Preparation of Formed Coke from Biomass by Sequence of Torrefaction, Binderless Hot Briquetting and Carbonization

Abstract

This paper proposes a method of preparing high-strength formed coke from woody biomass without binder. Chipped and pre-dried Japanese cedar was heat-treated in an inert atmosphere (i.e., torrefied) at 225–325°C (T_t), pulverized to sizes in three different ranges, molded into briquettes (in the form of thick disk with diameter/thickness ≈ 2.5) at temperature up to 200°C by applying mechanical pressure of 128 MPa. The torrefied/briquetted cedar (TBC) was then converted into coke by heating to 1000°C in an inert atmosphere at normal pressure. This process sequence enabled to prepare coke having indirect tensile strength (S_t) of 8–32 MPa, which was much higher than that without torrefaction, below 5 MPa. The torrefaction greatly improved pulverizability of the cedar, which was further promoted by increasing T_t. S_t of TBC and that of coke both increased as the particle sizes of TBC decreased, but this explained only a minor part of significant effect of T_t on S_t of the coke. S_t was maximized at T_t = 275°C regardless of the degree of pulverization. The T_t effects on physicochemical properties of TBC and coke were investigated in detail. The difference in S_t of coke by T_t was mainly due to that in the increment of S_t along the carbonization at 500–1000°C. Fracture surfaces of the coke had particular morphologies that had been inherited from the original honeycomb structure of the cedar.

1. Introduction

Production of blast furnace coke from biomass is a potential technology in the future ironmaking with minimized or no emission of coal-derived CO₂. So far, not few studies have been studied on this technology, as outlined below, under a common assumption of replacement of a limited portion of the current feedstock, i.e., coal, by biomass. Matsumura et al.¹⁾ added woody biomass to a slightly caking coal with addition rates of 0.5–2.0 mass% and carbonized mixtures at 1000°C. The I-type strength of coke from the mixture was equivalent to that from the coal alone in the case that the biomass had been molded by hot compression at 200°C. But the coke strength decreased by adding the biomass without densification prior to mixing.

Montiano et al.²⁾ investigated effects of addition of chestnut or pine sawdust to general coal blends on properties of coke from carbonization at 1100°C, and obtained results similar to that of the previous study.¹⁾ They confirmed reduction of apparent density and strength by adding the sawdust, and attributed this mainly to loss of fluidity by adding the sawdust at a rate of 0.75–3.0 mass%. Diez et al.³⁾ studied effects of adding woody biomass (sawdust), its particular components (cellulose, xylan and lignin) and charcoal to a coking coal (logarithm of maximum fluidity (logF_max) ≈ 2.5). They showed that logF_max decreased greatly to 0.5 by adding cellulose or sawdust at a rate of only 1–3 mass%, while slightly to 2.1 with 6–9 mass% addition of charcoal, lignin or xylan. These effects strongly suggested that the loss of fluidity was caused by a mismatch in the content of volatile matter or temperature range for its release. Such loss of the fluidity was also reported by Díaz et al.⁴⁾ They found that heating rate as high as 180°C/min minimized the fluidity loss even at a biomass addition rate of 5 mass%, mainly due to promoted fluidity by rapid heating. Ueki et al.⁵⁾ reported significant reduction of coke strength (on a tumbler strength index), by 5 mass% addition of sawdust (particle sizes < 0.15 mm), but also effectiveness of its semi-carbonization at 400–500°C, which minimized the loss of strength. These results are in broad agreement with those by Diez et al.²⁾

Very recent world-wide strategies of ‘decarbonization’ with zero emission of fossil-fuel-derived CO₂ from industries/societies by 2050–2060 may strongly restrict the use coal but encourage that of biomass as the main carbonaceous feedstock of coke as well as the main alternative to PCI fuel. Production of formed coke solely from biomass was studied by Kudo et al.,⁷⁾ who pretreated woody and herbaceous biomasses in hot-compressed water (HCW) at 200–300°C, dried, briquetted by hot-compression molding at 130–200°C, and then carbonized at 900°C. The resulting coke had a tensile strength of 20–50 MPa, which was several to ten times that of commercially available blast furnace coke. Such high strength of coke was attributed to high thermal plasticizability in the briquetting and no thermal swelling/fluidity in the subsequent carbonization, which gave highly densified briquette and high strength of coke, respectively. The implementation of the HCW treatment of biomass may be, however, difficult due to high capital and operational costs, unless the HCW treatment and subsequent processes produce high added-value products with substantial yields.

