke is fuel and a reductant in the blast furnace. However, another large industry, the coal tar industry, exists to deal with the secondary products or by-products produced during coal pyrolysis. In this process, coal tar is produced in 3-5% yield. Although the yield is rather low, about 1,920,000 tons of coal tar was produced in Japan in 1995, because a large amount of coal was used to produce blast furnace coke. It is very important to utilize coal tar effectively from the viewpoint of coal utilization. The technology for producing coke has been clearly established for many years.

132

3 Pyrolysis

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 material 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 development 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 coking and costly coals but use binder material. These processes are in various stages of development, and it is only a matter of time, politics, and/or economic pressure before coke-making 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.

3.2.1 Typical Yields of Bulk Products from High Temperature Pyrolysis of Coal Coal tar from the coke oven passes out the top end of the ovens through vertical standpipes, where a water spray lowers the temperature. The cooling water and coal tar are transported to a decanter for separation. The coal tar is refined by fractional distillation to obtain commercial products. Typical yields of bulk products from high temperature pyrolysis are shown in Table 3.1. Table 3.1 Yield of Bulk Products from High Temperature Pyrolysis of Coal Product Gas Liquor Light oils Tar Coke

Yield (wt% on dry coal) 17.2 2.5 0.8 4.5 75.5

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, naphthalene 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 % naphthalene, which is removed either by crystallization and hot processing, continuous fraction-


133

ation, or by crystallization and centrifugal washing. Pure naphthalene is used to make phthalic anhydride, an intermediate for plasticizers, polyesters and resins. Although petroleum-based phthalic anhydride obtained via o-xylene has largely replaced this source, naphthalene 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 presentation 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 creosote, and they may be a good source of carbon black oil. The pitch residue is used generally 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 industry. 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 primary 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 molecular 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 primary reaction and a freeboard for the secondary reaction in the gas phase. Since, in fluidized-bed pyrolysis, secondary reactions are known to occur in the dense bed to some extent (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 controlled as well. However, secondary pyrolysis of coal tar has not been investigated sufficiently. 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 influence 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 hydrocarbons 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-hydrogen 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 hydrocarbons in the presence of hydrogen and steam, using a naphthalene, toluene and benzene 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) described general mechanisms for the pyrolysis and carbonization of naphthalene, dimethylnaphthalene, and dibenzothiophene. Fitzer et al. (1971) gave a detailed mechanism for the pyrolysis of acenaphthylene. Studies using these aromatic compounds showed that polymerization 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 molecular 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 involved in coal tar and pitch pyrolysis. However, at present it is impossible to predict the reactivity 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 functionality 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 reaction 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 performance 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 1HNMR 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 functional 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.


135

A. Effects o f T e m p e r a t u r e a n d R e s i d e n c e T i m e o n P r o d u c t s Yields in Tar P y r o l y s i s Figure 3.3 shows the effects of residence time and temperature on the product yields from the secondary pyrolysis of coal tar. Products of the coal tar pyrolysis were separated by solvent fractionation into HS (hexane soluble), HI-CFS (hexane insoluble-chloroform soluble), CFI-THFS (chloroform insoluble-tetrahydrofuran soluble) and THFI (tetrahydrofuran insoluble, or coke). The yield of each product did not change significantly at 700 ~ for 22 s in comparison with the feed tar. However, the yield of each product changed remarkably at 800 and 900 ~ This result shows that the secondary pyrolysis of coal tar in the coal pyrolysis process is important at temperatures over 800 ~ At every temperature, the yields of CFI-THFS, THFI and gaseous products increased, HS decreased and HI-CFS did not change significantly with lapse of time. In particular, there was similarity in the trends of decrease in HS and increase in THFI. The yield of THFI increased remarkably at 900 ~ and reached 38% for 13 s. However, there was no linearity between the yield of THFI and residence time. The rate of the formation of THFI decreased over a yield of THFI of 30%. This indicates that the reactivity or the amount of precursor of THFI decreased with lapse of time. 80

.

80

80

b: 800 ~

60

~- 60

60

40

~ 40

40

20

~

0

0

5 10 15 20 25 Residence time (sec)

20 0

c: 900 ~

>= 20 0

' 5 10 15 20 25 Residence time (sec)

o

0


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 Temperature (~

900

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 pyrolysis 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 elevated 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 ~

B. Aromatic Compounds in Hexane-soluble Fractions Figure 3.5 shows the variation in the amount of one- to five-ring aromatic compounds with or without substituents in the HS fraction on a feed basis. This was calculated by the amount of each component and the yields of HS fraction in the pyrolysis. The feed tar is abundant in two- to four-ring unsubstituted PAHs including naphthalene, phenanthrene, anthracene, pyrene, benzophenanthrene, benzoanthracene, chrysene and so on, and in tworing substituted PAHs which include methylnaphthalene, dimethylnaphthalene, and so on. The amounts of unsubstituted and substituted PAHs did not change significantly at 700 ~ for 22 s in comparison with the feed tar. This shows that the effect of secondary pyrolysis of coal tar in coal pyrolysis processes is important at temperatures over 800 ~ The amount of unsubstituted and substituted PAHs decreased remarkably at 900 ~ Although it has been proposed that the reactivity of model compounds of unsubstituted PAHs rises with increase in the number of aromatic rings, the ratios of reduction of two- to five-ring unsubstituted PAHs were nearly equal. Naphthalene, which is considered to be the most stable compound in unsubstituted PAHs in the HS fraction, also converted to the condensed compound under these conditions. Badger et al. (1996) proposed the pyrolysis mechanism of naphthalene shown in Fig. 3.6. The carbon-hydrogen fission gives naphthyl radicals, which react with naphthalene to yield binaphthyls, and cyclodehydrogenation of the binaphthyls leads to the benzofluoranthenes and perylene. These pyrolysis reactions of naphthalene occur markedly at 900 ~ The amounts of substituted PAHs decreased more rapidly than those of unsubstituted PAHs. The conversions of substituted PAHs remarkably started even at 800 ~ These high reactivities of substituted PAHs are consistent with Jess's report (1996). A possible mechanism for pyrolysis of the HS fraction is shown in Fig. 3.7. The HS fraction is consid12 10

12 10

b: 800 ~

0

g6

~6 ~z 4 T

0

I

I

I

I


2 0

c: 900 ~

g8

~8 6 9~ 4 2

12 I 10

0


~ 4 2 0

I

0


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

1,1'-B inaphthyl

2,2'-B inaphthyl

1,2'-Binaphthyl

1

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)]

H3C

~

.

n

~

~

CH3

CH3

H3C

CH3 9 Coke

H3C

:"~

Fig. 3.7 Mechanism of pyrolysis of the hexane-soluble fraction. [From Godo. M. et al., J. Jpn. Inst. Energy, 77, 128 (1998)] ered to be abundant in alkyl-substituted PAHs such as dimethylnaphthalene, and alkyl-substituted PAHs produce radicals more easily in comparison with unsubstituted PAHs. The thermal instability of the HS fraction which contains alkyl-substituted PAHs results in high reactivity in the secondary pyrolysis of coal tar. THFI fraction is produced by the reaction of HS with HI-THFS in the initial reaction stage. C. Heterocyclic Compounds in Hexane-soluble Fractions A significant portion of the HS fraction in a study (Godo et al., 1998b) was made of heterocyclic compounds. Dibenzofuran was a major compound among the heterocyclic compounds containing oxygen. The HS fraction contained many kinds of compounds having phenolic -OH such as cresol, dimethylphenol, trimethylphenol, naphthol and others. However, the amount of these compounds was not so much. Fig. 3.8 shows the change in the yields of dibenzofuran in feed with the secondary pyrolysis of coal tar. The amount of dibenzofuran did not change greatly at 700 and 800 ~ and decreased rapidly at 900 ~

138

3 Pyrolysis

41

0 700 ~

t

A 800 oc

3~

-

I-1 900 ~

~2 ~

1

i

I

0 0

,

5

I

i

I

i

I

10 15 Residence time (sec)

i

20

25

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

0.0

|

0

I

5

,

10 15 Residence time (sec)

/

20

,

25

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 containing nitrogen. Figs. 3.9 and 3.10 show the change in the yields of quinoline and carbazole, respectively, in feed with the secondary pyrolysis of coal tar. The amount of quinoline did not change in comparison with feed tar at 700 and 800 ~ However, quinoline almost disappeared in 13 s at 900 ~ In contrast to quinoline, the amount of carbazole decreased 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


139

1.0

t

O 700 oc /~ 800 oc l-1 900 ~

0.8

~" 0.6

0.4

0.2

0.0

t

0

5

i

i

10 15 Residence time (wt%)

,

i

20

,

25

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. D. Aliphatic Compounds in Hexane-soluble Fractions Normal alkanes from C~5 to C30 were found in the feed and pyrolyzed coal tar. The total yields of aliphatics from the feed coal tar and the pyrolyzed coal tars at 700, 800 and 900 ~ at a gas flow rate of carrier nitrogen of 50 ml/min (residence time: 13-16 s) were 3.1, 2.1, 1.5 and 0.6 wt% on a feed coal tar basis, respectively. Their distributions are shown in Fig. 3.11. Although the total yields of normal alkanes decreased with rise in reaction tem0.8 ~O-------fl

Feed 700 ~ 800 ~

0.6

s

0.4

0.2

0.0 [ 15

20 25 Carbon number of straight-chain alkanes

30

Fig. 3.11 Distribution of straight-chain alkan.es in secondary pyrolyzed coal tar. [From Godo. M. et al., J. Jpn. Inst. Energy, 77, 129 (1998)]

140

3 Pyrolysis

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 secondary pyrolysis of coal tar do not change significantly. E. Hydrogen Distribution and Structural Parameters of the HI-CFS Fraction Hydrogen distributions of all samples shown in Table 3.2, determined by using ~H-NMR spectra data according to the assignment of chemical shifts shown in Table 3.3, are given in Table 3.4. After the secondary pyrolysis of coal tar within short reaction time range, the amount of aromatic hydrogen (Bar) increased and those of H,~, Ha, H~,, HF and Hoi~ deTable 3.2 Temperature (~

Residence time (sec)

Feed tar

Elemental Analysis of HI-CFS Elemental composition (wt%)

C

H

N

O (diff.)

90.4

5.1

1.8

2.7

H/C 0.677

700 700 700

10.5 16.2 22.1

90.8 90.7 90.9

5.0 5.0 4.9

1.7 1.7 1.7

2.7 2.5 2.5

0.661 0.662 0.647

800 800 800

5.6 9.6 14.7

91.1 91.4 91.3

4.9 4.8 4.9

1.5 1.5 1.5

2.5 2.3 2.3

0.645 0.630 0.644

900 900 900

4.8 8.8 13.3

91.5 91.4 91.5

4.9 5.0 4.8

1.4 1.3 1.4

2.2 2.3 2.3

0.643 0.656 0.630

[From Godo. M. et al., J. Jpn. Inst. Energy, 77, 236 (1998)] Table 3.3 Hydrogen

Chemical Shift of ~H-NMR

Chemical shift (ppm)

Har H,~

9.30-6.30 3.20-1.85

H~

1.85-1.00

Ho, HF

5.80--4.20 4.20-3.20

H~

Assignment Aromatic hydrogen Alpha hydrogen Beta hydrogen Gamma hydrogen Phenolic hydrogen Methylene hydrogen alpha to two aromatic rings

1.00--0.50

[From Godo. M. et al., J. Jpn. Inst. Energy, 77, 237 (1998)] Table 3.4 Temperature (~

Residence time (sec)

Feed tar

Hydrogen Distribution in HI-CFS Hydrogen distribution (wt%) H~

H~

Hr

HF

Ho.

82.8

11.4

2.31

0.68

2.65

0.24

Har

700 700 700

10.5 16.2 22.1

84.2 88.9 88.8

10.3 8.54 8.64

2.26 0.79 0.82

0.75 0.42 0.45

2.13 1.18 1.11

0.38 0.19 0.16

800 800 800

5.6 9.6 14.7

89.0 90.7 89.0

7.87 6.17 7.40

1.18 1.01 1.28

0.67 0.53 0.69

1.06 1.38 1.39

0.19 0.24 0.20

900 900 900

4.8 8.8 13.3

90.5 89.2 86.6

6.29 6.81 8.29

1.25 1.84 2.65

0.60 0.65 1.05

1.28 1.37 1.31

0.10 0.16 0.15

[From Godo. M. et al., J. Jpn. Inst. Energy, 77, 238 (1998)]


C H 3 ~ .

