Modeling the Chemical Structures of Coals with Different Classifications Using Mean Molecular Weights

Abstract

The chemical structure models for the extractions and residues of two types of bituminous coals, A and B, were constructed. The molecular weights of the extractions were determined via gel permeation chromatography (GPC). New standard materials with structures similar to those of coal extraction (i.e., 9, 10- diphenylanthracene, 5,6,11,12-tetraphenylnaphthracene, and chemical compounds A (Mw = 811) and B (Mw = 1135), which were synthesized using the coupling reaction) were adopted for GPC in order to obtain more accurate mean molecular weights than those in literature. Furthermore, a support program for constructing chemical structure models based on ¹H nuclear magnetic resonance (NMR) spectra was adopted. The coal models constructed suitably indicate the differences between the types of coal. In particular, it is found that a high pyridine-insoluble fraction extracted rate, which accounts for the most significant difference between the total extracted rates for coals A and B, enhance the coking property of coal A. In addition, the cluster size in the magic solvent-insoluble fraction might affect the softening property of coal.

1. Introduction

In previous research, two- and three-dimensional chemical structure models for coal were realized based on partial structures that were determined experimentally.^{1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29)} In particular, many researchers have devoted considerable attention toward the chemical structure of bituminous coal, owing to its unknown physical and chemical properties, as indicated by the phenomena of softening, melting, gasification, and dilatation. Therefore, it can be considered that previous research on the chemical structure models of coal is primarily composed of studies focusing on the chemical structure models of bituminous coal.^{10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27)} The first structural model for coal was constructed for bituminous coal by Fuchs et al.¹⁰⁾ In 1960, Given¹¹⁾ proposed a bituminous coal model based on X-ray diffraction, infrared spectrometry (IR), and nuclear magnetic resonance (NMR) data; this model mainly consisted of one or two aromatic rings. Subsequently, various chemical structure models for bituminous coal were proposed. For instance, the bituminous coal models constructed by Hill and Lyon¹³⁾ and Wiser^14,15,16) exhibited large distributions for the number of aromatic rings. Thereafter, several chemical structure models for coal were developed owing to the advances in analytical methods. Through their chemical structure model for coal, Kovac and Larcen¹⁷⁾ were the first to describe non-covalent bonding between such low-molecular-weight compounds as extracted fractions. This implies that low-molecular-weight compounds may serve as essential components for the coking property. Furthermore, Solomon¹⁹⁾ constructed a thermal decomposition model for bituminous coal, and Lazarov and Marinov²²⁾ described the hydrogen bonds in the chemical structure model for coal. The most prominent model for bituminous coal was constructed by Shinn in 1984; this model described low-molecular-weight components in polymer-like network molecules based on the pyridine solvent extraction ratio.

Early research has advocated γ component theories, which state that low-molecular-weight components dominate fluidity and coking properties.^30,31,32,33) Similarly, the metaplast theory^{34,35,36,37,38,39,40,41)} explains that the increase in low-molecular-weight components over temperatures of 360–420°C dominates the coking properties; this implies that these amounts influence the coking properties. Therefore, the amount of extraction in organic solvents can serve as an effective indicator for predicting the physical and chemical properties of coal. Iino et al.⁴²⁾ reported that the carbon disulfide and N-methyl-2-pyrrolidone (CS₂/NMP) mixed solvent afforded a high extraction yield for bituminous coal. Takanohashi et al.^43,44,45) constructed a chemical structure model for bituminous coal based on the extracted CS₂/NMP mixed solvent fractions and their residues.

Recently, the chemical structures of coals have been expanded to three dimensions owing to the developments in computer simulations;^26,27) this has facilitated descriptions of the non-covalent bond between aggregated molecules. To more accurately simulate the relaxation of aggregated structures, it is considerably important to describe molecular structures based on reliable information pertaining to the molecular structure obtained via experimental analyses. Gel permeation chromatography (GPC) has been employed to obtain the mean molecular weight during coal extraction, and NMR has been used to obtain the hydrogen and carbon structures of coal. In many analyses of coal, the mean molecular weight is the most obscure structural information. This is because GPC measurements do not directly yield the molecular weights; instead, these values need to be converted from the molecular volume by using standard materials. The chemical structures of polystyrenes, which are generally used as standard materials in GPC measurements, differ from the actual molecular structures of coals in terms of their aromatic rings and aliphatic branches. Therefore, in this research, new standard materials with structures similar to those of coals were used during the GPC measurements. It is significantly difficult to construct the complex molecular structure models for coal based on the structural information obtained via NMR, GPC, or ultimate analyses. In research pertaining to heavy oils, the supporting computer program developed by Sato⁴⁶⁾ is employed to obtain the average molecular structure based on analytical data. Accordingly, in this study, we constructed more reliable molecular structure models for coal extraction, as compared to those in previous research, by combining the mean molecular weight estimated using polycyclic aromatic hydrocarbons as the new standard materials in GPC measurements and by employing the support program for analyzing the average molecular structures with a personal computer.⁴⁶⁾

The purpose of this study was to concretely reveal the chemical structures of coal extractions and residues that function as key materials for the coking property.