Production of formed coke from biomass is potentially applicable if high-degree densification of biomass (or pretreated biomass) prior to carbonization with modestly suppressed volatile matter release is realized. Recent studies on production of formed coke from coal^8,9,10) or lignite^11,12,13) have demonstrated that a sequence of pulverization, hot- or cold briquetting and conventional carbonization can successfully produce formed coke with a tensile of 10–40 MPa from non-caking and slightly calking coals ranging from lignite to low-volatile bituminous ranks. Matoba et al.⁹⁾ and Uchida et al.¹⁰⁾ showed importance of pulverization of coal to sizes smaller than ca.100 μm for producing high-strength formed coke from the above-mentioned sequence. Its application to biomass encounters well-known difficulty of pulverization.¹⁴⁾ A most reasonable way to solve this problem is heat treatment that allows mild pyrolysis, i.e., torrefaction. Yu et al.¹⁵⁾ and Wang et al.¹⁶⁾ showed great improvement in pulverizability of biomass in terms of time or energy consumption by torrefaction at temperatures of 220–325°C. However, the torrefaction could have a disadvantage, that is, loss of plasticizability of biomass that is necessary for the densification in the hot briquetting.

The three major components of biomass, i.e., cellulose (but amorphous), lignin and hemicellulose (xylan) undergo glass transition at 160–180°C,¹⁷⁾ 110–150°C,¹⁸⁾ and 120–130°C,¹⁹⁾ respectively. It is thus expected that hot compression, if operated at a temperature higher than those temperatures, successfully densifies biomass more extensively than at lower temperature. The torrefaction causes not only covalent bond breaking²⁰⁾ leading to depolymerization but also cross-linking²¹⁾ that inhibits plasticization. It seems that these chemical reactions result in net loss of plasticizability, which is supported well by the improved pulverizability^15,16) due to physical embrittlement. Thus, the pulverizability and briquettability may be in a trade-off relationship. Torrefaction under optimized conditions potentially enables to convert biomass into high-strength coke by the pulverization, hot-briquetting and carbonization. There have so far been no studies on formed coke production from torrefied biomass alone.

The present authors studied production of formed coke from woody biomass by applying sequential torrefaction, pulverization, binderless hot briquetting and carbonization. This paper reports experimental data demonstrating preparation of high-strength coke, and also discusses the mechanisms for emergence of optimum torrefaction conditions and occurrence of coke strength.

2. Materials and Methods

2.1. Biomass Sample

Sawdust of a Japanese cedar (SD) with particle sizes smaller than 3.0 mm was supplied from The Iron and Steel Institute of Japan. SD was pulverized to sizes smaller than 100 μm in a cutter mill (Wonder Blender WB-1, Osaka Chemical) for a short period of ca. 1 min. The cutter milling of SD was necessary to prepare coke samples without torrefaction.

2.2. Torrefaction and Pulverization

A prescribed amount of SD (1.0 g) was heated in atmospheric flow of high-purity N₂ (flow rate; 300 mL-stp/min, purity; > 99.9999 vol%) in a horizontal tubular reactor, applying heating rate, peak temperature and holding time at the peak temperature of 5°C/min, 225–325°C and 60 min, respectively. The temperature for the torrefaction will be denoted by T_t. The gaseous products ranging from H₂ up to C₄ hydrocarbons were collected in a gasbag, while the liquid products (pyrolytic water, light oil and heavy oil) on an aerosol filter (150°C) or in cold traps (0°C, −70°C) connected to the reactor in series. The entire portion of the liquid product (except water) was defined as tar. Details of the product collection, quantification and analysis were reported elsewhere.²²⁾ The solid left in the reactor, i.e., the torrefied SD was then recovered. Hereafter, the torrefied SD with T_t = x°C is referred to as Tx (x = 225, 250, 275, 300 or 325). The yield, atomic composition and higher heating value (HHV) of SD and TSD are shown in Table 1.

Table 1. Yield and properties of TSD.

