2018 年 58 巻 4 号 p. 660-666
To investigate the dominant factors that affect carbonization process of coal, behaviors of caking and low-quality coals (i.e., non- or slightly caking coal) during carbonization were examined. The release of volatile matter of pulverized coal samples in carbonization was evaluated using thermogravimetric analyzer, and the fluidity of coal particles was measured by the Gieseler plastometry method. Furthermore, a single coal particle was heated under a nitrogen atmosphere, and the images of the samples were acquired. From the images, swelling onset temperature, maximum swelling temperature, and maximum swelling ratio were evaluated. The carbonized coal particles were imaged using X-ray Computed Tomography (CT), and their internal structure was investigated. Although the release behavior of volatile matter, fluidity, and swelling of each coal were different according to kinds of coal, many of the parameters associated with those behaviors would correspond to coal rank. Combined with the carbonization behavior of coal and pore structure of carbonized coal, the amounts of thermoplastic components and volatile matter may affect the softening and swelling of coal particles.
The strength of metallurgical coke is a key feature of coke in a blast furnace. Coke produced from caking coal is a porous material, and the strength of the coke is known to be determined by the pore structure rather than the hardness of coke matrix.1) In general, coal as a raw material is supplied into a coke oven, and coke is obtained by carbonization of the packed bed of the coals. In carbonization process, the coal particles are softened and resolidified by heating, and then coke is formed. A good understanding of behaviors of coal during its carbonization process is important because the pore structure of coke is formed during the carbonization. Many studies of coke have been conducted to correlate coke strength with pore structure. In the previous study, although those studies investigated characteristic of microscopic structures and origins of the fracture of coke which affect coke strength,2,3,4) their objects were fabricated coke, and little is known about dominant factors in the formation of pore structure during coal carbonization.5)
As a study on a packed bed of coal during carbonization, Hays et al.6) carbonized a packed bed of coal in a furnace, in which one side of wall was heated, and samples were obtained with temperature gradient inside them. And then, the samples were embedded in resin and polished, and the formation of pore structure in the packed bed was observed. They reported that the pores coalesced after pores were generated in the coal particle and concluded that open pores in semi-coke after resolidification were generated by coalescence of pores. Furthermore, Hayashizaki et al.7) produced cylindrical samples with temperature gradient inside them as with Hays et al.6) and observed changes in pore structure by using X-ray computed tomography (CT). In the cross section where temperature is 420°C, coal particles existed individually, whereas in the cross section with temperature of 430°C, pores in coal particles existed and some coal particles adhered to other particles. In the cross section with higher temperature, pores developed and coal particles could not be distinguished. Hays et al. and Hayashizaki et al. have successfully visualized pore formation in the packed bed of coal. However, the behavior of the packed bed of coal in carbonization resulted from the development of pores in each coal particle and swelling of the particles, and it is unknown how the swelling behavior relates the carbonization behavior of coal. Furthermore, Steel et al.8) investigated effects of changes in properties with softening of coal on pore structure formation of pellet samples by using rheometry and micro-CT analysis. They indicated that pores in samples coalesced with an increase of temperature, and the porosity of the sample increased. Unfortunately, since they studied with a packed bed of coal, the behavior of each particle would not be understood because that of the packed bed was complex.
In terms of a single coal particle, Takizawa et al.9) observed swelling behaviors of a single coal particle for four kinds of coals including caking and low-quality coals and reported that each behavior was different according to kinds of coal. And then, the swelling behaviors were numerically represented. Toishi et al.10) measured changes in swelling ratio and pore diameter of single particles of two kinds of caking coals and indicated that coal particles were swelled by development of pores. However, they did not experimentally investigate factors that affect pore development in coal particles. Besides, because they used only caking coals, they cannot investigate pore formation in cokes produced from different kinds of coals. Moreover, Arima and Nomura11) concluded that with respect to fluidity and swelling ratio of coal, there is no definite correlation between those indexes because those test methods were completely different. For this reason, fluidity and swelling of coal during carbonization were not discussed well enough. In the present study, the dominant factors that affect carbonization behavior were discussed using three kinds of caking coals and one kind of low-quality one. To evaluate releasing rates of volatile matter, thermogravimetric analyses were conducted for pulverized coal samples. The Gieseler plastometer method was used for evaluation of fluidity of coal particles. Furthermore, a single coal particle was carbonized, and its swelling behavior was observed. X-ray CT imaging was conducted for carbonized coal samples to evaluate their internal structure, and the carbonization behavior of coal was investigated in relation to the structure.