2. Experimental

2.1. Preparation of Extracted Coal Samples

Bituminous coals A and B, which varied considerably with different maximum dilatation (MD) values, were analyzed according to JIS M 8801:2008. These coals were ground to sizes smaller than 150 μm and used for the subsequent analyses. The results of the proximate and ultimate analyses and the MD values are listed in Table 1. After drying in vacuum at 80°C for 12 h, these coal samples were extracted with the CS₂/NMP mixed solvent under ultrasonic irradiation at room temperature (20–25°C). The residues were labeled as the magic solvent-insoluble (MI) fraction. The obtained extracts were fractionated with acetone and pyridine to form acetone-soluble (AS), pyridine-soluble (PS), and pyridine-insoluble (PI) fractions (These fractionation flows are shown in SI.1). The PS, PI and MI fractions were washed by acetone and the AS fraction was washed by mixed solvent which consists of acetone and water (1:4 volume). All the fractions, including AS, PS, PI, and MI, were dried under vacuum at 120°C for 12 h, in order to eliminate the solvent from the extracts. The ratio of each fraction is shown in Table 2.

Table 1. The proximate, ultimate analysis, maximum dilatation (MD) values and maximum fluidity (MF) by gieselar plastometer of coal samples (d.a.f.; dry ash free, d.b.; dry base).

Coal	Ultimate analysis (mass% d.a.f.)					Proximate analysis (mass% d.b.)		MD (%)	MF (log DDPM)
	C	H	N	S	Odir	Ash	VM
A	84.6	5.0	1.9	0.8	7.8	8.9	28.8	218	3.88
B	81.8	5.6	2.0	0.5	10.0	8.6	36.3	22	2.34

Table 2. Fractionation ratios of Acetone-Soluble (AS), Pyridine-Soluble (PS), Pyridine-Insoluble (PI) and Magic solvent-Insoluble (MI) fractions.

Coal samples	Residue ratio (% d.a.f.)	Extract fraction (% d.a.f.)
	MI	AS	PS	PI
A	61.8	4.7	15.1	20.3
B	73.6	7.1	17.8	0.4

2.2. ¹H NMR Measurements

The AS, PS, and PI fractions were evaluated using liquid state ¹H NMR in CDCl₃, pyridine-d5, and CS₂/d6-NMP solvents, respectively. These ¹H NMR spectra were measured using single pulse excitation with a 90° pulse length of 11.6 μs and a pulse repetition time of 10 s, under a static magnetic field with the strength of 9.4 T; the JEOL ECA400 NMR spectrometer was used. The chemical shift was referenced to the ¹H signal from tetramethylsilane (TMS) at 0 ppm. The solid state ¹H NMR spectrum of the MI fraction was obtained using the Agilent INOVA500 spectrometer under 11.8 T. Single pulse excitation combined with Depth⁴⁸⁾ was applied to eliminate the background signal originating from the probe. Single pulse excitation with a 90° pulse length of 11.1 μs and pulse repetition time of 5 s was employed. The magic angle spinning rate was set to 58 kHz to reduce the peak broadening caused by the ¹H–¹H dipole interaction.

2.3. ¹³C NMR Measurement

The AS fraction was measured using liquid state ¹³C NMR in the CDCl₃ solvent, with the JEOL ECA400 NMR spectrometer under 9.4 T. The 90° pulse length was 4.9 μs and the pulse repetition time was set to 30 s. A small amount of Chromium (III) acetylacetonate (Cr(acac)₃) was added to the sample as a paramagnetic relaxation reagent. The chemical shift was referenced to TMS. The solid state ¹³C NMR spectra of the PS, PI, and MI fractions were obtained using the Agilent INOVA500 spectrometer under 11.8 T. The spin-echo pulse sequence was used, and a repetition rate of 100 s was adopted for all samples in order to obtain quantitative spectra under the 20 kHz magic angle spinning. The chemical shifts were referenced to adamantine as an external standard.

2.4. GPC Measurements for Coal Extraction

The mean molecular weight of each extracted fraction was determined via GPC measurements; four types of polycyclic aromatic hydrocarbon compounds with structures similar to those of coal were used as the standard materials for calibration. GPC measures the retention time, which is based on the molecular volume of the solute in the solvent. The relationship between the molecular volume and molecular weight of the solute in the solvent was determined based on the polarity and steric hindrance of the solute. Therefore, the measured solute and standard materials should have similar chemical structures. The chemical structures of these standard materials are presented in Fig. 1. Reagents (a) 9, 10- diphenylanthracene and (b) 5,6,11,12-Tetraphenylnaphthracene, used as the standard materials, were procured from Tokyo Chemical Industry, while (c) compound A and (d) compound B were synthesized using the coupling reaction.⁴⁹⁾ All the samples were dissolved in the CS₂/NMP solvent. The concentrations of the AS and PS fractions and standard materials ranged from 0.50 to 2.00 mass% and that of the PI fraction ranged from 0.25 to 0.75 mass%. A CS₂/NMP mixed solvent was used as the eluent in all GPC experiments.