Tt, °C	yield, kg/kg-dry-SD	H/C atomic ratio	O/C atomic ratio	HHV, MJ/kg-dry
SD	100	1.45	0.68	19.7
225	91.2	1.29	0.59	20.4
250	87.3	1.29	0.58	20.6
275	74.2	1.18	0.53	21.1
300	55.2	0.92	0.34	23.5
325	42.2	0.86	0.28	24.6

The pulverization of TSD was performed in two steps. The first step was cutter milling in the same way as mentioned in 2.1. The particle sizes of the cutter milled TSD were smaller than 100 μm, as shown later. The cutter-milled TSD was further pulverized by gentle ball milling for 10 or 20 h. Details of the ball milling were reported previously.⁹⁾ The ball milling further decreased the particle sizes to smaller than 40 or 20 μm.

2.3. Briquetting by Hot-compression Molding

The pulverized SD and TSD were molded into a circular plate by hot-compression employing a mold with 15 mm inner diameter and 20 mm depth, and applying the following conditions: mass of SD or TSD; 1.0 g-dry, temperature (T_m); 40–200°C, mechanical pressure (P_m); 128 MPa, period of mechanical pressure loading; 8 min. These conditions followed the authors’ previous studies.^9,10) No binder was used in the molding. The dimensions (diameter and thickness) of each circular plate were measured. In the cases of T_m = 200°C, the diameter was within a range of 14.0±0.1 mm, independent of both T_t and degree of pulverization, but these influenced the thickness within another range of 5.1–6.9 mm.

2.4. Carbonization

Four to eight specimens of TSD briquettes, which had been prepared from the same TSD and under the same conditions, were heated together in atmospheric flow of N₂ (flow rate; 300 ml-stp/min, purity; > 99.9999 vol%) at a heating rate of 5°C/min to the peak temperature (T_c) of 1000°C with a 10 min holding time at T_c. The carbonization was followed by natural cooling of coke to ambient temperature in the same N₂ flow as above, and its recovery. Cutter-milled SD and TSD were also carbonized under the same conditions as above, but without briquetting.

2.5. Characterization

Coke, TSD briquettes and SD briquettes were subjected to strength tests for measuring indirect tensile strengths (S_t). Details of the measurement were reported previously.^9,10) Four to twelve specimens were tested for each coke or briquette. It was confirmed that the deviation of the measured S_t was within ±10% from the average. The fracture surfaces of cokes and briquettes were observed by scanning electron microscopy with a JEOL model of JSM-IT700HR. The particle size and distributions of TSD were measured with an analyzer (Malvern Panalytical, Morphologi 4). In each analysis, 55000–74000 particles were observed with a microscope, and the properties of the individual particles such as circle-equivalent (CE) diameter, solidity, aspect ratio and circularity were determined. The porosities of coke samples were measured by a mercury porosimetry over a range of pore size of 0.1 to 100 μm.⁹⁾

Cutter milled SD and TSD were analyzed by thermogravimetry in a thermobalance (Hitachi High-Technologies, model STA7200) employing initial sample mass, heating rate, peak temperature, carrier gas and its flow rate of around 2 mg, 5°C/min, 830°C, atmospheric N₂ (purity > 99.9999 vol%) and 200 mL-stp/min, respectively. The rate of mass release was analyzed in order to estimate the chemical composition of SD or TSD. The analysis was performed assuming the following according to previous reports^23,24,25) (i) SD consists of three primary components, i.e., cellulose, hemicellulose and lignin. (ii) Each primary component is converted into volatiles and the secondary component (common among the three components), following first-order kinetics with respect to the mass of the unconverted portion. (iii) Each primary component behaves as a single reacting component with a single set of Arrhenius parameters (pre-exponential factor and activation energy). An example of the rate analysis is shown in Fig. S1. The assumptions (i)–(iii) were successful, because the mass release from SD at temperature up to 500°C was expressed quantitatively as the sum of the conversions of the three components into volatiles by optimizing the individual kinetic parameters. The mass release at >500°C is due to that from the secondary component formed by polymerization of the primary components. Thus, the kinetic analysis enabled to estimate the mass fractions of the primary and secondary components as those of the volatile matter. In the case of SD, the fractions of cellulose, hemicellulose, lignin and the secondary components were 35.1, 27.2, 15.3 and 4.7 mass%-dry, respectively, while that of the solid residue (char) was 17.8 mass%. The fraction of char consisted of those from the primary and secondary components. The compositions of TSD and optimized kinetic parameters are available in Tables S1 and S2, respectively.