141

CH3

n

~

CH3 ~

~

~

HCH3

CH3

CH3

CH CH3

H

-

S

CH

H

C

H

H

~ H

C

CH3

H

~

CH3

H

Fig. 3.12 Possiblemechanismfor pyrolysis of the HS fraction. [From Godo. M. et al., J. Jpn. Inst. Energy, 77, 238 (1998)] Table 3.5 Brown-Ladner'sStructure Parameters in HI-CFS Temperature Residence (~ time (sec) Feed

fa

cr

p

0.952

0.098

0.653

700 700 700

10.5 16.2 22.1

0.957 0.968 0.968

0.089 0.078 0.078

0.653 0.659 0.643

800 800 800

5.6 9.6 14.7

0.969 0.976 0.971

0.074 0.063 0.069

0.640 0.623 0.635

900 900 900

4.8 8.8 13.3

0.975 0.970 0.963

0.062 0.066 0.076

0.636 0.646 0.612

[From Godo. M. et al., J. Jpn. Inst. Energy, 77, 239 (1998)] creased. This indicates that the aliphatic substituents, the methylene bridge and the hydroxy group decomposed at the initial stage of secondary pyrolysis. However, in long reaction 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 temperatures. 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 containing 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 determined using 1H-NMR spectra data. F. Concentrations of Functional Groups of HI-CFS The concentrations of functional groups in the HI-CFS fraction of raw and pyrolyzed coal tar were determined using the algorithmic method described above and shown in Table 3.6. The HI-CFS fraction of raw coal tar mainly contains molecules which consist of condensed two- to six-ring aromatics substituted by aliphatic chains, and some condensed aromatics were joined by a methylene bridge. After secondary pyrolysis of raw coal tar, the distribution of condensed aromatics, which consisted of molecules in HI-CFS shifted to the large ring number side. At 700 ~ although the changes in the product yields were very small, the peak of the distribution of condensed aromatics shifted from the four-ring to the fivering. At 800 and 900 ~ the distribution of condensed aromatics changed from the threering to the seven-ring. Further, the distributions of condensed aromatics shifted to the large ring number side with lapse of time, especially at 900 ~ From these results, a possible mechanism for the condensation of aromatics in HI-CFS is shown in Fig. 3.13. ,CH3

r

Fig. 3.13 Possible mechanism for the condensation of aromatics in HI-CFS. [From Godo. M. et al., J. Jpn. Inst. Energy, 77, 240 (1998)]

It was determined the components in the HS fraction of raw and pyrolyzed coal tar using 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 secondary 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 obtained from hydrogen distributions and average structural parameters, and provided more effective information in comparison with the average structural parameters.

'"~

0

0

0

E
, o

20-

.,.~

~-

80-

~

60-

(b)

"~N

40-

,,~

=

201

Feed

Blank

800-50

900-50 950-50

Run No. [~: Yield of naphthalene I~: Radioactivity of naphthalene Fig. 3.18 Change in the ratio of recovered amount and the radioactivity of naphthalene. (a): radioactivity of 3H; (b): radioactivity of 14C [Reproduced with permission from Kabe, T. et al., Fuel, 77, 818, Elsevier (1998)]


149

The coal tar used in the study contained unsubstituted, alkyl-substituted and hydrogenated derivatives of PAHs. If dealkylation of alkyl-substituted PAHs and dehydrogenation 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 naphthalene generated. E. Mechanism of Pyrolysis and Hydrogen Exchange of Naphthalene Figure 3.18a shows the changes in the amounts of recovered naphthalene and its tritium radioactivities relative to those in the feed tar. The results in the reaction under the same conditions using 14C-labeled naphthalene described in the previous section are also shown in Fig. 3.18b. As described in the previous section, the decrease in the 14C radioactivity of naphthalene was greater than that in the amount of naphthalene, indicating that naphthalene without 3H was newly formed during the pyrolysis of coal tar parallel to the transformation of 3H-naphthalene to other components. If dealkylation of alkyl-substituted PAHs and dehydrogenation of hydrogenated derivatives of PAHs with two tings occur, the amount of naphthalene increases. Thus, 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 naphthalene formed. If hydrogen exchange of naphthalene does not occur, the decrease in the amount of tritium radioactivity of naphthalene would be equal to that of the 14C radioactivity of naphthalene because the naphthalene formed contains neither tritium nor ~4C. The raT R.

T

T

+

9 T

T T

+ T

T

~ T

RT

(q.l]

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

Scheme. 3.2 [Reproduced with permission from Kabe, T. et al., Fuel, 77, 819, Elsevier (1998)]

(3.4)

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 recovered 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 mobility 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 because 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 generates radicals Ro easily. When an R~ is formed under the operative conditions, a 3H-naphthyl 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 hydrogen transfer reaction depending on the concentration of the naphthyl radicals. A 3H-naphthyl radical reacts with another radical such as an anthryl radical to yield 9-1-naphthylanthracene in Eq. (3.3). The dehydrocyclization of 9-1-naphthylanthracene leads to the formation 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 polymerization 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 functional groups and has wide ranging of molecular weight. Functional groups with heteroatom in coal tar such as hydroxy groups, imino groups, etc. play a critical role in the pyrolysis of coal tar because they constitute the more polar fraction of the coal tar and stabilize free radicals (Attar and Hendrickson, 1982; Shinn, 1984). Consequently, it is very important 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 conversion 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 dissolved 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


151

Table 3.7 Properties of Raw Coal Tar Weight (%) Elemental analysis C H N O (Difference) FTIR spectroscopy analysis Phenolic oxygen Pyrrolic nitrogen

91.2 5.5 1.2 2.1 0.33 0.20

[From Kabe. T. et al., J, Jpn. Petrol. Inst., 41, 165 (1998)] 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 hydrogen mobility in its macromolecular structure are discussed. The examination was made by hydrogen exchange reaction of coal tar or its representative model compounds with tritiumlabeled 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.

A. Hydrogen Exchange of Coal Tar with Tritiated Water Figure 3.19 shows the variation in the hydrogen exchange ratio (HER) of coal tar at the reaction temperatures of 50 and 100 ~ with reaction time. Here H E R means the percent ra-

12t 1.o

lOO ~ C)

--

0.8

~

0.6 0.4 0.2 0.0

0

5

10 15 20 Reaction time (h)

25

30

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 obtained 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 considered 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 temperatures 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

&. m 10

3h

0 0

100

200

300

400

Temperature (~ Fig. 3.20 Effect of temperature on the ratio of hydrogen exchange. [From Kabe. T. et al., J. Jpn. Petrol. Inst., 41, 165 (1998)] Table 3.8 Hydrogen Exchange Ratio of Model Compounds with Tritiated Water (%) Experimental data b Compounds Naphthalene Naphthol Indole

Predicted HER a 0.0 12.5 14.3

100 ~

300 ~

0.04 13.1 15.0

0.39 38.6 49.7

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)] a

3.3 Pyrolysis of Coal Tar Pitch

153

over 100 ~ B. Hydrogen Exchange of Model C o m p o u n d s with Tritiated Water Heteroatom-containing compounds such as naphthol and indole are considered to be representative model compounds of the main functional groups present in coal tar. The hydrogen exchange of these compounds with tritiated water was conducted at 100 ~ and 300 ~ for 6 h to identify the position of the exchanged hydrogen. In addition, naphthalene was also used as a model compound with nonsubstituted aromatic ring. The results are presented in Table 3.8. Since hydrogen in naphthalene barely exchanged with water even at a higher temperature of 300 ~ the aromatic hydrogen in a nonsubstituted aromatic compound is considered not to exchange with hydrogen in water. In contrast, hydrogen in naphthol and indole readily exchanged with water at 100 ~ and the HERs of these compounds are approximately equal to the ratios of hydrogen in functional groups to total hydrogen derived from the stoichiometry of the model compounds. Therefore, the results suggest that the exchangeable hydrogen in coal tar with tritiated water at 100 ~ is hydrogen from the heteroatom of functional groups. On the other hand, the HERs of naphthol and indole at 300 ~ increase remarkably to about three times higher than at 100 ~ indicating that hydrogen other than that in the functional groups can exchange with hydrogen in water. In the hydrogen exchange reaction of coal and coal related compounds with tritiated water, Ishihara et al. (1993c) showed that hydrogen in an aromatic ring attached by a heteroatom such as phenol was exchangeable through electrophilic substitution. The hydrogen exchange of the aromatic hydrogen in indole proceeded with a similar mechanism at 300 ~

3.3 Pyrolysis of Coal Tar Pitch Carbon fibers are black fibers used as yarn, felt, or powder-like short monofilaments of diameter smaller than 10/.tm. They are mainly applied to reinforce polymers. For instance, carbon fiber reinforced thermoplastic and epoxy resin are utilized in many industries. Because the composites reinforced by carbon fibers such as PAN-based carbon fiber are different in density, strength and modulus, they are generally called "advanced composites" or "high performance" composite materials. The progress achieved with carbon fibers, as compared with traditional glass fibers, is based on the superior stiffness of carbon fibers, combined with high strength and low density. Pitch-based carbon fibers (PBCFs) have been recognized as a strategic material because of high Strength per weight, and composite materials offer a major growth market for high performance fibers (Otani et al., 1983). The principle of their preparation has been established (Otani, 1965; Idem, 1983) and commercial production of PBCFs has begun (Otani et al., 1983). However, industrial production of PBCFs of high quality at reasonable cost suffers from several serious problems (Mochida and Korai, 1985; Otani and Oya, 1993). On the other hand, mesophase pitch, recognized as an essential precursor for high performance carbon fiber (Otani, 1965), must be spinnable at low temperature, highly oriented, reactive for oxidation to promote thermosetting, and of high coking value. To satisfy these requirements mesophase pitch should be prepared as described by Korai and co-workers (1983, 1985). Since mesophase transformation occurring in the early stage of carbonization of pitch is a key step in determining the physical and chemical properties of resultant coke and carbon products (Yamada et al., 1984), much attention has so far been focused on the preparation of mesophase pitch (Yamada, 1986; Park et al., 1986) as well as the formation mecha-

154

3 Pyrolysis

nism (Mochida and Korai, 1985). So-called mesophase is a carbonization intermediate which shows optical anisotropy, initially discovered by Borks and Taylor (1965, 1983). A definition of mesophase based on scientific characterization has been given by Mochida and Korai (1985). Mesophase is a liquid crystal state in which molecular groups are regularly oriented so as to show optical anisotropy. The preparation methods of mesophase pitch have been extensively reported, and can be classified into two categories: physical and chemical methods. The chemical methods mainly include polycondensation of raw materials consisting of low molecular weight molecules, hydrotreatment, alkylation, dealkylation and deheteroatom, etc. Among these, hydrotreatment is particularly effective for the preparation of a mesophase pitch from coal tar pitch derived from a coke oven as a by-product since it has undergone severe thermal hysteresis. By hydrotreatment, the fluidity and melting point of pitch can be lowered and thereby the optical texture of the resulting carbon products can be improved. So far, hydrogenation using molybdenum-based supported catalysts, hydrogen donor solvent, and BenKeser reaction has been investigated by many researchers (Mochida et al., 1974, 1987; Yamada et al., 1981). It has been shown that hydrogenation can saturate the polynuclear aromatic tings in pitch to improve the carbonization reactivity of pitch. Therefore, developing a more effective and economical method for the hydrotreatment of pitch is necessary. Hydrogen exchange is the other form of hydrogen transfer and has also been extensively studied to better understand the mechanisms of various processes (Benjamin et al., 1982a, 1983; Davis and Garnett, 1975; Garnett and Kenyon, 1971). It has been recognized by these studies that the exchangeability of a compound is strongly related to its molecular structure, and further to its capacity as hydrogen donor or acceptor. From this knowledge, it is possible to estimate the structural feature of pitches by investigating the hydrogen exchange. In investigations on the mechanism of hydrogen transfer, the isotope tracer method has been utilized extensively as an effective means. For example, in coal liquefaction, this method makes it possible to determine which structural positions in coal react with hydrogen during liquefaction by labeling reactive sites with deuterium or tritium. Most of these investigations were performed using deuterium tracer. Since 1976, several research groups have used deuterium to investigate the mechanism of coal hydrogenation (Schweighardt et al., 1976; Gaines and Ytirtim, 1976; Kerskaw and Barrass, 1977; Franz, 1979) and reaction of coal-related model compounds (Cronauer et al., 1980; Benjamin et al., 1978). In these researches, intensive effort has been made to obtain a better understanding of coal hydroliquefaction mechanisms. Such knowledge may lead to the improvement of hydrogen 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 to solvents and the lack of quantitative data from 2H NMR, it is difficult in these studies to conduct the quantitative analysis of hydrogen transfer among the gas phase, solvent and coal. The investigation of coal liquefaction mechanisms using radioactive tritium as tracer started in recent years. The representative investigations were 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 et al. (1983, 1986a, 1986b, 1991). The results have shown that the tritium tracer method has several distinct advantages over the deuterium tracer methods, especially for a complex reaction system. Carbonization is the other important process for the production of carbon material from pitches. One of the main objectives to develop an understanding of the carbonization phenomena in this field is to achieve the manufacture of satisfactory carbon products. It was