Fig. 1.

Polycyclic aromatic hydrocarbon compounds used as the standard materials for calibration: (a) 9, 10- diphenylanthracene, (b) 5,6,11,12-Tetraphenylnaphthracene, (c) compound A, and (d) compound B; their molecular weights were 330, 507, 811, and 1135, respectively.

2.5. Model Construction

Chemical structural models for each extraction were constructed using the support program developed by Sato,⁴⁶⁾ where the quantitative information obtained via ultimate analyses, ¹H and ¹³C NMR spectra, and the mean molecular weight evaluated using GPC was used as the input. The chemical structural model of the residue (MI) fraction was constructed based on the number of aromatic ring carbons, estimated based on the ¹³C NMR results (refer to the chemical shifts in Table 3);⁵⁰⁾ the mean molecular weight was assumed as 3000, according to previous research.²⁵⁾ For all the chemical structural models, the functional groups of oxygen were determined via ¹³C NMR chemical shifts,⁵¹⁾ as shown in Table 3.

Table 3. Assignments for ¹³C NMR chemical shift.

Chemical shift (ppm)	Functional groups
185–240	Aldehyde and Ketone group
165–185	Carboxyl group
150–165	Aromatic O group
135–150	Aromatic C group
90–135	Aromatic H group
90–50	Aliphatic O group
50–22	Methylene and Methine group
22–0	Methyl group

3. Results and Discussion

3.1. ¹H NMR, ¹³C NMR, and Ultimate Analysis

The structural parameters obtained via the NMR measurements for each fraction of coals A and B are listed in Tables 4 and 5. The solid state ¹H NMR spectrum of MI fraction was curve-fitted and separated into 4 types (Ha, Hα, Hβ, and Hγ) of hydrogens, as in the liquid state ¹H NMR spectra. These values of f_a, H_ar/C_ar, and σ were calculated according to Eqs. (1), (2), and (3), respectively:   

$$fa=\frac{C-\frac{H_{\alpha }}{2}-\frac{H_{\beta }}{2}-\frac{H_{\gamma }}{3}}{C}$$(1)

  

$$\frac{H_{ar}}{C_{ar}}=\frac{H_{ar}+\frac{H_{\alpha }}{2}+O_{OH}+2\left(O_{in}+N+S\right)}{C-\frac{H_{\alpha }}{2}-\frac{H_{\beta }}{2}-\frac{H_{\gamma }}{3}}$$(2)

  

$$\sigma =\frac{\frac{H_{\alpha }}{2}+O_{OH}+2\left(O_{in}+N+S\right)}{H_{ar}+\frac{H_{\alpha }}{2}+O_{OH}+2\left(O_{in}+N+S\right)}$$(3)

where

C: total carbons; H_α: α hydrogens; H_β: β hydrogens; H_γ: γ hydrogens; H_ar: aromatic hydrogens; C_ar: aromatic ring carbons; C_OH: phenolic oxygen; and O_in, N, S: heterocyclic oxygen, nitrogen, and sulfur.

Table 4. ¹H NMR and ¹³C NMR analysis of AS, PS, PI, and MI fractions of coal A.

Fraction	1H NMR chemical shift (%)				fa	Har/Car	σ
	Ha	Hα	Hβ	Hγ
AS	32.1	33.0	27.6	7.4	0.69	0.78	0.44
PS	34.6	37.3	21.5	6.6	0.75	0.71	0.50
PI	28.6	51.9	14.9	4.6	0.75	0.66	0.59
MI	44.6	32.7	9.4	13.3	0.75	–	–

Table 5. ¹H NMR and ¹³C NMR analysis of AS, PS, PI, and MI fractions of coal B.

Fraction	1H NMR chemical shift (%)				fa	Har/Car	σ
	Ha	Hα	Hβ	Hγ
AS	22.86	33.93	33.51	9.70	0.64	0.78	0.55
PS	33.21	34.14	22.90	9.75	0.72	0.75	0.52
MI	29.64	12.77	48.00	9.59	0.64	–	–

All the extractions likely contained small amounts of the NMP solvent used for extraction; this is because a few sharp ¹H NMR peaks originating from the NMP were observed in each liquid state ¹H NMR spectrum. In particular, the AS fraction included the largest amount of the NMP solvent among all the extractions (approximately 8% ¹H derived from NMP). Therefore, the ultimate analysis values were corrected based on the amount of NMP solvent, which was calculated using the liquid state ¹H NMR spectra. The corrected ultimate analysis results for these fractions are shown in Tables 6 and 7. The heavy fractions exhibit a high percentage of oxygen. This implies that the hydrogen bonding in the aggregated structure of the heavy fractions is stronger than that in the lighter fractions.

Table 6. Ultimate analysis for AS, PS, PI, and MI fractions of coal A, with correction based on liquid state ¹H NMR.