3. Results and Discussion

3.1. Pulverizability of TSD

Figures 1(a) and 1(b) show the effects of T_t on the distribution of particle size (represented by CE diameter) for TSDs after the cutter milling and 10-h ball milling. It is seen that increasing T_t shifts both distributions toward smaller CE diameter side. This is also confirmed in Fig. 2(a) that demonstrates monotonous decreases in average CE diameter with increasing T_t. Figure 2(b) reveals that the solidity, circularity and aspect ratio all increase toward unity (theoretical maximum) as T_t increases. The increasing aspect ratio and solidity result from more extensive disintegration of particles and grinding of their surfaces, respectively, and are both consistent with increasing circularity. The data shown in Figs. 1 and 2 thus demonstrate an improvement of pulverizability by increasing T_t. The improvement in pulverizability of SD is remarkable because SD was hardly pulverized by the ball milling.

Fig. 1.

Particle size distributions of TSDs with different T_t of 225–325°C. (a) after cutter milling (CM), (b) after 10-h ball milling (BM10).

Fig. 2.

Effects of T_t on (a) average particle size as CE diameter of TSD after cutter milling (CM), 10-h ball milling (BM10) and 20-h ball milling (BM20), and (b) Shape factors for TSD after CM.

Another important feature of torrefaction was found. It was spontaneous pulverization of SD during or after the torrefaction, which was induced thermochemically. Figure 3 exhibits typical SEM photographs of TSDs with T_t = 225, 250, 275 and 300°C. The torrefaction with higher T_t results in smaller TSD particles^23,24,25) even before the intentional pulverization.

Fig. 3.

Spontaneous pulverization of SD particles by torrefaction without milling.

3.2. Strength of Briquette

Figure 4 shows the effect of T_t on S_t of briquette. S_t is hardly influenced by T_t at 225–275°C, but decreased by that at 300–325°C. This result is discussed by focusing on the size and thermomechanical properties of pulverized TSD before the briquetting. Figure 4 also shows that a smaller particle size is favored for greater S_t. This trend is well known in molding of particles to densified bulk material. In fact, for each TSD, the apparent density of briquette increases with decreasing particle size (Fig. 5(a)). Figure 5 expresses that the apparent densities of T300 and T325 briquettes depend more strongly on the particle size than those of T225–T275. Such dependency is probably related to the plasticity of TSD in the hot compression. It was believed that SD lost much of its original plasticity during the torrefaction at 300–350°C but less at 225–275°C. The remaining plasticity of particles allowed their deformation, inducing particles’ contact and boding, leading to more degree of densification. The size was thus less important for particles with more plasticity. Figure 5(b) plots the logarithmic S_t of T225–T325 briquettes against average particle size. The dependencies of S_t of the briquettes on the particle size are similar to one another with respect to the slope of the lines drawn in the figure. Decrease in St with increasing particle size is consistent with the trends seen in Fig. 5(a).

Fig. 4.

Effects of T_t on S_t of briquette from STD pulverized by cutter milling (CM), ball milling for 10 h (BM10) or that for 20 h (BM20) before hot-compression molding with T_m = 200°C.

Fig. 5.

Effects of CE diameter on apparent density (a) and S_t (b) of TSD briquettes. Note that the vertical axis of Fig. 5(b) is logarithmic.

The above-discussed particle size effect is important for S_t of briquette. It is, however, not reasonable to attribute the variation with T_t of briquette’s S_t only to the particle size effect. As mentioned previously, it was believed that a more or less portion of the original plasticity had been left in T225, T250 and T275, and that the plasticity helped particles’ contact/bonding and densification of briquette. The densification of T300 and T325 was not successful due to their less plasticity than T225–T275. The data shown in Figs. 4–5 thus demonstrate that the torrefaction under appropriate conditions greatly improve the pulverizability of SD with no or little loss of briquettability (moldability). Though not shown in Figs. 4–5, S_t of the briquette from SD was as low as 3.2 MPa. The torrefaction and subsequent milling even improved the briquette strength, which was reported by none of the previous works.

