Regular Article
Effects of Rapid-heating and/or High-pressure Conditions on Coke Making, Examined Using a Synthetic Model Compound
2023 Volume 63 Issue 9 Pages 1502-1507
Details
2023 Volume 63 Issue 9 Pages 1502-1507
This study investigated the mechanisms by which rapid-heating and/or high-pressure conditions can improve the thermal fluidity of coal, using a synthetic compound as a model for low-molecular weight (Mw) compounds in coal. The compound had one quinoline ring and two naphthalene rings, C36H33N, Mw of 479, and a boiling point (bp) of ~520°C. Rapid heating (> 10°C/min) overcame the evaporation rate of the compound, whereas high pressure (> 1 MPa) increased the bp and suppressed the pyrolysis reaction. These conditions allowed the compound to remain until temperatures > 400°C, although it completely evaporated at 370°C under general heating conditions in a coke oven (3°C/min and 0.1 MPa). The effects of increasing the heating rate from 3 to 10°C/min at 0.1 MPa corresponded to the effects of increasing the pressure from 0.1 to 1.0 MPa at 3°C/min. The compound remaining at temperatures > 370°C can act as a mobile phase and hydrogen donor, thereby increasing the fluidity of coal. It can also serve as a reactant in the coking reaction and increase the coke yield.
The production of high-quality coke is essential for stable operation of blast furnaces during the iron-making process. Various manufacturing conditions have been investigated to produce high-strength coke that can increase the percentage of low-cost, low-grade coal in blended coal. To achieve this, the softening and melting (i.e., thermoplastic) properties of coal at approximately 370–460°C must be controlled1) because low-grade coals have low thermoplastic properties at such temperatures. Without the use of hydrogen, there are two coking conditions that will improve the thermoplastic properties: rapid heating and high pressure.
Rapid heating affects the thermoplastic properties of coal and increases the coke yield. Sakawa et al. examined the effects of heating rate (3, 10, and 20°C/min) on thermal weight decrease of coal in a nitrogen atmosphere.2) The coke yields at 900°C increased from 65% at 3°C/min to 68% at 20°C/min. Fukada et al. observed a reduction of coal viscosity with increasing heating rate (10, 100, 1000, and 1800°C/min), indicating significant improvement of coal softening and melting by rapid heating.3) Takanohashi et al. confirmed that the proportion of coking coal in blended coal can be reduced by this effect.4) The SCOPE21 process developed through a Japanese national project from 1994 to 2003 is a successful industrial process involving the application of rapid heating.5,6,7) Rapid heating of low-grade coals to 330–380°C increases their softening and melting properties at temperatures > 400°C, enabling an increase in their ratio to coking coal. The mechanism has been explained using a two-component system of coal with small molecules (mobile phase) embedded in a macromolecular three-dimensional crosslinked network (immobile phase).8) Based on the results of in situ nuclear magnetic resonance (NMR) imaging, Saito et al. suggested that the rapid heating caused thermal structural relaxation of coal, thereby increasing the amount of mobile component and improving its propagation and melting at high temperatures in the coal particles.9) Those authors attributed the effects of rapid heating to changes in physical properties during heating; they did not consider chemical reactions.
High pressure also influences thermal fluidity at approximately 400–450°C, increasing the coke yield. Kaiho et al. performed Gieseler plastometry using non-coking coal at a heating rate of 3°C/min in a pressurized nitrogen atmosphere.10) The maximum fluidity substantially increased from zero at ambient pressure to 6500 ddpm at 4 MPa. The authors of that study presumed that pressurized gas suppressed the evaporation of low-molecular weight (Mw) compounds from the coal system and/or enhanced the dissolution of gases in liquid products, resulting in a decrease in coal system viscosity. Lancet et al. also observed fluidity enhancement by pressurized nitrogen, with Pittsburgh No. 8 coal exhibiting maximum fluidities of 5300 dppm at 0.1 MPa and 36900 dppm at 2.4 MPa based on a heating rate of 6°C/min.11) Additionally, the pressure increased the coke yield from 65% to 67.5% when heated at 800°C for 1 h. Koba et al. heated various coals to 950°C in an autoclave under a nitrogen atmosphere, and the coke yields increased from 61–77% at ambient pressure to 71–79% at 10 MPa.12) Using a needle penetration method, Matsuoka et al. reported that nitrogen pressures of 1–3 MPa caused a non-coking coal to develop melting properties at 425–450°C during heating at 20°C/min.13) They suggested that the pressure caused chemical changes in coal during pyrolysis.