Fraction	Ultimate analysis (mass% d.a.f.)
	C	H	N	S	Odir
AS	86.1	6.8	1.9	0.8	4.4
PS	83.7	5.4	2.5	0.8	7.7
PI	84.9	5.0	1.9	0.8	7.4
MI	83.3	4.8	1.9	0.8	9.1

Table 7. Ultimate analysis for the AS, PS, PI, and MI fractions of coal B, with corrections based on liquid state ¹H NMR.

Fraction	Ultimate analysis (mass% d.a.f.)
	C	H	N	S	Odir
AS	83.7	6.8	2.1	0.5	6.9
PS	81.3	5.4	2.6	0.5	10.2
MI	81.2	5.4	2.1	0.5	10.8

3.2. GPC Measurements for Coal Extraction

The GPC chromatograms of the four types of polycyclic aromatic hydrocarbon compounds used as the standard materials for calibration are shown in Fig. 2. In the chromatogram of compound A, two peaks appeared around 13.7 and 15.6 min. The peak with the longer retention time was attributed to the impurities in compound A, which was synthesized using the coupling reaction. This was confirmed via field desorption-mass spectrometry (FD-MS) measurements on compound A. The GPC chromatograms of the AS, PS and PI fractions are also shown in Figs. 3, 4, and 5, respectively. The calibration lines C1 and C2 in Fig. 6 were drawn based on the retention time for the four types of compounds. This new calibration line C2 exhibits smaller values of the slope and intercept than the conventional calibration line C1 for polystyrene. If the retention time of an extract is smaller than the intersection point of the two lines, the mean molecular weight of that converted by new calibration line C2 is smaller than that converted by conventional calibration line C1. The mean molecular weights of these fractions were estimated based on the calibration line C2. As shown in Tables 8 and 9, the mean molecular weights of the extracted fractions depend on their concentrations. As shown in Fig. 3, the leading edge of the side with the short retention time in the chromatogram exhibits a broader profile with an increase in the concentration of the solution. This implies that the solutes agglomerate with each other when the concentration exceeds 0.50 mass% AS fraction in the CS₂/NMP solvent. Obtaining the GPC chromatogram of the low-concentration solution was almost impossible owing to the distortion of the base line; consequently, determining the optimized concentration of the samples for dispersion in the solvent was difficult. We considered that the mean molecular weight of the AS fraction can be obtained using the calibration lines in Figs. 7(a) and 7(b), based on the assumption of infinite dilution, although a well-dispersed solution with finite concentration might be present. As a result, the number-average molecular weight and the average molecular weight for the AS fraction in coal A were estimated as 672 and 796, respectively. By contrast, considering the tendency of the leading edge in the GPC chromatograms of the PS and PI fractions to exist on the side with the short retention time, the agglomeration of solutes in the solution seemed negligible. Owing to significant dilution, the 0.25 mass% solution of the PI fraction exhibited an undulating baseline. The mean molecular weights of the PS and PI fractions in coal A were calculated using the GPC chromatograms for the 0.5 mass% solutions. The physical properties of polymer compounds are typically dominated by the weight-average molecular weight. These calculation process also have been performed in coal B. As a result, we determined that the mean molecular weight was 796 for the AS fraction, 2148 for the PS fraction, and 2164 for the PI fraction in coal A and 1054 for the AS fraction and 2560 for the PS fraction in coal B. For the same fractions, the mean molecular weight of coal B extraction was greater than that for coal A; however, the difference between the mean molecular weights of the AS and PS fractions was larger than that of the coals.

Fig. 2.

GPC chromatograms of 9, 10- diphenylanthracene, 5,6,11,12-Tetraphenylnaphthracene, compound A, and compound B. (Online version in color.)

Fig. 3.

GPC chromatograms for AS fraction of coal A.

Fig. 4.

GPC chromatograms for PS fraction of coal A.

Fig. 5.

GPC chromatograms for PI fraction of coal A.

Fig. 6.

Calibration line 1 (C1: ●) and 2 (C2: ▲) drawn based on plots of the polystyrene and polycyclic aromatic hydrocarbon compounds shown in Fig. 1.

Table 8. Mean molecular weight of extractions for coal A.

Sample	AS			PS			PI
Concentration	0.50%	1.0%	2.0%	0.50%	1.0%	2.0%	0.25%	0.50%	0.75%
Mn	693	669	713	1197	1210	1214	1073	1133	1104
Mw	859	923	1048	2148	2416	2455	2073	2164	2111

Table 9. Mean molecular weight of extractions for coal B.

Sample	AS			PS
Concentration	0.50%	1.0%	2.0%	0.50%	1.0%	2.0%
Mn	735	769	747	1341	1300	1203
Mw	1073	1093	1079	2597	2532	2605

Fig. 7.

(a) Number-average molecular weight and (b) weight-average molecular weight of AS fraction for each GPC concentration.