The effect of T_t on the briquette S_t is considered from a chemical viewpoint. Figure 6 illustrates change in the chemical composition of SD by the torrefaction. The cellulose (CL) seems to be stable up to T_t = 275°C, while the hemicellulose (HC) is converted at T_t ≥ 250°C. The fraction of the lignin (LG) unexpectedly increases with T_t of 225–300°C probably due to conversion of HC into components that undergo pyrolysis following kinetics very similar to that of LG. It is noted that the total fraction of the CL, HC and LG (including HC-derived LG) decreases but very slightly up to T_t = 275°C. This is consistent with that the briquettes of T225, T250 and T275 had equivalent S_t. It is clear that the torrefaction at 300°C and that at 325°C respectively convert HC and CL completely, leaving LG, HC-derived LG and the secondary component. More importantly, the torrefaction at 300–325°C greatly increases the fraction of non-volatile matter, which is given by the difference between unity and the total fractions of CL, HC, LG and secondary component. This is clearly from the progress of polymerization and carbonization during the torrefaction. It is hence expected that the torrefaction at 300–325°C causes loss of the inherent plasticity of SD. The torrefaction at 300–325°C thus enhances the pulverizability but diminishing plasticity at the molding temperature (200°C). S_t of T300 and T325 briquettes and their chemical compositions indicate that negative effect of plasticity loss on the briquette S_t is more than the positive effect of enhanced pulverizability.

Fig. 6.

Chemical compositions of SD and TSDs with T_t = 225–325°C.

3.3. Strength of Coke from Carbonization with T_c = 1000°C

This section presents and discusses the effects of T_t on coke properties. The coke samples were prepared by the carbonization of briquettes that had been prepared with T_m = 200°C. This temperature was decided based on preliminary results for the effects of T_m on S_t and apparent density of coke. Figure 7 shows that S_t of coke increases greatly with T_m within the range of 40–200°C, demonstrating the effectiveness of employing higher T_m unless TSD undergoes pyrolysis. Such a positive effect of T_m on S_t is arisen mainly from greater apparent density of briquette with higher T_m. It is believed that TSD, regardless of T_t, is plasticized under the heating and mechanical pressure of 128 MPa.

Fig. 7.

Effects of T_m on (a) S_t of cokes from T225, T250, T275 and T300 and (b) apparent density of briquettes of T225, T250, T275 and T300. Conditions: pulverization; CM, T_c; 1000°C.

Figure 8 illustrates the effects of T_t on S_t of cokes from the cutter-milled and ball-milled TSDs, showing some important features of the sequential processes of torrefaction, pulverization, hot briquetting and carbonization. Firstly, the coke from the culler-milled TSD is greater than that from SD, 4.7 MPa, regardless of T_t. The torrefaction is thus necessary for not only improvement of pulverizability of SD but also production of high-strength formed coke with S_t greater than 5 MPa. Secondly, more degree of pulverization leads to coke with greater S_t. This is in broad agreement with existing knowledge in formed-coke production.^9,10) The maximum S_t of the coke from the 20-h ball-milled TSD is as high as 32 MPa. Thirdly, regardless of the degree of pulverization of TSDs, S_t is maximized at T_t = 275°C. The effect of T_t on S_t of the coke is clearly different from that of the briquette. There is no significant effect of T_t at 225–275°C on S_t of the briquette, as seen in Fig. 4. For each of the cokes from cutter-milled (CM), 10-h ball-milled (BM10) and 20-h ball-milled (BM20) TSDs, the S_t is maximized and minimized at 275°C and 325°C, respectively. The ratio of the maximum S_t (T_t = 275°C) to minimum S_t (T_t = 325°C) is 2.1, 2.0 and 2.8 for the cokes from CM, BM10 and BM20 TSDs, respectively.

Fig. 8.

Effects of T_t on S_t of cokes. Conditions: T_m; 200°C, T_c; 1000°C.

3.4. Physical Properties of Coke from Carbonization with T_c = 1000°C

The above-mentioned third feature is considered based on coke properties. Figure 9 plots the apparent density and porosity of coke against T_t. The density and porosity decrease and increase monotonously with T_t, respectively. As seen in Fig. 9(c), these properties are related linearly to each other. The straight dashed line drawn in this figure indicates a relationship if the density of coke, which is given by assuming that the density after elimination of the volume of macropores with sizes greater than 0.01 μm (in other words, based on the total volume of mesopores, micropores and non-porous carbonaceous matrix) is 1.43 g/cm³. The linearity suggests that the effects of T_t on the true density and micro-/meso-porosity of coke were, if any, insignificant.