Coal A, Coal B, and Coal C (caking coal) and Coal D (low-quality coal, i.e., non- or slightly-caking coal) were used as raw materials. The proximate, ultimate, and petrographic analyses of the coal samples are listed in Table 1. In any sample, the amount of ash was almost the same, and that of volatile matter (VM) increased on the order of Coal D, Coal A, Coal B, and Coal C. Reflectance in oil (Ro) represents coal rank, and the index number of Coal D was the smallest, which was followed by Coal A, Coal B, and Coal C.
| Coal | Proximate analysis [daf wt%] | Ultimate analysis [daf wt%] | Petrographic analysis [%] | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Ash | VM | C | H | N | S | Odiff | TI | Ro | |
| A | 9.1 | 26.5 | 89.9 | 5.29 | 1.48 | 1.09 | 2.24 | 13.4 | 1.17 |
| B | 9.4 | 23.1 | 89.9 | 5.14 | 1.92 | 0.52 | 2.52 | 31.6 | 1.27 |
| C | 8.8 | 18.9 | 90.4 | 4.80 | 2.07 | 0.57 | 2.16 | 25.8 | 1.56 |
| D | 9.3 | 37.7 | 84.7 | 5.75 | 1.78 | 0.71 | 7.06 | 16.2 | 0.80 |
where VM is volatile matter, TI is total inert, and Ro is reflectance in oil.
To investigate release behavior of volatile matter during carbonization, weight loss of samples was measured by a thermo gravimeter (STA449 F1 Jupiter, NETZSCH). The coal samples were crushed to below 212 μm in diameter. Approximately 15 mg of coal sample corresponding to a particle with 3 mm in diameter was held at a temperature of 25°C for 15 min under a nitrogen atmosphere, heated up to 105°C with a heating rate of 10°C/min, and the temperature was kept for 60 min. And then, the sample was heated up to 1000°C with a heating rate of 3°C/min, and the temperature was kept for 60 min. Release rates of volatile matter were calculated from weight loss measurements. Note that the measurements were performed at three times, and the reproducibility was confirmed.
2.3. Gieseler TestTo investigate fluidity of coal particles, a fluidity test was conducted by the Gieseler plastometer method defined by JIS M8801. Coal samples were crushed to below 414 μm in diameter, and 5 g of the crushed one was heated with a heating rate of 3°C/min from room temperature.
2.4. Evaluation of Swelling of a Single Coal ParticleA coal particle with a diameter of approximately 3 mm was used as sample, and the single coal particle swelled in the same way of Toishi et al.10) The coal particle was placed on a quartz rest which was inserted into a quartz tube and heated in a nitrogen atmosphere using an infrared gold image furnace (RHL-P610, SHINKU-RIKO Inc.). Here, to control the temperature of the coal particle, an R-type thermocouple through a graphite block (2 mm×2 mm×2 mm) was heated beside the coal sample because infrared emissivity of graphite is almost the same as that of coal. To dry the coal particle in accordance with the industrial standard method (JIS M 8812), the sample was heated up to 105°C with a heating rate of 10°C/min in nitrogen atmosphere, and the temperature was kept for 60 min. After that, the sample was heated up to 700°C with a heating rate of 3°C/min. All the time, the sample was imaged at intervals of 20 s using a digital camera (CX-5, Ricoh Co. Ltd.). After the temperature reached 700°C, the heating was stopped, and the sample was quenched by the large flow of nitrogen with a flow rate of 100 L/min. The projection area of the sample in the image was measured using image analysis software (WinROOF ver. 6.1, Mitani Corp.). Swelling ratios were calculated by the following Eq. (1), and swelling onset temperature, maximum swelling temperature, and maximum swelling ratio were evaluated from the images.
| (1) |
The coal particle after carbonization was imaged by using microfocus X-ray computed tomography (CT) system (Scan Xmate D160TS110, Comscantecno Co. Ltd.), with the tube voltage of 100 kV and tube current of 80 μA. The display resolution was 12 μm/pixel. The CT images obtained were visualized and observed by using image analysis software (Exfact VR, Nihon Visual Science Inc.). The element of the image was classified into pores and coke matrix by the discriminant analysis method.