3.3. Chemical Structure Models for Coals A and B

The average chemical structure models for the AS, PS, PI, and MI fractions of coals A and B are shown in Figs. 8 and 9. These models of the extracts were constructed based on the chemical structural parameters (Tables 7 and 8) obtained using the support program developed by Sato.⁴⁶⁾ As shown in Fig. 8, the AS fraction model for coal A consists of two clusters and features five aromatic rings in each cluster. This is similar to the models reported in literature.⁴³⁾ By contrast, the AS fraction model for coal B, shown in Fig. 9, comprises three clusters and features three aromatic rings in each cluster. It was considered that the aromatic ring development depends on the degree of coalification. This tendency can also be applied to the other fractions. The PS fraction model for coal B exhibited clusters with five or six aromatic rings, whereas the MI fraction model only featured three aromatic rings in a single cluster. Given that the PS and MI fractions of coal A exhibited a maximum of six and five aromatic rings in one cluster, all the fractions of coal A had more aromatic rings per cluster than those of coal B. As the number of clusters for coal A was greater than that for coal B, it was concluded that the number of clusters increases with the progress of coalification. The MI fraction of coal B, in particular, had a large number of clusters with two or three aromatic rings in a single molecule. The chemical structure was associated with that of lignin.

Fig. 8.

Chemical structure models for coal A constructed using the support program developed by Sato,⁴⁶⁾ and chemical structural model for residue constructed based on the number of aromatic ring carbons, estimated mainly from the ¹³C NMR results.⁵⁰⁾

Fig. 9.

Chemical structure models for coal B constructed using the support program developed by Sato,⁴⁶⁾ and chemical structural model for residue constructed based on the number of aromatic ring carbons, estimated mainly from the ¹³C NMR results.⁵⁰⁾

The differences between the models of different fractions for the same coal were larger than those between the same fraction models for different coals. Thus, these models clearly indicated the differences between the types of coals.

Coals A and B exhibited different fluidity and considerably different dilatation abilities, as shown in Table 1. First, it was considered that the superior coking property of coal A originated from the high extraction rate, as discussed under the metaplast theory,³⁴⁾ and the high-temperature solvent fractionation.⁵³⁾ It was also concluded that the high extraction rate of the PI fraction, which accounts for the largest difference between the extraction rates of coals A and B, might enhance the coking property of coal A. In other words, the PI fraction in coal might be an essential component in coal softening. According to the chemical structure of the PI fraction model for coal A, the number of aromatic rings in a cluster is 6–7, which is the highest among all the fraction models. This implies that, among all the fractions, the strongest van der Waals force acts between the aromatic rings of the PI fraction, although the PI fraction can be dissolved in an organic solvent. Therefore, the agglomeration of large amounts of low-molecular-weight compounds at room temperature is essential for coal softening at high temperatures. Furthermore, from the point of view of self-dissolution model in which heavier components dissolved into lighter melted components continuously, it is considered that the expansion phenomenon cannot be maintained because the amount of PI fraction is small in coal B. In other words, it is suggested that a highly viscous film that can swell while retaining gas is not able to form in coal B. This is the probable reason why the expansion rate is low even though the fluidity of coal B shown in Table 1 is not so low.

Furthermore, it was also presumed that the chemical structure of the MI fraction in coals, which was the largest fractionation component, affects the coking property; this is because the cluster sizes of the MI fractions in coals A and B were considerably different. The MI fraction of coal A, featuring 4–5 aromatic rings per cluster, appears to yield certain components with structures similar to those of the AS and PS fractions during the coking process. These products of thermal decomposition might function as a binder-like extracted fraction. However, it is difficult to conclude that the MI fraction of coal B, featuring 2–3 aromatic rings per cluster, affords components that function as a binder during the coking process. These decomposition compounds originating from the MI fraction in coal B must possess low boiling points and sublimability owing to their low molecular weights.

Table 10. Chemical structural parameters for AS, PS, and PI fractions of coal A, obtained using a support program.⁴⁶⁾