155

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 carbonization 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 carbonization of pitch has not yet been achieved. In this section, the hydrogen mobility or reactivities of various pitches are first estimated to clarify the relationship among hydrogen mobility of pitch, molecular structural features, 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 performance 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 order to introduce hydrogen into pitch (Mochida et al., 1981, 1982, 1987). However, the reactivity 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 hydrogen with pitches. Attention was focused mainly on the interdependence of hydrogen mobility 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 tritium 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 temperature for all three pitches. The tritium was transferred from gas phase to pitch via hydrogen 100

g 9= 80 0

N 60

O 300 ~ [] 350 ~ A 400 ~

40

,

,

I

0

,

I

,

I

120

60

,

180

I

,

240

I

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

0.3

300 ~

0.2 0.1

0.0~ .~ 0.4 d=

[] Addition

(b)

9 Exchange

~ o.3

350 ~

0

r~ ~ 0.2 ~

O.1

~_____~---r---7-~- - - - - - - - - - ~ ,

0.0 .~ 0.4

I

,

I

~

A Addition 9 Exchange

= 0.3

I

(c) 400 ~

0.2 0.1 0.0

,

0

,

60

i

120

i

180

|

240

300

Nominal reaction time (min) Fig. 3.22 Amount of hydrogen transferred from gas phase to coal tar pitch (HP-A pitch). (Wang, X. 1993)


157

Table 3.9 ActivationEnergies of Hydrogen Addition and Exchange in Hydrogenationof Coal Tar Pitches Pitch

Hydrogen addition kcal/mol

Hydrogen exchange kcal/mol

HP-A GP-A GP-B

9.3 9.8 9.0

15.4 2.9 3.7

[Reproduced with permission from Wang, X. et al., Fuel Process. Technol. 38, 75, Elsevier (1994)] Table 3.10 Rate Constants of Hydrogen Addition and Exchange in Hydrogenation of Coal Tar Pitches Pitch

Reaction temperature (~

Hydrogen addition (min-l)

Hydrogen exchange (min- ~)

HP-A

300 ~ 350 ~ 400 ~

1.5 X 10-4 3.1 X 10-4 5.0 X 10-4

3.2 X 10-s 8.3 X 10-5 2.4 X 10-4

GP-A

300 ~ 350 ~ 400 ~

4.5 X 10-5 1.1 X 10-4 1.6 X 10-4

1.1 X 10-4 1.3 X 10-4 1.6 X 10-4

GP-B

300 ~ 350 ~ 400 ~

5.0 X 10-s 1.3 X 10-4 1.6 X 10-4

1.1 X 10-4 1.3 X 10-4 1.8 X 10-4

[Reproduced with permission from Wang, X. et al., Fuel Process. Technol, 38, 76, Elsevier (1994)] 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 deviated slightly from the straight line. This may show that side reactions such as dehydrogenation and polycondensation occurred significantly at 400 ~ It was assumed that tritium was transferred from gas phase to pitch by a first order reaction depending on the concentration of tritium in gas phase. Thus, the rate constants of hydrogen addition and exchange reactions at each temperature were obtained as shown in Table 3.9. Further, the apparent 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 given 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 hydrogen 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 exchange 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 kinetic parameters for the hydrogen transfer may reflect the differences in the structures and reactivities 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 mobility 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 mainly with the bond energy of C-H, O-H, N-H etc. in pitch molecules. Consequently, information 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 evidence 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 exchange (Derbyshire and Whitehurst, 1981). Such peri-condensated aromatics would be present 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 exchangeable components in the pitches, detailed composition analyses are required. However, the kinetic parameters obtained from the hydrogenation may be used to comprehensively 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 considered 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 behavior 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 carbonization, 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~

3 .~

60

tD O

~

0 [] A 9 []

40

20

0

Methane Ethane Ethylene Propane Partial pressure

60 120 180 240 Nominal reaction time (min)

2

r~

=

1

.~

0 ~

300

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

Table 3.11 Compositionsof Raw and Hydrogenated Coal Tar Pitch Pitch

HP-A

GP-A

GP-Bc

Reaction Reaction temp. time (~ (min)

Composition (wt%)a THFIA

BIS-THFS

HIS-BS

HS~

300 300 350 350 400 400 400

-120 300 120 300 0 120 300

27.2 32.6 32.3 38.1 36.8 36.3 38.7 50.6

5.5 2.5 3.4 2.4 1.3 3.0 3.1 4.0

52.6 42.9 40.2 35.3 38.4 53.8 48.2 33.9

14.7 22.0 24.1 24.2 23.5 9.9 10.0 11.5

300 300 350 350 400 400 400

120 300 120 300 0 120 300

7.7 21.3 15.3 14.6 29.3 33.0 56.7 70.0

5.4 17.2 7.6 13.3 19.2 9.4 10.4 7.1

69.6 48.8 49.1 64.9 42.6 49.9 27.0 13.9

17.3 12.7 28.0 7.2 9.0 7.7 5.9 7.5

300 300 350 350 400 400 400

-120 300 120 300 0 120 300

45.0 45.3 44.2 41.9 39.1 42.6 47.4 55.4

9.5 14.7 10.6 15.1 13.6 14.7 17.0 14.4

45.0 39.9 45.3 42.9 47.3 42.7 35.6 30.2

HS: Hexane soluble; HIS-BS: Hexane insoluble but benzene soluble; BIS-THFS: Benzene insoluble but tetrahydrofuran soluble; THFIS: Tetrahydrofuran insoluble. b A small amount of liquid products such as naphtha and light oil are contained. c In HIS-BS fraction, a small amount of BS fraction and liquid products are contained. [Reproduced with permission from Wang, X. et al., Fuel Process. Technol., 38, 76, Elsevier (1994)] a

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 activation 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 molecules 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 results suggest that, in the hydrogenation of coal tar pitch in the absence of catalyst and vehicle 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 hydrocracked 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 HPA 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

simultaneously during hydrogenation. HS fraction of HP-A pitch increased with temperature or reaction time below 350 ~ and HS fraction of GP-B pitch increased with temperature or reaction time only below 300 ~ However, the HS fractions of HP-A and GP-A pitches suddenly decreased and the corresponding HIS-BS fractions also decreased when the temperature reached 400 ~ for HP-A pitch and 350 ~ for GP-A pitch, respectively, especially for longer reaction times. The change in B IS-THFS fraction with temperature or reaction time did not appear to be a good reference for HP-A and GP-A pitches, since the B IS-THFS is an intermediate fraction and its content is the lowest in the two pitches. These results suggest that, at lower temperatures, the pyrolysis and polycondensation of pitch may occur as a side reaction at the same time, but at higher temperatures the polycondensation of pitch is dominant among the side reactions under such hydrogenation conditions. In addition, it should be noted that the increase in the THFIS fraction for GP-A pitch was more remarkable than that for HP-A pitch and the decreases of HS and HIS-BS fractions for GPA pitch started at lower temperatures than those for HP-A pitch. These results suggest that the GP-A pitch easily polycondensates through the reaction of thermally produced radicals to form a semi-coke, in comparison with the HP-A pitch. These characteristics are not so obvious for GP-B pitch. It seems that because the GP-B pitch has been heat-treated prior to the hydrogenation, the subsequent hydrogenation did not cause considerable change in its composition, especially at lower temperatures. B. Structural Analysis of Hydrogenated Coal Tar Pitch Using IH-NMR and Elemental Analysis Here, the effects of different hydrogen donor solvents on hydrotreatment of coal tar pitch and the mechanism of hydrogen transfer from solvent to coal tar pitch through establishing concentrations of functional groups in pitches, utilizing IH-NMR and elemental analysis are addressed. Regarding structures of pitches, analyses have been performed using IH/~3C NMR by Dickinson (1985), and earlier by Seshadri et al. (1980). In their works, hypothetical average molecular structures for the whole pitch and its fractions were determined by ~H and 13C NMR spectroscopy. In Japan, a similar method for the structural analysis of heavy oil derived from coal hydrogenation has also been developed by Hasegawa et al. (1980). A fundamental consideration in the method is the selection of an appropriate combination of aliphatic and alicyclic structures substituted to aromatic rings. A list of various probable structures is prepared and stored in the computer memory. For a given molecular structure, the total number of carbons and the number of aromatic carbons are part of the input data. A combination of an aliphatic group is chosen by the computer such that the difference between the total number of molecular carbons and the number of aliphatic carbons will match the number of aromatic carbons. By this method, an audio-visual average chemical structure for a given mixture can be provided. However a single average chemical structure so obtained is deficient in its physical meaning. Here, the method of functional group analysis was utilized to characterize the structure of coal tar pitch. The first step in determining the desired group concentration is to propose a set of functional groups representing the possible structures present in the pitches. Based on information from the literature as well as experimental results obtained from IR analysis, etc. (Wang et al., 1994a), a set of functional groups representing the possible structures present in coal tar pitch has been proposed as shown in Fig. 3.24 (Takeuchi et al., 1994). Once the functional groups have been specified, concentrations which satisfy the available analytical data must be found. This is done by relating the concentrations of the func-


161

tional groups to the experimental data through a set of balance equations. The set of all balance equations can be expressed in the matrix form (3.5)

~ AijYj = bi(i= l ..... m) j=l

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 stoichiometric coefficients. If the functional groups of Fig. 3.24 are combined with the data of elemental 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 solution 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 solution, then the proposed set of functional groups is insufficient to describe the observed data

1

0

4

9

(3------ CH3

10

~

CH3

11

D

CH3

12

6 Notation:

13

O------ CH2---C)

14

C)

OH

O----- Bound directly to an aromatic ring. Bound to act carbon or to an aromatic ring. Bound to a t3 carbon or farther from an aromatic ring.

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 limited, even through the number of solutions is mathematically defined as infinite. The equations dealt with in the present work belong to this case. To select a single solution from the feasible region, we can use the following computational 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 minimization 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 tetrahydroquinoline (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) No.

Sample

Conditions of hydrogenation a Solvent

1

2 3 4 5 6 7 8 9 10 11

12 13 14 15 16 17 18 19 20 21 22 23 24

Temperature (~

R-BI

m

__

H-T410-B I H-T430-BI H-T450-BI H-T470-BI R-BS H-T410-BS H-T430-BS H-T450-BS H-T470-BS R-HS H-T410-HS H-T430-HS H-T450-HS H-T470-HS H-Q410-BI H-Q430-BI H-Q450-BI H-Q410-BS H-Q430-BS H-Q450-BS H-Q410-HS H-Q430-BS H-Q450-BS

Tetralin Tetralin Tetralin Tetralin ~ Tetralin Tetralin Tetralin Tetralin ~ Tetralin Tetralin Tetralin Tetralin THQ THQ THQ THQ THQ THQ THQ THQ THQ

410 430 450 470 ~ 410 430 450 470 ~ 410 430 450 470 410 430 450 410 430 450 410 430 450

Time (min)

60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60

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.

o~

[-

o

~n

0

o

u~

o

o o

o

0 o o

9

c~

?