These previous studies suggested that high pressure and rapid heating affect the thermal behavior of low-Mw components. However, the extents of those effects could not be quantitatively predicted because the components were present as complex mixtures. Here, we synthesized a compound as a model for the low-Mw components in coal, then investigated its thermal behavior to clarify the effects of rapid-heating and/or high-pressure conditions.
Coal contains small molecules with Mw of 200–500;14) some of these molecules have a naphthalene structure (two aromatic rings),15,16) which can enhance the softening and melting properties of coal.17) Considering the purpose of this study, the model compound was required to be solid at room temperature and liquid during heating; it was also required to have a boiling point of approximately 500°C. A less crystalline molecule with an asymmetric chemical structure and heteroatom was preferable. No commercial products were found that fulfilled these requirements.
Accordingly, we synthesized a model compound with one quinoline ring and two naphthalene rings connected to three and four alkyl carbons, 2-(4-(naphthalen-1-yl)butyl)-3-(3-(naphthalen-1-yl)propyl)quinoline (2). Figure 1 shows the synthesis scheme. The target molecule is new, but the synthesis method was previously reported.18,19,20)
Synthesis scheme.
Column chromatography was conducted using an auto-flush chromatograph (Isolera™ One; Biotage) with silica gel column (Sfär Silica HC D 20 μm; Biotage) or a high-performance liquid chromatograph (LC-10; Shimadzu Corp.) with a silica gel column (Inertsil SIL-100A 5 μm; GL Sciences Inc.). Elemental analysis was performed using a CHNS analyzer (FLASH2000; Thermo Fisher Scientific). NMR spectra were recorded at room temperature on a 500 MHz instrument (ECA-500; JEOL Ltd.). Chemical shifts were referenced to residual solvent peaks (δ in parts per million [ppm] CHCl3 1H: 7.26 ppm; 13C: 77.0 ppm). High-resolution mass spectra were recorded on instruments (JMS-S3000, MALDI-SpiralTOF-MS; JEOL Ltd.) with 2-[(2E)-3-(4-tert-butylphenyl)-2-methylprop-2-enylidene]malononitrile (DCTB) as the matrix. The melting point was estimated by differential scanning calorimetry (DSC-60APlus; Shimadzu Corp.) at a heating rate of 3°C/min under a flow of nitrogen gas. The boiling point (bp) was measured by the simulated distillation (SimDis) method using a gas chromatograph (7890B(CP-SimDist); Agilent Technologies), with a paraffin mixture as the bp standard (C12H26–C60H122 for bp 216–620°C).
3.2. Synthesis5-(Naphthalen-1-yl) pentanal (1). 1-Iodonaphthalene (8.55 g, 33.6 mmol), palladium acetate (227 mg, 1 mmol), tetrabutylammonium chloride (18.69 g, 67.3 mmol), lithium chloride (1.43 g, 33.6 mmol), and lithium acetate (5.55 g, 84.1 mmol) were diluted in dry and degassed dimethylformamide (85 mL) in a 3-neck round-bottomed flask under an inert atmosphere. To the mixture was added 4-penten-1-ol (2.43 g, 33.6 mmol), and the solution was stirred at room temperature for 10 days. The reaction was stopped by dilution with deionized water and washing with ethyl acetate. The organic layer was separated, washed with brine, and dried over MgSO4. The solution was filtered, and the solvent was removed under reduced pressure. The resulting solid was flushed with a gradient solvent ratio from 100:0 to 85:15 hexane to ethyl acetate using the auto-flush chromatograph. The solvent was then removed, and a portion of the solid (1.1 g, 82.3% purity determined by 1H-NMR) was purified by preparative high-performance liquid chromatography with hexane/ethyl acetate (98/2 vol%) as the eluent to give the target product (0.75 g, yield 51.2%). 1H NMR (CDCl3, 500 MHz): δ 9.77 (t, J = 1.7 Hz, 1H), 8.02 (d, J = 8.4 Hz, 1H), 7.87–7.85 (m, 1H), 7.73–7.71 (m, 1H), 7.53–7.26 (m, 4H), 3.11 (apparent dd, second order coupling, Japp = 7.2, 7.2 Hz, 2H), 2.49 (td, J = 7.0, 1.6 Hz, 2H), 1.81–1.76 (m, 4H). 13C NMR (CDCl3, 126 MHz): δ 202.6, 138.0, 133.9, 131.7, 128.8, 126.7, 126.0, 125.8, 125.5, 125.4, 123.6, 77.3, 77.0, 76.7, 43.8, 32.8, 30.2, 22.1. The Supporting Information includes the 1H/13C-NMR spectra (Figs. S1 and S2).