INPUT DATA		AS	PS	PI
Total Carbon	Ct	59	155	159
Total Hydrogen	Ht	54	120	116
Aromatic carbon	Ca	42	116	118
Inner aromatic C	Caj	16	44	52
Aromatic H	Ha	18	44	35
α-H	Hα	18	42	56
Other non-aromatic H	Hβali	14	26	20
Terminal CH3	Hγ	4	8	5
PARAMETERS
Unit	M	2	5	5
Bridgehead carbon	Cai	16	52	56
Aromatic carbons attached side chain	P	4	9	12
Carbons attached side chain	L	6	14	17
Carbons in fused rings	Ctr	50	136	145
Total CH3 carbon	CH3	4	6	9
Total CH2 carbon	CH2	11	25	22
Total CH carbon	CH	2	8	10
TerminalCH3	Cr3	1	0	0
Branched-ethyl CH3	Cr2	0	0	0
Branched-methyl CH3	Cr1	0	3	2
PARAMETERS for RINGS
Total rings	Rt	12	38	43
Aromatic rings	Ra	10	31	33
Naphthenic rings	Rn	2	7	10
attached aromatics	Rna	2	6	10
PARAMETERS for AROMATIC CARBONS
Peripheral carbon	Cap	26	64	62
Substituted carbons	Cac	8	20	27
Tertiary carbon	Ha	18	44	35
Quaternary carbons	Caq	24	72	83
Inner bridgehead carbon	Caj	2	18	24
PARAMETERS for FUSED CARBONS
Carbons in fused rings	Ctr	50	136	145
Peripheral carbon	Ctp	30	70	69
Bridgehead carbon	Cti	20	66	76
PARAMETERS for NAPHTHENIC CARBONS
Naphthenic carbon	Cn	8	20	27
C at α	Cna	4	11	15
Bridgehead carbon	Cni	0	3	5
Bridgehead carbon at α	Cnai	0	1	5
Peripheral carbon	Cnp	8	17	22
CH2 at β	Cnb2	2	2	7
CH at β	Cnb1	2	5	5
Naphthenic hydrogen	Hn	14	32	44
Hydrogen at α	Hna	8	21	25
PARAMETERS for SIDE CHAIN
Aliphatic carbon	Cc	9	19	14
Aliphatic hydrogen	Hc	22	44	37
Number of side chains	N	5	10	13
Total CH3 carbon	CH3	4	6	9
Terminal CH3 carbons	Cr	1	3	2
Terminal CH3 carbon(CH3CH)	Cr3	1	0	0
Branched ethyl	Cr2	0	0	0
Branched methyl	Cr1	0	3	2
CH3 at α	C3ap	2	3	7
CH at α (Naphthen included)	C1ap	0	0	0
CH3 in ethyl at α	C3b	1	0	0
Carbons at β	Cb	4	7	0
Hydrogens at β	Hb	8	15	1
CH2 at β	C2b	3	7	0
CH at α,β	C1b	0	0	0

Table 11. Chemical structural parameters for AS and PS fractions of coal B, obtained using a support program.⁴⁶⁾

INPUT DATA		AS	PS
Total Carbon	Ct	75	179
Total Hydrogen	Ht	74	146
Aromatic carbon	Ca	42	130
Inner aromatic C	Caj	0	0
Aromatic H	Ha	18	52
α-H	Hα	24	48
Other non-aromatic H	Hβali	26	32
Terminal CH3	Hγ	6	14
PARAMETERS
Unit	M	3	6
Bridgehead carbon	Cai	12	52
Aromatic carbons attached side chain	P	4	10
Carbons attached side chain	L	8	14
Carbons in fused rings	Ctr	68	160
Total CH3 carbon	CH3	4	5
Total CH2 carbon	CH2	15	35
Total CH carbon	CH	14	9
TerminalCH3	Cr3	0	2
Branched-ethyl CH3	Cr2	0	0
Branched-methyl CH3	Cr1	2	3
PARAMETERS for RINGS
Total rings	Rt	18	42
Aromatic rings	Ra	9	32
Naphthenic rings	Rn	9	10
attached aromatics	Rna	5	10
PARAMETERS for AROMATIC CARBONS
Peripheral carbon	Cap	30	78
Substituted carbons	Cac	12	26
Tertiary carbon	Ha	18	52
Quaternary carbons	Caq	24	78
Inner bridgehead carbon	Caj	0	10
PARAMETERS for FUSED CARBONS
Carbons in fused rings	Ctr	68	160
Peripheral carbon	Ctp	38	88
Bridgehead carbon	Cti	30	72
PARAMETERS for NAPHTHENIC CARBONS
Naphthenic carbon	Cn	26	30
C at α	Cna	8	16
Bridgehead carbon	Cni	10	4
Bridgehead carbon at α	Cnai	2	4
Peripheral carbon	Cnp	16	26
CH2 at β	Cnb2	6	10
CH at β	Cnb1	4	4
Naphthenic hydrogen	Hn	38	52
Hydrogen at α	Hna	14	21
PARAMETERS for SIDE CHAIN
Aliphatic carbon	Cc	7	19
Aliphatic hydrogen	Hc	18	42
Number of side chains	N	8	9
Total CH3 carbon	CH3	4	5
Terminal CH3 carbons	Cr	2	5
Terminal CH3 carbon(CH3CH)	Cr3	0	2
Branched ethyl	Cr2	0	0
Branched methyl	Cr1	2	3
CH3 at α	C3ap	2	0
CH at α (Naphthen included)	C1ap	0	0
CH3 in ethyl at α	C3b	0	0
Carbons at β	Cb	1	4
Hydrogens at β	Hb	2	8
CH2 at β	C2b	1	3
CH at α,β	C1b	0	1

Table 12. Chemical structural parameters for AS, PS, and PI fractions of coal A. Values in parentheses are calculated according to Formula (2’).

fraction		Mw	fa	Har/Car	σ	Ultimate analysis (% d.a.f)
						C	H	N	S	O
AS	analysis	796	0.69	0.78 (0.66)	0.44	86.1	6.8	1.9	2.6	4.4
	model	796	0.71	0.62	0.39	87.5	6.7	1.8	0	4.0
PS	analysis	2148	0.75	0.71 (0.61)	0.50	83.7	5.4	2.5	2.0	7.7
	model	2145	0.72	0.62	0.42	84.6	5.4	2.6	0	7.5
PI	analysis	2164	0.75	0.66 (0.57)	0.59	84.9	5.0	1.9	1.9	7.4
	model	2186	0.74	0.56	0.53	85.7	5.0	1.9	0	7.3