,--~ C~ Cxl fxl Cxl C~ ~-~ ~ q

Cxl C~ O~ C~

o o o o o o o o o o o ~ o o ~

o

o

~E

o

0 ,.c~

o o

163

164

,,..a

~Z [..

0t)

o

~z

o

.=. . o,...~ ..a

o o

0

o

..,.q

r

[..

az 9

+ ? +

r,.)

+

8

o

o

0

o

-~/ o

o

o

O

o

0

o

0

r

o

"

o

~

e,i o

,,6 ~

v

o

o

O

~

~

o,, e,i , - ,,6

o,--Z ~

0

0~ ~

0

0~ r

0

o

O

o

~

o

o

o

o

o,.~

o

o

o


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 samples 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, provide 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-chemical characteristics of these samples. The results reported in this work show that the structural 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 parameters. Examination of the functional group concentration reveals that the effect of hydrogenation by tetralin was different from that by THQ. The pyrolysis or hydropyrolysis of aliphatic 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 aliphatic 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 radicals 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 concentrations 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 suggested that tetralin is a more effective solvent for hydrotreatment of coal tar pitch, especially 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 accordance 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 increases. 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 development of anisotropic texture during heat treatment. Bond scission of strongly bonded pitch structure occurs when the hydrogenation temperature 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 decreased 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 considered 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 encountered 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 carbonization. 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 hydrotreatment 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 HISBS 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 Gas 0.7% -633500 dpm/g

167

G

Naphtha 3.1% 186500 dpm/g

HS 11.2% 52000 dpm/g

Product - - _Light oil 7.5% 65900 dpm/g

N.

0.13\

0.07

HIS-BS 63.7% 40800 dpm/g

Hydrotreaed -pitch 88.6% 41600 dpm/g

LO m

BIS-THFS 12.9% 37800 dpm/g THFIS 12.1% 31600 dpm/g

G: Gas; N: Naphtha; LO: Light oil; S: Solvent; HP: Hydrotreated pitch. (Hydrogen (g)/Pitch (30 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 aromatic 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, indicating 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 unhydrotreated 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 practically 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 .,~ ~-

9-~ 80

)40 ~ ,-,

,.t=

9, ~

9~

. ,...~

,a=

60

03 m

4O 20 I

00

I

I

200

I

I

400

I

I

600

I

I

800

1000

Temperature (~

0 Tritium

A Hydrogen

[] Weight

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

0

i

0

i

200

400

!

i

600

i

800

1000

Temperature (~ 0 Raw pitch

9 Hydrotreated pitch

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, simultaneously. 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 HS

1O0

169

HIS-BS

100

~, 80 .~ 60

.~ 60

~ 4o

~ 4O

20

20

0

0

1O0

200

400 600 ' 800 ' 1000 Temperature (~ (a) BIS-THFS

0

.

0

80

9~ 60

60

.

.

.

.

' 2;0 ' 4()0 ' 6;0 ' 8;0 Temperature (~ (b)

1000

THFIS

100

~, 80

.

40

"~ 40 nr

20

20 00

' 2()0 ' 4()0 ' 6()0 ' 8()0 ' 1000 Temperature (~ (c) 0 Raw pitch

0

' 2()0 ' 400 ' 660 ' 8()0 ' 1000 Temperature (~ (d) 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)] difficult to polycondense and transfer to mesophase. The thermogravimetric analyses of pitches and each c o m p o n e n t of the pitches were used to investigate the kinetics of non-isothermal heat treatment assuming that first order for the rate of weight loss corresponds to the integral m e t h o d used by Lee and Beck (1984). The kinetic parameters for pitches and their components obtained by the above m e t h o d are given in Table 3.15. The activation energy of pyrolysis of hydrotreated pitch was 40.9 kJ/mol, which was slightly less than that of raw pitch (41.3 kJ/mol). The activation energies of four fractions of raw pitch were higher than those for the corresponding fractions of hydrotreated pitch except the HS and THFIS fractions. The activation energies and the corresponding frequency factors of the THFIS fractions of both pitches are the greatest. The HIS-BS and B I S - T H F S fractions which are the main components of H P C F have lower pyrolysis activation energies, and hydrotreatment further decreased them. These results of kinetic analyses indicate that the HIS-BS and B I S - T H F S fractions are the most reactive components in pitch, and can be easily converted into mesophase through thermal polymerization. These would be easily converted to more reactive components by hydrotreatment.

170

3 Pyrolysis Table 3.15 Kinetic Parametersfor Heat Treament of Pitch Sample

Temp. range

Kinetic parameter

(~

E (kJ/mol)

A (min- 1)

Raw pitch

Whole HS HIS-BS BIS-THFS THFIS

250-650 150-650 150-650 200-600 450-800

41.3 26.4 31.7 37.9 50.8

7.597 X 10 2.979 2.094 X 10 4.163 X 10 1.035 X 103

Hydrotreated pitch

Whole HS HIS-BS BIS-THFS THFIS

250--650 150-550 150-650 200-600 450-800

40.9 28.2 30.3 34.7 52.9

8.100 X 10 2.551 1.675 X 10 2.442 X 10 1.754 X 103

[From Wang. X. et al., J. Jpn. Petrol. Inst., 34, 318 (1991)] In commercial production, heat treatment of pitch is usually conducted in the range of 350 to 450 ~ in which the primary processes consist of removal of smaller molecules and formation of small mesophase spheroids (Lewis, 1987). Pitches, when heat-treated at higher temperatures, undergo a series of physical and chemical transformations to form an infusible hydrocarbon polymer designated as coke. Since the initial pitch feedstocks are complex mixtures that are composed of multiple aromatic clusters with branched chains or connected to each other with methylene bridges and/or biaryl linkages (Shu and Li, 1990), as pitches are heat-treated, pyrolytic component and low molecular components are removed, and reactive aromatic components are coked by repetition of dehydrogenation and polymerization. The volatilization of low molecular components and the decomposition of pyrolytic components are basically completed before reaching 600 ~ The dehydrogenation and polymerization of reactive aromatic compounds are sustained at higher temperatures. After the hydrotreatment, content of hydrogen in the pitch increased from 4.70 to 5.15 wt% (Table 3.16). In addition, some differences on the pyrolytic character of pitch were also observed (Fig. 3.28), except that the composition by solvent fractionation changed (Fig. 3.25). Those results suggest that aromatic tings in the pitch were partially saturated, except for the hydrocracking of side chains in aromatics during hydrotreatnent of pitch. Since the parts saturated by hydrogen are more easily pyrolyzed than the unsaturated aromatic tings during the heat treatment, the pyrolysis activation energies of the HIS-BS and BIS-THFS fractions in the hydrotreated pitch were decreased slightly. In contrast, the pyrolysis activation energies of the HS and THFIS fractions increased after hydrotreatment. It Table 3.16 ElementalAnalysis after Heat Treatments of the HydrotreatedPitch Heat treament Temp. (~ None 400 500 600 700 800 900 1,000

Elemental composition (%) C

H

N

O (diff.)

H/C

92.78 93.35 93.65 94.73 94.57 94.69 95.25 97.27

5.15 4.97 4.46 3.57 3.03 2.53 1.87 1.20

1.23 1.22 1.16 0.92 0.96 0.97 0.92 0.88

0.84 0.45 0.73 0.78 1.44 1.81 1.95 0.65

0.055 0.053 0.048 0.038 0.032 0.027 0.020 0.009

[From Wang. X. et al., J. Jpn. Petrol. Inst., 34, 319 (1991)]


171

is considered that the increase of pyrolysis activation energy of HS fraction is due to the increase 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 saturate its aromatic rings. B. Pyrolysis Mechanism of Pitch The carbonization mechanism of coal- or petroleum-derived pitch in relation to the properties of the resultant carbon has been studied extensively over the past twenty years (Ehrburger and Lahaye, 1984). The main objective in understanding the carbonization phenomena in this field is to achieve the manufacture of satisfactory carbon products such as carbon fibers and metallurgical coke. It has been emphasized that the formation and growth conditions of the mesophase are important for the development of the anisotropic texture of a carbon product (March and Weaker, 1979). Further, it is recognized that the optical texture of mesophase formed during the carbonization of pitches is closely related to the structure and physical properties of the resulting carbons (Tillmans, 1985). The optical texture of mesophase depends to a great extent on the chemical composition of the parent material. As mentioned above, hydrogen transfer reactions in pitches play an important role in determining the development of the optical texture of mesophase during the early stage of carbonization (Wang et al., 1994b). Further, hydrotreatment has been recognized as an effective means to improve the property of raw pitch for high performance carbon fiber (Mochida et al., 1974). In particular, the distribution and mobility of hydrogen in pitch, which are important factors in determining the fluidity of the system, can be improved by hydrotreatment. It has also been reported that activation energies of pyrolysis of the pitch and their major components (HIS-BS and BIS-THFS fractions) decreased when pitch was hydrotreated (Wang et al., 1991). Thus, here, the carbonization mechanism in relation to the properties of raw pitch, particularly the hydrogen transfer during carbonization of pitches, which are the dominant reaction and would affect the property of resulting carbon, is described. In order to study the carbonization behavior of hydrogenated pitches, coal tar pitch (HP-B) and naphthalene pitch (NP), which represent two kinds of pitch with significant difference in the structural features of their component molecules, were chosen as starting materials. Fig. 3.29 shows the weight loss and decreases in hydrogen and tritium concentrations of HP-B hydrogenated by tritiated gaseous hydrogen using Ni-Mo/A1203 at carbonization temperature. The residual weight at 1000 ~ was 53.0%, which is lower than that (58.6%) of HP-B tritiated by isotope exchange. This result has been observed with coal tar pitch hydrogenated by hydrogen donor solvent, and an explanation for the weight loss has been reported (Wang et al., 1991). The rates of dehydrogenation and detritiation of the hydrogenated HP-B are all slower than the corresponding rates of HP-B tritiated by isotope exchange. However, the relative rate of dehydrogenation was faster than that of detritiation. This is the opposite of the result from the pitch tritiated by isotope exchange where the rate of dehydrogenation was slower than that of detritiation. These results suggest that hydrogenation can improve the thermal reactivity of pitch, and thereby the original hydrogen in pitch becomes easy to release. However, the hydrogen introduced into pitch by hydrogen addition will be more difficult to release than the original hydrogen in pitch. The carbonization of naphthalene pitch (NP) hydrogenated under the same conditions as those used for hydrogenation of HP-B showed a behavior similar to that of HP-B hydrogenated (Fig. 3.30). The weight loss at 1000 ~ was 86.5% lower than that (88.4%) of NP

172

3 Pyrolysis 100 I

~-

~

~

80

~40 .~

9 Weight 20

I-l Hydrogen A Tritium I

00

i

200

400 600 Temperature (~

800

1000

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 tritiated 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 naphthalene pitch, although the structural features of naphthalene pitch may be significantly different from those of coal tar pitch. The start and end of the carbonization process are indefinable. To facilitate comprehension 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 degradation 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 products are removed completely by thermal degradation. 100 [7~, o ~

80

.O.O ~

.,.

o ".~ 60 O

~ 40 = ~

9 Weight

20

D Hydrogen A Tritium

0

,

0

I

200

,

I

,

I

400 600 Temperature (~

,

I

800

,

1000

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)]


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 experimental 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 performance 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 exchangeable hydrogen and the optical texture of pitch developed at the early stage of carbonization 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 pitches were heat-treated under hydrogen atmosphere at 300-400 ~ the amount of THFIS fraction 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 dehydrogenation 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 development 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 molecules 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 exchangeability 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 exchangeable 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 behavior of hydrogenated pitches was quite different from that of pitches tritiated by isotope exchange. The original hydrogen in hydrotreated pitch was released easily. This phenomenon was also observed with Eureka pitch hydrogenated with Ni-Mo/A1203 (Mochida et al., 1987). It is generally assumed that hydrogenation improves thermal decomposition characteristics of pitch, hence the carbonization of hydrogenated pitches begin at lower temperatures. 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-

3 Pyrolysis

174

CH3 Isotope addition

~.1

---

H T

T

H

CH3 Isotope exchange

Ni-Mo/AIzO3, T2

H

H

(C) Fig. 3.31

H

H

CHT2

Pt/A1203, T20 H

(A)

H(T)

H

(B)

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 exchanged. 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 H

H(T)

or H

H(T)

H

etc.