2-(4-(Naphthalen-1-yl)butyl)-3-(3-(naphthalen-1-yl)propyl)quinoline (2). 1 (0.730 g, 3.44 mmol), aniline (0.153 mg, 1.64 mmol), and zirconocene dichloride (0.479 g, 1.64 mmol) were stirred with dry dichloromethane (40 mL) at 40°C in a round-bottomed flask for 1 day. The reaction mixture was diluted in dichloromethane and quenched with aqueous NaOH. The organic layer was separated, washed with brine, and dried over MgSO4. The crude product was dried, and the resulting solid was flushed through a silica gel column with a gradient solvent ratio from 100:0 to 90:10 hexane to ethyl acetate using the auto-flush chromatograph. The solvent was then removed and dried to give a sticky yellow solid (0.416 g, yield 52.9%). 1H NMR (CDCl3, 500 MHz): δ 8.04–7.99 (m, 3H), 7.87–7.81 (m, 3H), 7.72–7.70 (m, 3H), 7.64–7.61 (m, 1H), 7.51–7.30 (m, 9H), 3.21 (apparent dd, second order coupling, Japp = 7.5, 7.5 Hz, 2H), 3.07 (apparent dd, second order coupling, Japp = 7.7, 7.7 Hz, 2H), 2.94 (apparent dd, second order coupling, Japp = 7.8, 7.8 Hz, 2H), 2.88 (apparent dd, second order coupling, Japp = 7.9, 7.9 Hz, 2H), 2.17 (quint, J = 7.7 Hz, 2H), 1.92–1.78 (m, 4H). 13C NMR (CDCl3, 126 MHz): δ 161.8, 146.5, 138.6, 137.7, 134.8, 133.9, 133.8, 133.4, 131.8, 131.7, 128.9, 128.7, 128.5, 127.2, 126.9, 126.8, 126.4, 126.1, 125.8, 125.6, 125.5, 125.4, 123.9, 123.6, 35.7, 33.0, 32.7, 32.1, 31.2, 30.9, 29.6. High-resolution mass spectra (matrix-assisted laser desorption/ionization): m/z calculated for C36H33N ([M+H]+): 480.2691, found: 480.2203. Elemental analysis calculated for C36H33N: C, 90.15; H, 6.93; N, 2.92, found: C, 89.56; H, 6.85; N, 2.76. Estimated melting point: 64°C (differential scanning calorimetry), bp: 521°C (SimDis). The Supporting Information includes the 1H/13C-NMR spectra (Figs. S3 and S4), high-resolution mass spectra (Fig. S5), and differential scanning calorimetry profile (Fig. S6).
3.3. Thermal AnalysesA pyrolysis gas chromatograph–mass spectrometer (EGA/PY-3030D-7820A(DB-5MS UI)-5977E; Agilent Technologies) was used to analyze the products generated by flash pyrolysis, in which the sample was fed into a furnace heated to 400°C and reached the target temperature in 0.1 s. The weight loss behavior of compound 2 (2 mg) during the temperature increase (3, 10, or 50°C/min) under a flow of nitrogen gas at ambient pressure was investigated using a thermogravimetric analyzer (DTG-60AH; Shimadzu Corp.). The weight loss behavior under 1 MPa up to a certain temperature (350, 375, 400, 425, or 450°C) was measured using a high-pressure heat chamber (STJ-0123-1000NT; ST Japan Inc.). Importantly, the heating rate of coal in the real coking process is approximately 3°C/min.
The pyrolysis product of 2 (60 mg) was prepared using an autoclave (25 mL, ID 20 mm) with a thermocouple inside and an inner pressure monitor. Compound 2 was heated to 450°C at a heating rate of 10°C/min in a nitrogen atmosphere. The internal pressure of 0.8 MPa at room temperature increased to 1.8 MPa at 450°C. For reference, an experiment at 0.1 MPa and 450°C was also conducted using an autoclave. After cooling, the contents were recovered by washing with dichloromethane, and the dichloromethane was removed using a rotary evaporator.