Weight loss curves of each coal are shown in Fig. 1, and temperatures of the maximum release rate of volatile matter are listed in Table 2. The temperature of the maximum weight loss rate of Coal D was the lowest, which was followed by Coal A, Coal B, and Coal C. Here, it is known that the temperature of maximum release rate of volatile matter increased with an increase of carbon content in coal, i.e., coal rank.12) With respect to coal rank (Ro) shown in Table 1, the index increased on the order of Coal D, Coal A, Coal B, and Coal C. Thus, the order of coal rank corresponded to that of temperature of the maximum release rate of volatile matter in our experiment. The molecular structure of high-rank coal is aromatized, and the amount of the components decomposed at low temperature such as chain hydrocarbon in coal would be small. Consequently, the onset temperature of weight loss would increase for higher rank coal.
Weight loss curve of each coal.
| Coal | Temperature of maximum DTG [°C] | Maximum DTG [wt%/min] |
|---|---|---|
| A | 461 | 0.626 |
| B | 469 | 0.534 |
| C | 481 | 0.367 |
| D | 433 | 0.929 |
where DTG is rate of weight loss.
Results obtained by Gieseler plastometer are shown in Fig. 2. Initial softening temperature, maximum fluidity temperature, resolidification temperature, and maximum fluidity obtained from the fluidity test are listed in Table 3. With regard to the initial softening temperature, the temperature of Coal A was the lowest, which was followed by Coal D, Coal B, and Coal C. The maximum fluidity temperature increased on the order of Coal D, Coal A, Coal B, and Coal C, and the resolidification temperature increased on the order of Coal D, Coal B, Coal A, and Coal C. Focusing on the maximum fluidity, the fluidity index of Coal A was maximal, those of Coal B and Coal D were almost the same, and that of Coal C was minimal. These facts indicated that the order of the maximum fluidity temperature of each coal was the same as that of coal rank (i.e., Ro). Among caking coals, because the initial softening temperature increased and the maximum fluidity decreased with an increase in Ro, only the order of the initial softening temperature and maximum fluidity of each coal corresponded to that of coal rank. Therefore, these parameters related to coal rank, and this fact was consistent with the previous report of Kidena et al.13) Further, Takanohashi et al.14) described that softening behavior of coal is known in connection with coal rank, and the behavior was dominated by structural factors of coal. Miura explained that changes of coal particles during heating were considered as a process that forms coke matrix due to fusing and cohering of a mixture of active and inactive components.15) From above-mentioned facts, in the thermoplastic state, the fluidity of coal can be shown by thermoplastic components that can soften in coal. Focusing on low-quality coal, i.e., Coal D, some of the fluidity parameters were different from the order of coal rank. Kidena et al.16) concluded that coal with a high content of aromatic cluster had fluidity, and aliphatic compounds decreased fluidity of coal. They also reported that with regard to low-rank coal, the larger ratio of aliphatic components to aromatic compounds in metaplast led to less fluidity. Therefore, the behavior of low-quality coal can be different from that of three caking coals.
Gieseler fluidity of each coal.
| Coal | Initial softening temperature [°C] | Maximum fluidity temperature [°C] | Resolidification temperature [°C] | MF [log ddpm] |
|---|---|---|---|---|
| A | 384 | 456 | 495 | 3.96 |
| B | 416 | 464 | 494 | 2.46 |
| C | 437 | 477 | 510 | 1.26 |
| D | 396 | 434 | 462 | 2.21 |
where MF is maximum fluidity.