Table 13. Chemical structural parameters for AS and PS fractions of coal B. Values in parentheses are calculated according to Formula (2’).

fraction		Mw	fa	Har/Car	σ	Ultimate analysis (% d.a.f)
						C	H	N	S	O
AS	analysis	1054	0.51	0.78 (0.64)	0.55	83.7	6.8	2.1	0.5	6.9
	model	1081	0.53	0.80	0.56	83.3	6.7	2.6	0.0	7.4
PS	analysis	2560	0.71	0.75 (0.65)	0.52	81.3	5.4	2.6	0.5	10.2
	model	2569	0.69	0.77	0.50	81.4	5.4	3.3	0.0	10.0

4. Conclusions

The chemical structure models of the extractions and residues of two types of coals were constructed using a support program. The mean molecular weight of the coal extractions, which is an important structural parameter when constructing models, was obtained via GPC measurements; polycyclic aromatic hydrocarbons with structures similar to those of coal molecules were used as standard substances to realize a calibration line. In this manner, more accurate extraction models, as compared to those in literature, with structural parameters that were similar to the analytical values were obtained.

These models can suitably indicate the differences between the types of coals. In particular, they indicated significant differences in the number of aromatic rings and cluster sizes. Thus, it was concluded that the PI fraction in coal, which has a mean molecular weight of approximately 2000 and 6–7 aromatic rings in one cluster, might serve as an essential component for the coking property and the cluster size in the MI fraction, while also affecting the softening property of the coal.

Supporting Information

Fractionation procedure for the extracts.

This material is available on the Journal website at https://doi.org/10.2355/isijinternational.ISIJINT-2021-459.