9 Further

H(T)

(B)

H Fig. 3.32

T(H)

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)

@

CH~+ Fragments etc.

9 Further

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 hydrogen 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 temperatures. Based on the above results, the carbonization mechanism of coal tar pitch was estimated using the model system shown in Figs. 3.31-33. Methylnaphthalene (A) which is a typical 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 tritiumlabeled pitch. The orientation of tritium in the product after exchange reaction has been determined 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-


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 given 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 isotope exchange, where tritium was released more rapidly than hydrogen. In Fig. 3.33, however, thermal dehydrogenation or cleavage of an unsaturated ring may also occur as side reactions. 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 unsaturated 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.

Anealing time at carbonization temperature W]] Uncarbonizedregion [--] Carbonizedregion Fig. 3.34 Modelfor carbonization of pitch (Wang, 1993).

3.3.3

H y d r o g e n a t i o n a n d C a r b o n i z a t i o n o f C o a l Tar P i t c h in the P r e s e n c e of Catalysts

Organometallic complexes are expected to be transformed into highly-dispersed metals, and hence provide no influence to succeeding spinning, oxidation and carbonization processes. Further, the fine metal particles which are formed from a metal complex during the hydrogenation of precursor pitch at lower temperatures could also act as catalysts for succeeding carbonization or graphitization at higher temperatures (Yasuda and Miyanaga, 1989). In this section, the preparation of a new type of pitch which contains dispersed metal is discussed and the effect of several metal complexes on the hydrogenation of the coal tar pitch described. Further, carbonization of the hydrotreated pitch was carried out and the effect of fine metal particles derived from the decomposition of metal complexes added into the pitch is described. (Wang et al., 1992, Ishihara et al., 1993a, b, 1996).

176

3 Pyrolysis Table 3.17 Effectsof Various Catalysts on Hydrogenation of Coal Tar Pitch Product distribution (wt %)

Run 1 2 3 4 5 6 7 8

Catalyst(Amount)

Gas

Liquida

Pitchb

None Ni-Mo-A1203(2.06 g) Co-Mo-AI203(1.95 g) Fe3(CO)12(0.51 g) Fe3(CO)12(0.51 g)+ S (0.20 g) Mo(CO)6(0.79 g) Ru3(CO)12(0.08 g) Ru(acac)3(0.15 g)

1.0 1.3 0.9 1.2 1.0 0.8 1.7 3.9

9.5 22.8 17.1 14.1 12.6 10.8 12.6 15.6

89.5 75.9 82.0 84.7 86.4 88.4 85.7 80.5

Hydrogen transferred (g)C nadd (Aadd)

0.15 0.32 (57) 0.29 (49) 0.32 (58) 0.23 (42) 0.30 (51) 0.42(732) 0.35(552)

nex (Aex)

0.23 0.37 (46) 0.35 (39) 0.25 (7) 0.25 (7) 0.30 (51) 0.41(476) 0.26 (83)

Amount of hydrogen added from gas phase to pitch. Hex : Amount of hydrogen exchanged between gas phase and pitch. Aadd : Specific activity of catalyst based on hydrogen addition, g/mol. Aex: Specific activity of catalyst based on hydrogen exchange, g/mol. a Liquid : Light fraction(< 350 ~ bPitch( > 350 ~ c Specific activities are given in parentheses and they represent the following: Specific activity -- (Heat-H,o)/Weat (1) Heat:Amountof hydrogen transferred with catalyst (g). Hno: Amount of hydrogen transferred without catalyst (g). Wear:Amount of active metals in catalyst (mol). [From Wang. X. et al., J. Jpn. Petrol. Inst. 34, 453 (1992)] Hadd :

The results with several organometallic complexes and molybdenum-based supported catalysts used as references are summarized in Table 3.17. In the absence of a catalyst (Run 1), the yield of light fraction was 9.5 wt% (d.a.f, coal tar pitch); the amounts of hydrogen transferred between gas phase and coal tar pitch were 0.15 g (hydrogen addition) and 0.23 g (hydrogen exchange), respectively. In the reactions using commercially available m o l y b d e n u m - b a s e d supported catalysts, the catalysts were added into pitch in equal amounts with active metals (Ni + Mo or Co + Mo -- 3.0 mmol) (Run 2 and Run 3). When supported catalysts were used, the amounts of hydrogen added from gas phase to pitch increased to 0.32 g and 0.29 g for Ni-Mo/A1203 and Co-Mo/A1203, respectively. The amounts of hydrogen exchanged between gas phase and pitch increased to 0.37 g for the Ni-Mo/AlzO3 and to 0.35 g for the Co-Mo/AI203. The conversions into light fraction increased to 22.8 and 17.1 wt% for the Ni-Mo/AlzO3 and Co-Mo/A1203 catalysts, respectively. When Fe3(CO)12 (Run 4) and Mo(CO)6 (Run 6) were used, the amounts of hydrogen transferred were similar to those of supported catalysts, but the yields of light fraction were much lower than those of supported catalysts. In particular, the product distribution in the presence of Mo(CO)6 was nearly the same as that with no catalyst. Further, more active organometallic complexes for hydrogen transfer from gas phase to pitch were found to be ruthenium carbonyl (Run 7) and ruthenium acetylacetonate (Run 8). When Ru3(fO)le (0.37 mmol of Ru) was added to pitch, the amount of hydrogen added from gas phase to pitch increased to more than 0.4 g. The use of Ru3(CO)12 gave lower yields of light fractions than supported catalysts, while the use of Ru(acac)3 gave slightly higher yield of light fraction than other metal carbonyls. From these results, it is suggested that the organometallic complexes, especially Ru3(CO)12, give higher hydrogenation activity and lower hydrocracking activity relative to the commercially available molybdenum-based supported catalysts. Further, the specific activities of hydrogen addition (Aadd) and hydrogen exchange (Aex) for organometallic complexes have been compared with those for supported catalyst as shown in Table 3.17. The specific activities of hydrogen-added ruthenium complexes were about 10 times higher than those of the molybdenum-based supported catalysts.


177

Table 3.18 Compositionsof Raw Pitch and Hydrogenated Pitches by Solvent Extraction Run Raw pitch Run 1 Run 2 Run 3 Run 4 Run 5 Run 6 Run 7 Run 8

Composition (wt%) HS

HIS-BS

14.7 16.3 6.8 5.9 8.1 3.7 5.2 9.5 2.0

52.6 33.6 49.5 54.5 56.2 62.2 63.9 62.5 70.2

BIS-THFS 5.5 3.7 17.9 12.2 11.2 15.7 5.6 10.2 8.3

THFIS 27.2 46.4 25.8 27.4 24.5 18.4 25.3 17.8 19.5 (Wang, 1993)

The raw pitch and hydrogenated pitches were fractionated with solvents and the compositions of those are shown in Table 3.18. The changes in pitch composition after hydrogenation with catalysts (Runs 2-8) were quite different from that after hydrogenation without 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 hydrogenated pitch showed that the organometallic complexes were well dispersed into pitch and converted into fine metal particles of reduced states. These metal particles are considered to be active for hydrogenation of polynuclear aromatics in pitch. It is well known that the carbonization behavior of pitches is influenced by the presence 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

Naphthalene pitch

IF

v

~h, W

A W

9~ 70 Coal tar pitch

o . ,...~

O o

.~ 5o O 30 0.0

I

,

1.0

I

2.0

Content of Mo (wt%) Fig. 3.35 Effectof content of molybdenumon carbonizationyield of pitch (9 Method 1, 9 Method2). [From Ishihara. A. et al., Ind. Eng. Chem. Res., 32, 1724(1993)] 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 periodic 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 conventional supported catalysts. As the hydrogenated coal tar pitches containing fine metal particles 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 particles 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 H T~.,T

H

3

I

T-

T H

HT.

~,

H

H(T)

H

H(T)

~

H

(T)~.

T

H

T

179

H

H(T) D

H(T)

H

T

H

T

H

1

H

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

H(T) ~

C

T

H

3

H

T

H

H T

H

H

H

H

f T

~ H

- Further condensation H(T)

H

T

H(T) ~

f T

H(T)

Fig. 3.36 Carbonizationscheme for coal tar pitch. Note: The ratio of tritium vs. hydrogenin the aromatic ring could be inferred to be about 50% according to the results reported (Garnett et al., 1967, 1971) if it is assumed that all the hydrogen in ot-CH2 was isotopicallyexchangedbut the tritium orientation in the aromatic rings is unclear. [From Ishihara. A. et al., Ind. Eng. chem. Res., 32, 1726 (1993)] 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 absence 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-2methylnaphthalene 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 platinum 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 considered 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 reaction 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 benzene, toluene and xylene yields can be grouped into three categories: (1) pyrolysis in reactive 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 development 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 future. The local distribution of petroleum in the world has resulted in several crises and supply interruptions, especially since 1973 when the Organization of Petroleum Exporting Countries (OPEC) first achieved sufficient power to have a major impact on world oil markets (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 attractive 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 reserves. Therefore, coal will be more widely available than crude oil in the future. It is necessary to develop new processes for transformation of solid coal into clean liquid from the viewpoint of energy security in the future. 4.1.1

Coal Liquefaction

Many scientists and engineers believe that in the future coal liquefaction will be required to meet the demand for liquid transportation fuels. Coal liquefaction means the conversion of solid coal to fuel liquids (Whitehurst et al., 1980). Coal composed mostly of carbon and hydrogen and has lower hydrogen content than petroleum. Therefore, the process of coal liquefaction produces liquid compounds containing hydrogen at levels of approximately 10 to 15% by weight (Wu and Storch, 1968). Typical compositions for coals, liquid fuels, and some hydrocarbons are given in Table 4.1. Table 4.1 Compositionof Typical Fuel Oils and Hydrocarbons Fuel Typical crude oil Fuel oil Gasoline Bituminous coal Subbituminous coal Benzene Naphthalene

Carbon 86.0 86.0 85.0 78.0 71.9 92.3 93.7

Element, wt% (moisture and ash-free) Hydrogen Oxygen Sulfur 11.0 13.4 15.0 5.7 6.1 7.7 6.3

0.7 0.2 0.0 11.6 20.2 0.0 0.0

1.5 0.3 0.1 3.3 0.6 0.0 0.0

Nitrogen 0.5 0.1 0.0 1.0 1.0 0.0 0.0

182

4 Liquefactionof Coal

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 hydrogenolysis 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 capabilities. 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 reactor to process a low H2/CO ratio syngas from advanced coal gasifier could offer significant cost advantages over gas phase reactors (Gray et al., 1980, Brown et al., 1982). In cooperation with industrial organizations, DOE in 1981 began to support a research and development 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 development was successfully completed at the proof-of-concept (POC) scale in 1989, and advanced 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 production 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 producing 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 petroleum 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 future. 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 Coal

Liquefaction

183

Light oil Hydrogenation

f

r"~

Slurry preparation

Liquid products

Hydrogen

~ v

Distillation

M._._ Distillation} Heavy oil

Fig. 4.1 A flow diagramof the coal liquefaction process. 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 hydrogen 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 solvent 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 hydrogen-atom donor-solvent molecules, DH2, to form other radicals. A variety of recombination reactions terminate the chain reactions. Coal --->2Ri~

(4.1)

Ri" + DH2 ~ Rill q- DH-

(4.2)

Ri~ -Jr- Coal--) Rill -I-- Rj.