Target compound 2 was successfully synthesized with sufficiently high purity as confirmed by NMR, mass spectrometry, and elemental analysis. The compound was sticky and did not crystallize. The properties fulfilled the requirements described in Section 2. Figure 2 shows the relationship between bp and carbon number for polycyclic aromatic hydrocarbons and n-alkanes. The bp of 2 (C36H33N, 521°C) was close to the bps of coronene (C24H12, 525°C) and tetracontane (C40H82, 523°C). Pyrolysis gas chromatography-mass spectrometry confirmed that the compound produced 1-methylnaphthalene and acenaphthene (Fig. S7), which are abundant in coal tar. Therefore, compound 2 was a reasonable model of the low-Mw components of coal.
Relationship between bp and carbon number for polycyclic aromatic hydrocarbons, n-alkanes, and model compound.
The heating rate substantially influenced the weight decrease of the model compound. The lines in Fig. 3 show the weight decrease of the model compound at 3, 10, and 50°C/min under a flow of nitrogen gas at 0.1 MPa. The compound began to evaporate above 250°C; evaporation was complete at a temperature below the bp (~520°C) because the evaporated molecules at below its bp were ejected from the system by the stream of nitrogen gas. Therefore, longer residence time (lower heating rate) caused a greater amount of evaporation (greater weight decrease) at a particular temperature. For example, the residue yields at 350°C at 3, 10, and 50°C/min were 0.39, 0.90, and 0.95, respectively; the residence times from 250°C to 350°C were 33, 10, and 2 min, respectively. Table 1 shows the temperature of complete evaporation (Tcev) observed in the thermogravimetric curves under each condition, which is the temperature at the crossing point of the thermogravimetric curve and the horizontal axis (zero weight). The Tcev values for 3, 10, and 50°C/min at 0.1 MPa were approximately 370, 420, and 485°C, respectively. The results indicated that the rapid heating conditions allowed low-Mw compounds to be present at higher temperatures.
Weight decreasing behavior of model compound.
| Heating rate/°C min–1 | Pressure/MPa | Tcev/°C |
|---|---|---|
| 3 | 0.1 | 370 |
| 10 | 0.1 | 420 |
| 50 | 0.1 | 485 |
| 3 | 1.0 | 410 |
| 10 | 1.0 | 430 |
At the same heating rate, high pressure increased the weight decreasing temperatures because of the increased bp. The markers in Fig. 3 show the yields of the model compound at 3 and 10°C/min in a flow of nitrogen gas at 1.0 MPa. The residue yields at 400°C at 3 and 10°C/min were 0.24 and 0.64 at 1.0 MPa, respectively. Tcev values for 3 and 10°C/min at 1.0 MPa were approximately 410°C and 430°C, respectively, as shown in Table 1. Thus, the weight decreasing behavior at 3°C/min and 1.0 MPa was nearly identical to the weight decreasing behavior at 10°C/min and 0.1 MPa. The effect of increasing the pressure to 1.0 MPa at 3°C/min corresponded to the effect of increasing the heating rate from 3 to 10°C/min at 0.1 MPa under these experimental conditions.
High pressure also suppressed the pyrolysis reaction. Table 2 shows the elemental composition, the atomic ratios of H/C and N/C, and the yield of each pyrolysis product prepared at 450°C and 0.1 MPa using the autoclave heated at 10°C/min. The yield of pyrolysis product at 1.0 MPa was 0.95, and the elemental composition was nearly identical to the composition of the raw material. In contrast, the yield of pyrolysis product at 0.1 MPa was 0.71, and the atomic balance changed. Notably, the yields were higher than the results shown in Fig. 3 because the evaporated compound was confined in the autoclave. Figure 4 shows the 1H-NMR spectra of pyrolysis products. The chemical shifts of 8.2–7.3 ppm (a) and 3.3–1.7 ppm (b and c) correspond to the hydrogens attached to aromatic rings and alkyl carbons, respectively. After pyrolysis at 450°C and 1.0 MPa, the product was presumed to contain the unreacted compound (2), 3-methyl-2-(4-(naphthalen-1-yl)butyl)quinoline (3), and 2-methyl-3-(3-(naphthalen-1-yl)propyl)quinoline (4), as determined by peaks at approximately 2.7 ppm (d) and 2.4 ppm (e). At 450°C and 0.1 MPa, the product was presumed to contain mainly 4 and 2,3-dimethylquinoline (5) because the c peaks nearly disappeared and the peak heights of d, e, and f increased. Considering the thermal decomposition mechanism (i.e., beta scission), it is reasonable to expect that the alkyl chain of 3 (-C4H8-) decomposes more readily than the alkyl chain of 4 (-C3H6-). The small peaks in the range of 6–5 ppm presumably represented the hydrogens of unsaturated hydrocarbons (-CH=CH-) formed by pyrolysis, which were more intense at 0.1 MPa than at 1.0 MPa. These results indicated that the pressurized conditions suppressed the decomposition of low-Mw compounds.