The appearance of Coal A, Coal B, Coal C, and Coal D at temperatures of 300°C (i.e., before swelling) and maximum swelling temperature are shown in Fig. 3. Diameters of coal sample at the temperature of 300°C were approximately 3 mm, and then, these coal particles were swelled by heating. Although the shape of coal particle before swelling was angular, that was rounded after the swelling. Pores in coal particle can be developed and push the wall of the particle because the particle did not burst, which means that open pores did not occur. In particular, the present results of swelling behavior of caking coals (Coal A, Coal B, and Coal C) were consistent with the results of Toishi et al.10) Figure 4 shows swelling onset temperature, maximum swelling temperature, and maximum swelling ratio. The swelling onset temperature of Coal D was the lowest, and that of Coal A was slightly higher than that of Coal D. Following this, the temperature increased on the order of Coal B and Coal C. Because the order of swelling onset temperature corresponded to that of coal rank (Ro), swelling onset temperature of single coal particle would be related to coal rank of coal. Focusing on maximum swelling temperature, the temperature increased on the order of Coal D, Coal A, Coal B, and Coal C as well as the swelling onset temperature of coal. Because the maximum swelling temperatures of Coal A and Coal B were almost the same, even coals of the different kinds can have similar swelling behavior. Furthermore, compared with the maximum swelling ratio of each coal, the sample with the highest ratio was Coal A, followed by Coal B, Coal C, and Coal D. It is reasonable that the ratio of Coal D was low because Coal D is low-quality coal and does not swell well. With regard to the four kinds of coal, the order of the maximum swelling ratio did not correspond to that of coal rank. But, with respect to caking coals (Coal A, Coal B, and Coal C), the order of the ratio corresponded to that of coal rank. These facts suggested that parameter of swelling behavior of single coal particle was also related to coal rank.
Images of coal at temperature of 300°C (i) and at maximum swelling temperatures (ii). (Online version in color.)
Swelling behavior of single coal particle. Each coal was measured three times, and the plots are averaged values for each sample. Error bars represent maximum and minimum values. (Online version in color.)
From the above, the order of parameters of thermal decomposition characteristics and fluidity was the same as that of coal rank except as to Coal D, and that of swelling behavior of single coal particle was also the same as that of coal rank. Therefore, when low-rank coal such as low-quality coal was used, those parameters cannot be evaluated only by coal rank. However, when caking coal (i.e., high-rank coal) was used, those parameters would be related to coal rank.
3.4. Behaviors of Coal CarbonizationThe release behavior of volatile matter, fluidity, and swelling behavior of each coal, shown in Fig. 5, are discussed in relation to the internal structure of carbonized coals, shown in Fig. 6. Focusing on Coal A, with a rise in temperature, the volatile matter of coal was released, the single coal particle swelled, and Gieseler fluidity was shown. Toishi et al.10) showed that there were almost no pores in a coal particle before swelling, whereas after swelling many pores were generated. They concluded that because the volatile matter of coal flowed into pores originally existing in the coal and/or newly generated pores during carbonization, the coal swelled. Furthermore, Dohi et al.17) investigated permeation behaviors of coal in glass beads layer, and they indicated that onset temperature of permeation of coal correlated initial softening temperature in Gieseler plastometer method. Consequently, after the initial softening temperature, coal particles were softened, and thermoplastic components can be softened. With regard to carbonized single particle of Coal A, as shown in Fig. 6(a), a large pore and few pore walls existed. These results suggested that the coal particle was softened by thermoplastic components, pores in the particle were developed by releasing of volatile matter, and then, almost all of pores coalesced. Therefore, the volatile matter was released by heating, and thermoplastic components in a coal particle began to be softened by the releasing of volatile matter. When a sufficient amount of thermoplastic components is available, the swelling and softening of particles would progress simultaneously.
Behaviors of each coal during carbonization. Swelling of single coal particle with size of 3 mm and weight of about 15 mg (presented by red-color bars), releasing volatile matter in coal particles with diameter of less than 212 μm and weight of 15 mg (indicated by brown-color line), and softening of coal particles with size of 414 μm and weight of 5 g (presented by blue-color line). (Online version in color.)
X-ray CT images. The samples were carbonized at 700°C: (left) two-dimensional images; (right) three-dimensional images.