(4.3)

Ri9 -}- DH"---) Rill q- D

(4.4)

Ri ~ -4- Rj. ---->Rill -+- Argj

(4.5)

This view is supported by the general observation that free radical reactions control the pyrolysis chemistry of most organic substances. General kinetic features of coal liquefaction have also been used to support this view (Neavel, 1982). A detailed consideration of the chemical structure of coal and its reaction products also strongly suggests that free-radical reactions control coal chemistry. The aromatic and hydroaromatic units found in coal tars and liquids and presumed to be dominant structures in coal itself are known to be very reactive toward free radicals. Moreover, resonance-stabilized radicals derived from these

184


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 techniques 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 properties 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 temperatures 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 catalysts as cobalt-molybdenum has not altered process temperature or pressure requirements (Johnson, 1978). Research should be carded out to develop catalysts that will positively affect 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 distribution 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 determine 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 mechanism of the reaction involved in the catalytic effect, hydrogen transfer, etc.; and (4) develop 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 described.

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 integrated 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 second 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 molecules. Thus, hydrogen transfer reactions occur between the coal molecules and the components 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 catalysts. 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 attendant 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 fluid 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 transfer 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 liquefaction 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.

4.2 Coal Structure and Reactivity 4.2.1 Stages of Coal Liquefaction Prior to liquefaction, coal is often washed to remove inorganic minerals and dried. This process sometimes changes the structure and assemblage of coal macromolecules, which profoundly influences the reactivity of coal (Mochida and Sakanishi, 1994). In the preheater, coal with or without catalysts is rapidly heated to reaction temperature in the presence of solvent and pressurized hydrogen extensive decarboxylation, formation of carbonates, and dehydration take place in the preheater (Neavel, 1976). Coal is believed to be substantially dissolved in the preheater at this stage. Rapid heating of up to several hundred degrees per minute is believed to be very essential in obtaining high oil yields and preventing retrogressive reactions, which may take place at the same time. Catalysts are not expected to be effective in the preheater stage due to insufficient contact time. Hydrogen donor solvents play an important role in suppressing the retrogressive reactions at this stage, and it is important that the capacity of the donor not be exceeded in the preheater. The amount of hydrogen consumed from solvent has been considered to be relative to the heating rate. Slow heating rates allow more solvent dehydrogenation (Derbyshire et al., 1986b). It is well known that the viscosity increases very rapidly with bituminous coals that are dissolved in the solvent rapidly in the preheater. This sometimes causes problems of slurry transportation in narrow preheater tubes. The preheated coal slurry (essentially liquefied) is sent to the reactor, where thermal and catalytic cracking, hydrogenation and

186


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 conditions. 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 hydrocracked and/or hydrorefined products as well as rehydrogenated solvent. Various types of feeds, distillates, nondistillable liquids free of minerals, the catalyst, preasphaltenes and unreacted 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 series of reaction temperatures is expected to improve selectively specific reactions at different 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 liquefaction as shown in Fig. 4.2a-c. The first step in the liquefaction of solid coal is the formation 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 liquid phase (fusion and dissolution mechanisms). Other molecules undergo thermal fission at their weakest bonds, such as methylene and benzylether bonds, producing fragmented radicals (Whitehurstet et al., 1980). When the radicals are capped with hydrogen from the solvent 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 mechanism) (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 depolymerization. Coking takes place when such large molecules remain at elevated temperatures

4.2 Coal Structure and Reactivity

OH HO

O O

187

OH

H3C~

HO"~0 ( ~ ~ C H 3 ~"~"~-'~ 0 ~ . OH H O ' V N - ~ ~ ~ [ ~ )H ~.-.~'S'~~O-0 ~ . , ~ H O , ~ j [~J - ~ O I ~ ~ ~?H I ON ~ oO~ o ~O ~.H,~ ~.'~ o

a.f~o~.js~ ~

2.~ ~ e ~ ~

x-..,'m

" c~tT"_

H,C~ ~ ~_~ ~ HO~~OOH - ~"k ~ O

~ (H o ~ ~" N~O~L.~ C

O H

o"

~OOH "~"",Nvy

H, OH

o

H20

HO N HO O

O ~

H20

20 / NNN~ OHOH

/

3H8 (a)

2

H20 H20

CH4 H3C~~ ; ~ ~ ' ~ k H3C~' c~HH '" - ~ ~

CH3 l / - O ~ ~ Y ~ O ~ ~ ~ O % ~ H O ~ I ~ o H s ~ C ~ H 3 ~H20

\ .,o= ~ . . ~ L o ~ O - l / . .

~o't~,It'CH; H2S

H20 (b)

S

Fig. 4.2 The Shinn model of a bituminous coal structure [From Shinn, J.H., Fuel, 63, 1190, 1191, 1194 (1984)]

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 2,3-Diphenylbutane, C6HsC(CH3)H-C(CH3)HC6H5 2,3-Diphenyl-2, 3-dimethylbutane, C6HsC(CH3)2-C(CH3)2C6H5 9-(1-Phenylethyl) anthracene, 9-CI4H9CH2-CH2C6H5 9-Benzyl-9,10-dihydrophenanthrene, (9,10-C14HI1)-9-CH2C6H5 Bitetrayl, (1-CIoHIl)-(1-Cl0Hll) Benzyl phenyl ether, C6HsCH2-OC6H5 Benzyl phenyl thioether, C6HsCH2-SC6H5

1680 20 0.2 5 14 1 0.7 1

The 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)]

a

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 effective 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-related 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 relevant 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 conversion 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 already 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 reaction 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 suspended 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-dihydrophenanthrene to the reaction mixture inhibited the formation of radicals, whereas the addition 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 compatible 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

Rj,-+- ~

~

RjH + ~

(4.6)

(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 intermediate 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 dehydrogenation reactions. The fifth reaction in the basic sequence illustrates the notion that hydrogen atoms are transferred among coal molecules to yield both hydrogen-rich and hydrogenpoor compounds during the coal liquefaction process (Neavel, 1982). These processes are

~

+ Ri9

---> ~

--t--Rin

(4.8)

@

--t-Ri9

----) @ 9

-F-RiH

(4.9)

@o

+ Ri9

~

-+Rill

(4.10)

~

complemented by another series of reactions which are initiated by the abstraction of hydrogen 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


structures because the benzylic carbon-hydrogen bonds are relatively weak and the activation 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 reactions decrease the molecular weight of the coal molecules and the reactive products propagate the liquefaction reaction. C6HsCHCH2CH2C6H5 -'-) C6HsCH29 + 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 liquefaction, it is apparent that pericyclic reactions can make a major contribution to the degradation 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 temperatures 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 isomer, of 1,2-dimethylcyclohexane was obtained preferentially (6% yield) during the decomposition of the dihydronaphthalene. However, few other authentic examples of the reaction have been reported presumably because free radical chain reactions occur competitively obscuring 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 dihydroaromatic 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 hydrogentransfer 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

H

H3C

U3C H

(4.12)

Unimolecular dehydrogenation is almost certainly a more important process. Evidence for this reaction which leads directly to aromatic compounds is much more abundant (Derbyshire et al., 1982). These processes may be regarded as termination reactions in coal liquefaction. ~

@

+ H2

(4.13)

The notion that pericyclic reactions are important for carbon-carbon bond-cleavage reactions 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

simple hydrocarbons at the threshold temperature of liquefaction, 400 ~ (Virk, 1979). Tests of this proposal by the study of the rates and products of decomposition of 1,2diphenylethane, 1,3-diphenylpropane, and 1,4-diphenylbutane reveal that the retroene process is insignificant for the formation of the reaction products (Stein, 1981; Poutsma and Dyer, 1982; Hung and Stock, 1982). The situation is well illustrated by the pathway followed in the decomposition of labeled 1,4-diphenylbutane. The radical chain decomposition reaction predicts that unlabeled toluene [Eqs. (4.14)-(4.16)] be formed, whereas the C6HsCH2CD2CD2CH2C6H5 q- Ro --4 RD + C6HsCH2CDCD2CH2C6H5

(4.14)

C6HsCHzCDCD2CH2C6H5 ~

(4.15)

C6HsCH2CD=CD2-1-9

9CH2C6H5 + Tetralin --~ C6HsCH3 + 1-Tetralyl radical

(4.16)

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

H

+ C6HsCHzCD = CD2

(4.17)

D

[~

CH2 several steps > Toluene-2-d H

(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 _._)

-+- C6DsCD ----CH2

(4.19)

endothermic by no more than 38 kcal mol-1. However, the products formed in this reaction 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 endothermic by only 21 kcal mol-1. More important, the abstraction of the benzylic and CH2

CH2 ~ + C6HsCH--CH2

~

-q- C6H5CHCH3

(4.20)

192


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 support 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 ~

+Ri" ~ ~

-'l-Rill

(4.21)

Other pericyclic reactions such as the decarbonylation of phenolic aromatic compounds [Eq. (4.22)] and the fragmentation of tetralin [Eq. (4.23)] apparently are too slow to be important under the conventional conditions used in liquefaction reactions (Cypres and Bettens, 1974; Berman et al., 1980). Phenol

--+

Tetralin --+

~ ~

O -+ CO + Cyciopentadiene CH2

+ CH2 = CH2

(4.22) (4.23)

CH2

4.3 Catalysis in Coal Liquefaction The coal liquefaction proceeds even in the absence of the catalyst. There are essentially no use of catalysts in the processes developed in USA such as SRC, SRC-II, EXXON DONOR SOLVENT (EDS), in which the iron species present in ash of coal and hydrogen-donor solvent were utilized to enhance the liquefaction of coal (Suzuki and Ikenaga, 1998). On the other hand, the use of catalysts is considered to enhance the oil yield in coal liquefaction and has been spotlighted in a lot of studies, especially in Japan. In fact, the NEDOL process, which the catalyst and hydrogen-donor solvent was used, had showed higher oil yield than any present processes. The most conventional catalytic material is iron sulfide in various types. Iron-sulfur catalyst systems have been successfully employed for direct hydrogenation during coal liquefaction on a commercial scale (Wu and Storch, 1968). Nowadays, they are preferred because they are simple to use and because of economic reasons. Iron-sulfur catalyst systems have been widely investigated by several authors (Montano and Granoff, 1980; Montano et al., 1981; Cypres et al., 1981; Baldwin and Vinciguerra, 1983; Lambert, 1982; Stenberg et al., 1983; Mukherjee and Mitra, 1984; Trewhella, 1987). It was suggested that the highest conversion of coal to liquid products was associated with a pyrrhotite, which had the largest number of vacancies. Moreover, both H2S and pyrrhotite appeared to play a significant role in the coal liquefaction process (Montano and Granoff, 1980; Montano et al., 1981; Suzuki and Ikenaga, 1998). In most studies, it was recognized that the active form of iron-sulfur catalysts was pyrrhotite. There is an alternative interpretation that the catalyst was actually H2S produced from the reduction of pyrite (Lambert, 1982; Mukherjee and Mitra, 1984). Pyrite, pyrrhotite, and various nonstoichiometric sulfides are known, and pyrrotite is


193

postulated as the active form. Its precursors are red mud, residue of bauxite after the separation of alumina, iron ores of various sources, synthetic and natural pyrite, fine iron particles, 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 ironbased 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 petroleum 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 instability 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 hydrogenation and denitrogenation activities. Hydrogen sulfide in the reaction atmosphere has been reported to accelerate liquefaction directly, in addition to controlling the extent of sultiding 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 particle 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 diameter 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 hydrolysis 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/performance 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 distribution and acidic characteristics. The supporting procedure and the amount of supported sulfides are very influential in catalyst activity. Alternative supports to alumina are the focus 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 properties (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


the coal surface. Very high dispersion on the catalyst may allow direct interactions between 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 impregnated 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 highly 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 exchange, and dispersed impregnation of catalysts or catalyst precursors are combined to enhance 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 excellent 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 reported 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, suggesting 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 hydroclone; 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 distributed on the surface of coal as fine particles via impregnation showed higher yield of liquid 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 decrease of surface area even though its initial value was ca. 138 m2/g, resulting in lower catalytic 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., unground 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 surface 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 oxidation 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

+

2

0=

o 2:

I

I

I

I

0

+~

< z

o =

0 0 J

Pow~der [] CH2C12 soluble,

0 0 M

J

Y Impregnation

9 Heptane soluble

Fe203 5000 ppm, FeOOH 2500 ppm, Ammonium heptamolydate (AHM) 1500 ppm Fig. 4.3 Effect of several catalysts on liquefaction of Illinois #. 6 coal at 450 ~ and under 1000 psi of H2 for 1 h. [Reproduced with permission from Cugini, A.V. et al., Catal. Today, 19, 401 (1994)]

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 crystallization, 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 atmosphere for FeS2, ?'-FeOOH and limonite, it was found that ?'-FeOOH and limonite were completely 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 liquefaction of coal. It is important to elucidate the catalytic roles played by highly dispersed catalysts and sulfur which are added during coal liquefaction although these are not well defined. On the other hand, since the reactions involved in coal liquefaction include hydrocracking 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 Average radius (~m)

100

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 solvent, 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 dispersed iron catalysts (pyrrhotite and Fe(CO)5) and in the presence of sulfur, in which a tritium and radioactive 35S dual-tracer method was used. As we know, tetralin is one of the most simple and convenient model compounds because 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 between 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 Run

Catalyst

197

Sulfur

Feed

~:~:~:~1~ --

lg

3

Pyrrhotite

4

Pyrrhotite

lg

5

Fe(CO)5

lg

6

Fe(CO)5

2g

~iii!zi~

D Butylbenzene [-1 Methylindan

~: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)]

Run

Catalyst

Sulfur

1 2

-

-

9 []

lg

3

Pyrrhotite

4

Pyrrhotite

1g

5

Fe(CO)5

1g

6

Fe(CO)5

2g

Solvent Gas

Tritium equilibrium 0

I

I

I

20

40

60

'

I

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 therefore 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 solvent 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 dispersed 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 solvent 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 Table 4.3 HydrogenTransfer between Gas Phase and Solvent Run

Catalyst

Sulfur

1 2 3 4 5 6

m ~ Pyrrhotite Pyrrhotite Fe(CO)5 Fe(CO)5

m 1g ~ 1g 1g 2g

Amount of hydrogen Amountof H2S exchanged (g) generated(g) 0.013 0.156 2.032 2.428 0.275 0.410

0.000 0.800 0.121 1.011 0.012 1.041

[From Godo, M. et al., Energy Fuels, 11,726 (1997)] tained with pyrrhotite. In the case of Fe(CO)5, part of the added sulfur produced iron sulfide, 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 autoclave. 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 generate. 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 hydrocracking and the main product, naphthalene by dehydrogenation. Decalin by disproportionation 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 between tetralin-dl2 and hydrogen in coal at 400 ~ 1 h in a shaken autoclave system, protium 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 important 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 1tetralyl radical in this system may be the rate-determining step for both the conversion of

4.3 Catalysisin Coal Liquefaction ~

+ H.

(4.24)

T

HT or TSH

(4.25)

+ H. or SH. CH3

CH2 9

H9

+ H.

199

H9

.

(4.26)

(4.27)

+ SH9

(4.28)

+ H2S

(4.29)

+ SH9

(4.30)

-+- H2S

(4.31)

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 formation. 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 sulfur. 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 formation 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 simultaneously promoted both the thermal decomposition and polycondensation of coal. In Cat. --

T2

Cat.

+ H-Cat.

2T-Cat.

(4.32)

(4.33)

200


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 distribution in the solvent and decreased that in coal. The use of pyrrhotite only (Run 9) increased 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 distribution in the coal increased in spite of the presence of sulfur, suggesting that the catalyst derived from Fe(CO)5 and sulfur acted directly on the coal and increased both the rate of coal Run

Catalyst Sulfur

:o:.:***:.:,:o:.:.:.:.:.~ 8

-

-

9

Pyrrhotite

10

Pyrrhotite

1g

11

Fe(CO)5

1g

12

Fe(CO)5

2g

-

n THFIS VA BI-THFS im HI-BS IN HS F-! Light-oil ! ! Naphtha E! Gas

i !ii !!iii i ii!

lg -

:i:::::i::!:i:i::::::::!:~

'

0

I

'

20

I

'

40

I

60

'

I

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)] Run

8

Catalyst Sulfur

--

i I~

lg

9

Pyrrhotite

--

10

Pyrrhotite

1g

11

Fe(CO)5

1g

12

Fe(CO)5

2g

Coal Solvent Gas

iiiiiiii iiiiiiiiiiiii iiiii!iii i iiiiiiiill /iiiiiiii!ii!i!iiiiil 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)]


201

Table 4.4 HydrogenTransfer among Gas Phase, Solvent and Coal Amount of hydrogen transferred (g) Run

Catalyst

Sulfur

From gas to solvent

From gas to coal

7 8 9 10 11 12

--Pyrrhotite Pyrrhotite Fe(CO)5 Fe(CO)5

-1g -1g 1g 2g

0.229 0.301 0.320 0.510 0.187 0.213

0.129 0.116 0.328 0.207 0.335 0.397

Amount of hydrogenadded to coal (g) 0.264 0.335 0.330 0.348 0.354 0.415

[From Godo, M. et al., Energy Fuels, 11, 729 (1997)] conversion and the tritium transfer to coal. The amounts of hydrogen transferred from the gas to the solvent and from the gas to the coal, and the amount of hydrogen incorporated into the coal, are listed in Table 4.4. The amount of hydrogen incorporated into coal increased with the addition of catalysts and sulfur. In the absence of a catalyst (Runs 7 and 8), and in the presence of pyrrhotite (Runs 9 and 10), the amounts of hydrogen transferred from the gas to the coal in the presence of sulfur were less than those in the absence of sulfur. However, the amount of hydrogen incorporated into the coal in the presence of sulfur was greater than when sulfur was absent. It is therefore likely that sulfur promotes the hydrogen addition from the solvent to the coal. On the other hand, in the case of Fe(CO)5, the amounts of hydrogen transferred from the gas to the solvent were less than those when pyrrhotite was used, whereas the amounts of hydrogen transferred from the gas to the coal were greater; moreover, the amount of hydrogen incorporated into the coal increased significantly. This showed that Fe(CO)5 was effective in transferring the hydrogen from the gas to the coal, and suggests that the iron sulfide generated from Fe(CO)5 was dispersed successfully on the coal particles and was more effective in transferring the hydrogen from the gas to the coal than in transferring it from the gas to the solvent. In order to trace the behavior of added sulfur, the reactions were conducted using 35Slabeled sulfur. Fig. 4.9 shows the 35S distributions in the Taiheiyo coal liquefaction. In Run 4, using pyrrhotite and sulfur but no coal, 9% of the added sulfur was transferred into the catalyst after the reaction, corresponding to 5% of the sulfur atoms in the pyrrhotite. This shows that a sulfur e x c h a n g e reaction occurred b e t w e e n the added sulfur and pyrrhotite. In the presence of coal (Run 10), it is likely that a similar sulfur exchange would occur. In the presence of Fe(CO)5 and 1 g of sulfur (Runs 5 and 1 1), most of the 35S was distributed into the THFI fraction. The amount of sulfur used (lg) was not enough to produce H2S. In the presence of Fe(CO)5 and 2 g of sulfur (Runs 6 and 12), half of the 35S was distributed into the THFI fraction, and almost all the remaining 35S was distributed into H2S; the distribution into the solvent and the coal was negligible. The pyrrhotite catalyst is a nonstoichiometric iron sulfide which has a number of vacancies or defects on the catalyst surface. A possible mechanism for the sulfur exchange reaction is shown in Fig. 4.10. The added sulfur produced [35S]H2S, which was then assumed to dissociate into H and 35S-SH groups (SH group labeled by 35S) on the surface of the pyrrhotite catalyst. 32S-sulfur in pyrrhotite would generate 32S-SH, which would be in a state of equilibrium with [32S]H25. The H and SH groups are intermediates in the sulfur exchange reaction; this hydrogen atom may contribute to the promotion of the hydrogen trans-

202

4 Liquefaction of Coal Run Coal Catalyst Sulfur

10

12

--

Pyrrhotite 1 g

+

Pyrrhotite 1 g

--

Fe(CO)5

1g

--

Fe(CO)5

2g

+

Fe(CO)5

1g

+

Fe(CO)5

2g

II [-1 I VA

0

20

40

60

80

HzS Solvent Coal (THFS) Coal (THFI) + Cat.

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 exchange and addition via the radical reaction mechanism. [35S]HzS

[35S]H25

[35S]SH~

1

H, [

'~

[35515H.

([3'S]Fe'-xS.,

~

~..--...._.._.__..._Fig. 4.10 Possible mechanism of sulfue exchange. [From Godo, M. et al., Energy Fuels, 11,729 (1997)]

4,3.3

Non Fe-based Catalysts

H-coal method, by which higher yield of gasoline cut can be obtained with one-stage process by using a Co-Mo/A1203 catalyst to be used for the petroleum desulfurizafion, can be thought about to be a promising process if the loss of the catalyst can be prevented (Comolli et al.,, 1982). Molybdenum is one of higher active metals and is expensively studied. Some advances in recent research are introduced. Sakanishi et al. (1996) reported a result obtained in coal liquefaction using Ni-Mo catalyst supported on a hollow carbon micro particles (Ketjen Black: KB). The example of the result is shown in Fig. 4.11. When the catalyst, which is 3wt% of the coal, was added to the coal, oil yield obtained at 440 ~ for 60 min reached 54%, indicating Ni-Mo/KB catalyst was more active than the commercial Ni-Mo/A1203 catalyst and the synthesized FeS2 catalyst. The surface area of KB is 1270 m2/g and greater than that of A1203, Thus the active metal can more readily disperse on the former than on the latter. Further, the NiMo/KB catalyst is also suitable to upgrading of the primary liquefied oil of the coal (Sakanishi et al., 1997). Zhang et al. (1994) compared a Co-Mo catalyst supported on an alumina carrier, the surface of which was modified with TiO2 and ZrO2, or carbon by the thermal cracking of


203

60 50

O"

40 r~ "O - -

30

.,..~

9

20 10 ,

00

IFI

,

I

,

I

,

I

,

I

,

I

,

_..

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 effect. There is almost no difference among the catalysts, and the effect modified in the surface 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 sulfide catalyst when adding Mg. Especially, the conversion and oil yield increased respectively 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 redundancy as well. On the other hand, when large quantities of resolution radicals and aromatic 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 dehydrogenation 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


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 increased 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 naphthenate (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 conversion 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 catalysts 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 impregnating 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-micro-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 understanding of these reactions is regarded as significant for the process design. Generally, coal liquefaction consists of processes in which coal is thermally hydrocracked 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 experimental 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 system 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-


3H-Donation / z - - ~

205

~

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 products 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 extractions, HS, BS and PS represent hexane, benzene and pyridine soluble products, respectively; 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 butylbenzene produced during coal liquefaction are thought to be converted from tetralin solvent, 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


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 radioactivities 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 transfer 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 incorporation 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 reactions between tetralin and naphthalene was also slight. When the liquefaction was conducted 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 Coal (25.02 g) + 3H-Tetralin (75.18 g) + H2 (1.34 g) + Catalyst (5.08 g) + ~4C-Naphthalene (0.20 g) 3H: 63000 dpm/g E 14C:32500 dpm/g

I

Gas 50 kg/cm2 (9.36 g as H20) E total 142600 dpm 0 dpm

I Solid I Benzene Ext.

Liquid

,

I

I

Naphtha (2.34 g)

Solvent (68.01 g)

Light Oil (2.55 g)

SRC (11.15 g)

E 28500 0

] Tetralin (60.47 g)

E 22300 1200

E 9300 1500

HS

BS-HIS (4.88 g) r-" 9200 L 1100

E 58200 800 Naphthalene (7.54 g) E 40100 299000

(4.44 g) I- 9900 L 2100

I BIS I Pyridine Ext.