| C/wt% | H/wt% | N/wt% | H/C | N/C | Yield | |
|---|---|---|---|---|---|---|
| Model compound 2 | 90.1 | 7.0 | 2.9 | 0.92 | 0.028 | 1.0 |
| Product at 450°C and 1.0 MPa | 90.1 | 6.9 | 3.0 | 0.91 | 0.029 | 0.94 |
| Product at 450°C and 0.1 MPa | 89.8 | 6.8 | 3.5 | 0.90 | 0.033 | 0.71 |
1H-NMR spectra and peak assignments.
In previous studies, rapid-heating and/or high-pressure conditions influenced the thermoplastic properties of coal, with increased fluidity at temperatures > 400°C and increased coke yield. Assuming that model compound 2 represents the low-Mw compounds in coal, the mechanism can be explained as follows using the schematic model shown in Fig. 5.1)
Schematic model of the thermoplasticities of coals (Goonyella (GNY) and Witbank (WIT)) according to pyrolytic behavior and amount of transferable hydrogen. The size of the letter “H” is proportional to the amount of hydrogen in each form.
At a temperature of 370°C (above the plastic range of coal), low-Mw compounds act as hydrogen donors to stabilize radicals and suppress the cross-linking reactions among high-Mw compounds, leading to greater fluidity to form good coke (Fig. 5, top).21) When coal has lower amounts of low-Mw compounds in the plastic range, radical recombination reactions are likely to occur, causing lower fluidity and forming poor coke (Fig. 5, bottom). Therefore, the amounts of low-Mw compounds at temperatures > 370°C influence the fluidity and properties of coke. As mentioned above, the low-Mw compounds evaporate completely before 370°C at the general heating rate of 3°C/min and 0.1 MPa in a coke oven, resulting in the loss of transferable hydrogen. However, under rapid heating conditions > 10°C/min and/or at higher pressures > 1.0 MPa, some portions of the compounds persist at temperatures > 370°C and are involved in the coking reaction. The remaining compounds act as hydrogen donors and increase the fluidity of coal. Additionally, the compounds can act as reactants for coking reactions, resulting in an increase in coke yield. Importantly, the quality of coke can be improved by these effects, according to the mechanism shown in Fig. 5. Rapid heating and/or higher pressure allow the use of low-cost, low-grade coal while maintaining coke quality.
We synthesized compound 2 (C36H33N, Mw 479, and bp 521°C) with one quinoline ring and two naphthalene rings as a model of the low-Mw compounds in coal. Additionally, we examined it thermal properties to clarify the effects of rapid-heating and/or high-pressure conditions on coke making. The use of synthetic compounds with well-defined chemical structures and properties allowed us to clarify the effects of these conditions.
Rapid heating (> 10°C/min) and/or high pressure (> 1 MPa) conditions allowed the model compound to persist until temperatures > 400°C, although it completely evaporated before 370°C under general heating conditions in a coke oven (3°C/min and 0.1 MPa). The effects of increasing the pressure from 0.1 to 1.0 MPa at 3°C/min corresponded to the effects of increasing the heating rate from 3 to 10°C/min at 0.1 MPa under our experimental conditions. Rapid heating overcame the evaporation rate of the compound, whereas high pressure increased the bp and suppressed the pyrolysis reaction. At temperatures > 370°C, the remaining low-Mw compounds can act as a mobile phase and hydrogen donors, thereby increasing the fluidity of coal. The compounds can also serve as reactants in the coking reaction and increase the coke yield.
1H/13C-NMR spectra of 1, 1H/13C-NMR spectra of 2, MALDI TOF MS spectrum of 2, DSC profile of 2, and pyrolysis GC-MS chromatogram of 2.
This material is available on the Journal website at https://doi.org/10.2355/isijinternational.ISIJINT-2022-512.
This work has been done in the Research Group of “Cokemaking technology for low CO2 emission and high quality while extending available resources” in ISIJ. The authors are grateful to all the research group members. The authors also thank to Prof. Jeffrey M. Stryker, Dr. Robin J. Hamilton, and Dr. David E. Scott of University of Alberta for their teaches of synthesis methods, and to Dr. Sinya Sato, Dr. Kazumasa Yazu, and Mr. Toshihiro Kakinuma of AIST for their measurements.