On the other hand, for Coal D which is low-quality coal, although swelling of a single coal particle was started and Gieseler fluidity was shown due to releasing of volatile matter as with Coal A, the swelling behavior of Coal D was different from that of Coal A. In particular, the temperature of maximum release rate of volatile matter of Coal D was lower than that of Coal A. Furthermore, the temperature range of maximum swelling temperature was not overlapped with the range which showed Gieseler fluidity. With regard to pores in coke carbonized, pore walls of coke were thick, and large and small pores existed, shown in Fig. 6(d). When the pore walls were observed in detail, independent pores existed inside of the wall (Fig. 7). In the case of Coal D, although thermoplastic components were softened and swelling of coal was started by releasing of volatile matter as with Coal A, the amount of thermoplastic components would not be sufficient because volatile matter was released at low temperature, and the volatile matter would little contribute to swelling of coal even though the volatile matter flows in pores. Thus, the maximum swelling ratio of Coal D was the lowest because the amount of thermoplastic component was not sufficient. From the above facts, factors that affect carbonization process are different in caking coal and low-quality coal. Therefore, we will concentrate on caking coal.
Three-dimensional transmission images focusing on pore wall of Coal D. The sample was carbonized at 700°C. (Online version in color.)
For Coal B, as is the case with Coal A, the volatile matter was released, the single coal particle swelled, and Gieseler fluidity was shown (Fig. 5(b)). With regard to pore structure of carbonized sample of Coal B shown in Fig. 6(b), pore walls were thin, pores were connected with each other. Although large pores, as in coke of Coal A, did not exist, the size of pores in coke of Coal B was larger than that of Coal C. Because the release behavior of volatile matter in Coal B was similar to that in Coal A, Coal B would be softened and swell as with Coal A. However, compared with Coal A and Coal B, the maximum release rate of volatile matter, maximum fluidity, and maximum swelling ratio of Coal B were lower than those of Coal A. This is because the amounts of thermoplastic components and released volatile matter of Coal B were smaller than those of Coal A. As shown in Fig. 6(b), in the case of Coal B, pores separated by several walls existed, and when the amount of thermoplastic components was not sufficient, pores would not coalesce.
Although carbonization behaviors of Coal A and Coal B were almost the same, that of coal C was different from those of Coal A and Coal B in spite of caking coal, shown in Fig. 5(c). In particular, the Gieseler fluidity was shown at temperature below the swelling onset temperature of Coal C. Furthermore, Fig. 6(c) indicated that there were many small pores in coke of Coal C, and the pore structure of Coal C was different from those of Coal A and Coal B. The temperature of the maximum release rate of volatile matter of Coal C was higher than those of Coal A and Coal B, and the temperature which showed Gieseler fluidity was also higher. Therefore, although thermoplastic components were softened at high temperature, the amount of the components would be small. When there is not a sufficient amount of thermoplastic components, the volatile matter would not flow into pores and little contribute to swelling of coal.
From above all results, the amount of thermoplastic components and volatile matter would affect softening and swelling of coal particles.
In the present study, carbonization behaviors of three kinds of caking coals and one kind of low-quality one were investigated. Thermogravimetric analyses and fluidity tests for coal samples were conducted, swelling behavior of a single coal particle was observed, and the internal structure of the carbonized coal particles was imaged with X-ray computed tomography (CT). The orders of the temperature of the maximum release rate of volatile matter, the temperature of the maximum fluidity of each coal, and the swelling onset temperature of a single coal particle corresponded to that of coal rank. Furthermore, the order of the maximum swelling ratio of coal particles was the same as that of coal rank except as to low-rank coal. Therefore, factors for the releasing behavior of volatile matter, fluidity, and swelling of coal can be correlated with coal rank. With regard to the release of volatile matter, fluidity and swelling of coal particles, and pore structure of the carbonized coal particle, when the temperature of the release rate of volatile matter matches the temperature in which the sufficient amount of thermoplastic components existed, the swelling and softening of coal particles would progress simultaneously. Hence, large pores can be formed. However, when there is not a sufficient amount of thermoplastic components, large pores in the coal cannot be formed because pores would not coalesce. On the other hand, when the temperature of the maximum release rate of volatile matter was higher or lower than the temperature in which the sufficient amount of thermoplastic components existed, the release of volatile matter would have little contribution to swelling of coal. From above-mentioned, the amounts of thermoplastic components and volatile matter can be thought to affect the softening and swelling of coal particles.