I

PS-BIS (1.83 g) I-- 7700 L 500

PS-BIS (0.48 g) 15700 3400

I PIS (15.45 g) I

Residue (6.34 g) 4400 E 300

Ash (4.03 g)

I Catalyst (5.08 g)

Catalyst, Ni-Mo-A1203; Initial H2, 5.9 MPa; Reaction temperature and time, 400 ~ and 30 min. Fig. 4.13 Balances of material and radioactivity of 3H and ~4C in Taiheiyo coal liquefaction at 400 ~ and under initial pressure of H2 of 5.9 MPa for 30 min in the presence of Ni-Mo-A1203 catalyst. [From Kabe. T., J. Jpn. Petrol. Inst., 39, 346 (1986)]


207

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, therefore, 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 distributions for the experiments using 3H labeled tetralin and unlabeled gaseous hydrogen. The product distributions indicate that the catalyst enhances the hydrocracking of preasphaltenes 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 naphthalene 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 presence 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 ~., 1

0

Without catalyst 0 ~

lOOi

~

"~ so

x

• ~

With catalyst x x ~ Naphtha

~.~ 80 ~

x

zx Light oil

~,

Oil

O

~ 6o

~, 60

4o)

40

-.~

I

0

~ 2o i i 30 60 90 120 Nominal reaction time (min) 9 : Unconverted Coal, O : I D +Oil,

300

0

~ : 9 + Preasphaltene, A : O + Light Oil,

I I I 30 60 90 Reaction time (min)

i 120

ID : ~ + Asphaltene, • : A + Naphtha

Reaction temp. : 400 ~ Initial H2 pressure: 60 kg/cm 2 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

With catalyst

1O0

10C o ....

90

9C I

80 -,~

o

~

5 ...

~

a

0

30

60 90 120 300 Reaction time (min) O 9Tetralin,

| 9 SRC,

[] 9Gas,

0

30


• 9Naphtha,

120

A 9Light oil

Fig. 4.15 Effect of reaction time on the tritium 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, 348 (1986)]

151 !

~=,~ l O ~

-S

[

[]

Z 5,-

I

30

I

I


I

120

E-l, In the presence of coal, catalyzed; II, In the presence of coal, uncatalyzed; O, In the presence of oil (HS); O, In the presence of asphaltene (BS-HIS); A, In the presence of solvent (tetralin); Catalyst, Ni-Mo-A1203; Initial H2, 5.9 MPa; Reaction temperature, 400 ~ Fig. 4.16 Effect of reaction time on production of naphthalene in tetralin solvent at 400 ~ and under initial pressure of H2 of 5.9 MPa in the presence of Ni-Mo-A1203 catalyst. [From Kabe. T., J. Jpn. Petrol. Inst., 39, 348 (1986)]


209

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, regardless 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 primary 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 hydrocracked 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 reaction 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 further 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 result is shown by SRCcalc and compared with the result of the hydrocracking of the unseparated 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 hydrocracking 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

.s

I

I

I

BS BIS

SRC

50

i

I

I

I

I

I

I

I

I

BS

i

i

li i S~SRC I BIS I

I

I

I

I

I

'iL

BS

i

i

HS

BS

BIS SRCcalc

I

HS

BIS

IL I

100 I 9

I BS BIS

HS

BS

i

I I

L N ill

HS

i I

I

fllL

N

L

iI N

I

HS

I

I

N, Naphtha; L, Light oil; HS,HS" BS, BS-HIS; BIS, PS-BIS; SRCcalc, Calculated from the data of HS, BS-HIS and BIS and these contents in source SRC. Fig. 4.17 Productdistribution in liquids from hydrocracking of coal. [From Kabe. T., J. Jpn. Petrol. Inst., 39, 349 (1986)]

210


Figure 4.18 shows the distributions of products and 3H after the liquefaction in the system 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 liquid 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 catalyst. 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 directly, 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 dehydrogenated immediately by coal and converts to 3H-containing naphthalene. The hydrogen transfer cycle of naphthalene --~ tetralin ~ naphthalene is very effective in the liquefaction 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 0 10

Catalyst

20

30

40

50

60

70

80

90

% 100

Ni-Mo-AI203 None 3H Ni-Mo-A1203 /

None

m

T LN

Residue:

~

PS-BIS: ~

BS-HIS: ~

Light oil: ~

Naphtha: ~

Gas:

Solvent:

I

i ] N: Naphthalene.

HS:

T: Tetralin.

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 Products 0 10

Catalyst

20

30

40

50

60

70

80

90

211

% 100

Ni-Mo-A1203 None 3H Ni-Mo-A1203 None Residue: ~

PS-BIS: ~

BS-HIS: ~

Light oil: ~

Naphtha: ~

Gas: ~

HS: ! ~ Solvent: [

I

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 0 10

Catalyst

20

30

40

50

60

70

80

90

% 100

Ni-Mo-A1203 None 3H 1

I m

N

Ni-Mo-A1203

None Residue: ~

PS-BIS: ~

BS-HIS" ~

Light oil: ~

Naphtha: ~

Gas: ~

Solvent:

HS: Solvent: I

I

I~1 N: Naphthalene, T: Tetralin.

Initial H2, 5.9 MPa; Reaction temperature and time, 400 ~ and 30 min. 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

9~

10

,':2 Z

o

rj

Residue

t PS-BIS

BS-HIS

t HS

Liquefied products

Light oil

t

Tetralin

Naphthalene

t Toluene Decalin

Solvents

Fig. 4.21 Tritium concentrations in liquefied products and solvents. [From Kabe. T., J. Jpn. Petrol. Inst., 39, 350 (1986)]

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 solvent, 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 catalyst. From the data above, the amount of hydrogen transferred by addition or exchange reactions 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 hydrogen 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 unconverted 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 hydrogenation of solvents in the absence of coal. In other words, the hydrogen exchange between gaseous hydrogen and solvents is suppressed since solvents cannot easily be adsorbed 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 Gas phase Without 0.01 /,/'~1.22 , catalyst r . . . . 0.01 .... ~ Coal = Solvent 0.88

With catalyst

Solvent Gas phase

Gas phase 9' ~ 0"60/.,'100080 ',~ 0.04 p(~," 0.36 _ ~ Coal ~ Solvent 0.76 3H-Tetralin H2

Gas phase

213

Gas phase

0"20j~,,'~1.20 0"24///9' ,' 1.32 yr . 0.40 yr '0.36 Coal = -~ Solvent C o a l . . . . . . . . ~ Solvent 0.76 0.04 Gas phase

Gas phase 9' 0.88~,,,~.60 0.84~',~0.04 Wandoan > Datong, which shows that coals with higher carbon content are more difficult to liquefy. Fig. 4.38 shows the variations in the tritium concentrations of residue and tetralin with temperature. At 200-230 ~ the tritium concentrations were very low and hydrogen exchange hardly occurred. At 300 ~ tritium was introduced to residue, indicating that hydrogen exchange reaction between coal (residue and SRC and gaseous hydrogen occurred at this temperature (vide infra). The tritium concentration of residue increased with temperature. However, the tritium concentration of tetralin remained very low below 350 ~ and it was much smaller than that of residue throughout the entire temperature range. The hydrogen exchange ratio is plotted against reaction temperature in Fig. 4.39. Although the liquefaction scarcely proceeded at 300 ~ the hydrogen exchange reaction of coal occurred. With a rise from 350 to 400 ~ the hydrogen exchange ratio increased remarkably and nearly 50% of 100 80 O

~- 60 ,~ 40 20

300

350 Reaction temp erature (~

400

Fig. 4.37 Effectof reaction temperature on yields of residue and SRC in liquefaction of several coals for 120, min. O, A: Datong; ~,/k: Wandoan; O, A: Morwell 0, ~, O: Residue; A,/k, A: SRC [From Kabe. T. et al., Energy Fuels, 5, 460 (1991)]

232


- -

X

Datong Wandoan Morwell

10

Residue Q ~ 0

Tetralin II t-A I--!

"0 0

5 0 0

V

U

200

300 Reaction temperature (~

400

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 related 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 extremely 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 50

o

.,,.~

Reaction time: 2hr

_

Datong 9 Wandoan Morwell O

40

~ 30 ~= 20

~Z

10

0

--O 200

A

.. 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%, respectively. Although the extent of liquefaction does not necessarily follow the rank or the carbon 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 hydrogen 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 stabilized 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 depending 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 phases, 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, tri100

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

~

~Jlv

10

Wandoan ~ 400 ~ Morwell

0

0

O 350 ~

I

I

I

I

I

1

2

3

4

5

Reaction time (h) Fig. 4.41 Changes in the hydrogen exchange ratio of coal with gaseous hydrogen with reaction time. [From Kabe. T. et al., Energy Fuels, 5, 461 (1991)] 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 tritium transfer from coal to solvent was very fast. The reason for this result is not yet understood. 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 increased 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 1

0

m

0

~

A--

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 ~ 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)]

(Kabe et al., 1991b)

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 addition 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 ratio 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; however, 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 values, 4.5, 5.0 and 7.8%, respectively. The HER for Morwell coal was the largest and therefore the hydrogen exchange at 300 ~ may be related to the exchange of hydrogen in functional 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

Datong 9 A

Residue SRC

)

Coal and Coal-Related Compounds: Structures, Reactivity and Catalytic Reactions

Coal and Coal-Related Compounds, Volume 150: Structures, Reactivity and Catalytic Reactions (Studies in Surface Science and Catalysis)

Kinetics of Catalytic Reactions

Catalytic Carbonylation Reactions

Catalytic Carbonylation Reactions

Kinetics of Catalytic Reactions

Catalytic carbonylation reactions

Kinetics of Catalytic Reactions

Inorganic Reactions and Methods- Reactions Catalyzed by Inorganic Compounds

Advances in Friedel-Crafts Acylation Reactions: Catalytic and Green Processes

Advances in Friedel-Crafts Acylation Reactions.. Catalytic and Green Processes

Kinetics of Catalytic Reactions--Solutions Manual

Nitration and Aromatic Reactivity

Chemistry and Chemical Reactivity

Nitration and aromatic reactivity

Kinetics of Catalytic Reactions - Solutions Manual

Nitration and aromatic reactivity

Chemistry and Chemical Reactivity

Copper Amine Oxidases: Structures, Catalytic Mechanisms and Role in Pathophysiology

Ligand Coupling Reactions with Heteroatomic Compounds

Organic Chemistry Syntheses and Reactivity

Chemical and Catalytic Reaction Engineering

Compounds of Uranium and Fluorine (Chemical Compounds)

Solvent Effects and Chemical Reactivity (Understanding Chemical Reactivity)

Diffusion and Reactivity of Solids

Solvent Effects and Chemical Reactivity

Bioactive compounds and cancer

Fullerenes: Chemistry and Reactions

Fullerenes: Chemistry and Reactions

European coal geology and technology

Compounds and Aggregates

Coal and Coal-Related Compounds: Structures, Reactivity and Catalytic Reactions

Coal and Coal-Related Compounds, Volume 150: Structures, Reactivity and Catalytic Reactions (Studies in Surface Science and Catalysis)

Kinetics of Catalytic Reactions

Catalytic Carbonylation Reactions

Catalytic Carbonylation Reactions

Kinetics of Catalytic Reactions

Catalytic carbonylation reactions

Kinetics of Catalytic Reactions

Inorganic Reactions and Methods- Reactions Catalyzed by Inorganic Compounds

Advances in Friedel-Crafts Acylation Reactions: Catalytic and Green Processes

Advances in Friedel-Crafts Acylation Reactions.. Catalytic and Green Processes

Kinetics of Catalytic Reactions--Solutions Manual

Nitration and Aromatic Reactivity

Chemistry and Chemical Reactivity

Nitration and aromatic reactivity

Kinetics of Catalytic Reactions - Solutions Manual

Nitration and aromatic reactivity

Chemistry and Chemical Reactivity

Copper Amine Oxidases: Structures, Catalytic Mechanisms and Role in Pathophysiology

Ligand Coupling Reactions with Heteroatomic Compounds

Organic Chemistry Syntheses and Reactivity

Chemical and Catalytic Reaction Engineering

Compounds of Uranium and Fluorine (Chemical Compounds)

Solvent Effects and Chemical Reactivity (Understanding Chemical Reactivity)

Diffusion and Reactivity of Solids

Solvent Effects and Chemical Reactivity

Bioactive compounds and cancer

Fullerenes: Chemistry and Reactions

Fullerenes: Chemistry and Reactions

European coal geology and technology

Compounds and Aggregates

Recommend Documents