References

• 1)   I.  Wender: Catal. Rev. Sci. Eng., 14 (1976), 97.
• 2)   C. V.  Philip,  R. G.  Anthony and  Z.-D.  Cui: The Chemistry of Low-Rank Coals, ACS Symp. Series, Vol. 264, American Chemical Society, Washington, D.C., (1984), 287.
• 3)   E. A.  Wolfrum: The Chemistry of Low-Rank Coals, ACS Symp. Series, Vol. 264, American Chemical Society, Washington, D.C., (1984), 15.
• 4)   N.  Mallya and  R.  Zingaro: The Chemistry of Low-Rank Coals, ACS Symp. Series, Vol. 264, American Chemical Society, Washington, D.C., (1984), 133.
• 5)   P. J. J.  Tromp and  J.  Moulijn: New Trends in Coal Science, NATO ASI Series C, Mathematical and Physical Sciences, Vol. 244, Kluwer Academic Publishers, Boston, (1988), 305.
• 6)   K. J.  Hüttinger and  A. W.  Michenfelder: Fuel, 66 (1987), 1164. https://doi.org/10.1016/0016-2361(87)90319-X
• 7)   H.  Kumagai,  T.  Chiba and  K.  Nakamura: Prepr. Pap. Am. Chem. Soc., Div. Fuel Chem., 44 (1999), 633.
• 8)   Y. F.  Patrakov,  V. F.  Kamyanov and  O. N.  Fedyaeva: Fuel, 84 (2005), 189. https://doi.org/10.1016/j.fuel.2004.08.021
• 9)   J. H.  Shinn: Prepr. Pap. Am. Chem. Soc., Div. Fuel Chem., 41 (1996), 510.
• 10)   W.  Fuchs and  A. G.  Sandhoff: Ind. Eng. Chem., 34 (1942), 567. https://doi.org/10.1021/ie50389a010
• 11)   P. H.  Given: Fuel, 39 (1960), 147.
• 12)   L.  Cartz and  P. B.  Hirsh: Philos. Trans. R. Soc. Lond. A, 252 (1960), 557.
• 13)   G. B.  Hill and  L. B.  Lyon: Ind. Eng. Chem., 54 (1962), 36. https://doi.org/10.1021/ie50630a005
• 14)   W. H.  Wiser: Prepr. Pap. Am. Chem. Soc., Div. Fuel Chem., 20 (1975), 122.
• 15)   W. H.  Wiser: ACS Symp. Ser., 71 (1978), 29.
• 16)   W. H.  Wiser: Magnetic Resonance: Introduction, Advanced Topics and Applications to Fossil Energy, NATO ASI Series C, Vol. 124, D. Reidel Publishing Company, Dordrecht, (1984), 325.
• 17)   J.  Kovac and  J.  Larsen: Prepr. Pap. Am. Chem. Soc., Div. Fuel Chem., 22 (1977), 181.
• 18)   J.  Gibson: J. Inst. Fuel, 51 (1978), 67.
• 19)   P. R.  Solomon: New Approaches in Coal Chemistry, ACS Symp. Series, Vol. 169, American Chemical Society, Washington, D.C., (1981), 61.
• 20)   R. A.  Meyers: Coal Handbook, Marcel Dekker, Inc., New York, (1981).
• 21)   C. L.  Spiro and  P. G.  Kosky: Fuel, 61 (1982), 1080.
• 22)   L.  Lazarov and  S. P.  Marinov: Fuel Process. Technol., 15 (1987), 411. https://doi.org/10.1016/0378-3820(87)90062-2
• 23)   J. H.  Shinn: Fuel, 63 (1984), 1187.
• 24)   M.  Nomura,  K.  Matsubayashi,  T.  Ida and  S.  Murata: Fuel Process. Technol., 31 (1992), 169. https://doi.org/10.1016/0378-3820(92)90017-K
• 25)   T.  Takanohashi and  H.  Kawashima: Energy Fuels, 16 (2002), 379. https://doi.org/10.1021/ef0101154
• 26)   M. R.  Narkiewicz and  J. P.  Mathews: Energy Fuels, 22 (2008), 3104. https://doi.org/10.1021/ef700779j
• 27)   D.  Van Niekerk and  J. P.  Mathews: Fuel, 89 (2010), 73. https://doi.org/10.1016/j.fuel.2009.07.020
• 28)   A.  Vishnyakov,  E. M.  Piotrovskaya and  E. N.  Brodskaya: Adsorption, 4 (1998), 207.
• 29)   P. J.  Pappano,  J. P.  Mathews and  H. H.  Schobert: Prepr. Pap. Am. Chem. Soc., Div. Fuel Chem., 44 (1999), 567.
• 30)   A. H.  Clark and  R. V.  Wheeler: J. Chem. Soc., Trans., 103 (1913), 1704.
• 31)   C.  Cockram and  R. V.  Wheeler: J. Chem. Soc., 117 (1931), 854.
• 32)   T.  Shinmura: J. Fuel Soc. Jpn., 8 (1929), 379.
• 33)   T.  Shinmura: Rep. Fuel Inst., 14 (1932), 1.
• 34)   D. W.  van Krevelen,  C.  van Heerden and  F.  Huntjens: Fuel, 30 (1951), 253.
• 35)   D. W.  van Krevelen,  F.  Huntjens and  H. N. M.  Dormans: Fuel, 35 (1956), 462.
• 36)   H. A. G.  Chermin and  D. W.  van Krevelen: Fuel, 36 (1957), 85.
• 37)   H. N. M.  Dormans and  D. W.  van Krevelen: Fuel, 39 (1960), 273.
• 38)   I. G. C.  Dryden and  K. S.  Pankhurst: Fuel, 34 (1955), 363.
• 39)   I. G. C.  Dryden and  W. K.  Joy: Fuel, 40 (1961), 473.
• 40)   W. D.  Hertog and  N.  Berkowitz: Fuel, 37 (1958), 358.
• 41)   S.  Sarkar and  S. G.  Krishnan: Fuel, 42 (1963), 303.
• 42)   M.  Iino,  T.  Takanohashi,  H.  Ohsuga and  K.  Toda: Fuel, 67 (1988), 1639. https://doi.org/10.1016/0016-2361(88)90208-6
• 43)   T.  Takanohashi,  T.  Yoshida,  M.  Iino,  K.  Kato and  K.  Fukada: Energy Fuels, 12 (1998), 913. https://doi.org/10.1021/ef970239+
• 44)   T.  Takanohashi,  Y.  Terao,  M.  Iino,  Y.  Yun and  E. M.  Suuberg: Energy Fuels, 13 (1999), 506. https://doi.org/10.1021/ef980178v
• 45)   T.  Takanohashi,  H.  Kawashima,  T.  Yoshida and  M.  Iino: Energy Fuels, 16 (2002), 6. https://doi.org/10.1021/ef0101255
• 46)   S.  Sato: J. Jpn. Pet. Inst., 40 (1997), 46 (in Japanese).
• 47)   T.  Takanohashi,  M.  Iino and  K.  Nakamura: Energy Fuels, 12 (1998), 1168. https://doi.org/10.1021/ef980042x
• 48)   M. R.  Bendall and  R. E.  Gordon: J. Magn. Reson., 53 (1983), 365.
• 49)   K.  Sumi,  Y.  Niko,  K.  Tokumaru and  G.  Konishi: Chem. Commun., 49 (2013), 3893.
• 50)   T. H.  Fletcher,  S.  Bai,  R. J.  Pugmire,  M. S.  Solum,  S.  Wood and  D. M.  Grant: Energy Fuels, 7 (1993), 734. https://doi.org/10.1021/ef00042a006
• 51)   T.  Yoshida,  Y.  Nakata,  R.  Yoshida,  S.  Ueda,  N.  Kanda and  Y.  Maekawa: Fuel, 61 (1982), 824. https://doi.org/10.1016/0016-2361(82)90309-X
• 52)   Y.  Yun and  E. M.  Suuberg: Fuel, 72 (1993), 1245. https://doi.org/10.1016/0016-2361(93)90337-2
• 53)   R.  Ashida,  K.  Takashima,  K.  Miura,  N.  Okuyama,  M.  Hamaguchi and  T.  Shishido: Tetsu-to-Hagané, 100 (2014), 127 (in Japanese).