Fig. 9.

Effects of T_t on (a) apparent density and (b) porosity of coke from the carbonization (T_c = 1000°C) of cutter-milled TSD, (c) relationship between porosity and apparent density of coke from cutter-milled TSD, and (d) apparent density of coke from the carbonization (T_c = 1000°C) of ball-milled TSD.

In general, S_t of coke decreases with a decrease in the apparent density or an increase in the porosity, and these changes with T_t may explain the magnitude relation of S_t amongst the cokes from T275, T300 and T325 qualitatively, but not at all for the cokes from T225, T250 and T275. The coke from T275 has S_t greater than those of T225 and T250, while smaller apparent density and greater porosity. Figure 9(d) shows that the apparent density of coke from the ball-milled TSD changes in somewhat different manners from that for the cutter-milled TSDs. Nevertheless, the effect of T_t on the apparent densities can hardly explain the trend of S_t at T_t = 225–275°C.

Figure 10(a) illustrates the pore size distributions of the cokes from cutter-milled TSDs. The pore size ranges from ca. 0.1 to 2 μm commonly for the cokes except for that from T325, of which pore size ranges ca. 0.1 to 3 μm. Mechanical strength of porous solid is generally affected negatively by the pore size. However, the magnitude relation of S_t among the cokes with different T_t cannot be explained by that of pore size distribution. As shown in Fig. 10(b), the porosity increases with the mean pore size. This trend cannot explain that S_t of the coke from T275 is greater than those from T225 and T250.

Fig. 10.

Porous properties of coke from TSD. (a) pore size distributions of cokes from the carbonization (T_c = 1000°C) of cutter-milled TSDs, (b) relationship between porosity and mean pore size at which cumulative pore volume is a half of the total pore volume.

Thus, the above-described consideration of the effects T_t on the coke properties is not successful in drawing the mechanism for maximization of S_t at T_t = 275°C. It is also difficult to attribute the difference in S_t of coke with different T_t to that of the density of the ‘non-macroporous’ part of coke because the density seems to be hardly influenced by T_t (see Fig. 9(c)). Figure 11 shows the net and gross coke yield on the bases of the mass of SD and TSD, respectively. The gross coke yield increases as T_t increases, as expected. It is noted that the net coke yield is almost independent of T_t as 225–300°C, which is primarily due to that more volatile matter yield in the torrefaction is compensated by less yield in the carbonization. The net coke yield from T325 is slightly lower than the others. This is because of the less extensive suppression of volatile matter release during the carbonization by the previous hot briquetting. As shown in Fig. 12, the hot briquetting suppresses the release of tar, while such an effect diminishes for T_t = 300–325°C.

Fig. 11.

Mass yield of coke from the carbonization (T_c = 1000°C) of TSD. (a) gross yield based on that of TSD, (b) net yield based on the mass of SD.

Fig. 12.

Gross coke yields from the carbonization (T_c = 1000°C) of TSD briquette and TSD powder (without briquetting) and their difference (a), and its relationship with tar yield from torrefaction.

T325 has underwent more degrees of pyrolytic conversion, in other words, tar evolution and cross-linking (according to the yields of pyrolytic water and inorganic gases such as CO₂ and CO) than the other TSDs. The yields of pyrolytic water and CO/CO₂ are shown in Fig. S2. T325 has the smallest ability of evolving tar during the carbonization. Figure 12 suggests that the tar evolution in the torrefaction, if more than 20 mass%-dry-SD, results in less positive effects of hot briquetting on the coke yield. According to the net coke yield (Fig. 11), there is no significant effect of T_t (225–275°C) on the chemical nature of coke.

3.5. Occurrence and Development of Coke Strength during Carbonization

The physical and chemical properties of coke and characteristics of carbonization, discussed in the previous section are both insufficient to explain the T_t effect on the coke S_t, in particular, those from T225, T250 and T275. In this section, changes in S_t of TSD along the carbonization over the range of T_c from 300 to 1000°C are shown and discussed for understanding how the effect of T_t on S_t of coke (T_c = 1000°C) occurs, and find out a main reason why T_t = 275°C leads to the maximum S_t.

Figure 13 plots S_t against T_c for T225–T300. The plots at T_c = 200°C indicate S_t of briquettes before the carbonization for expedience. S_ts of cokes from the four different TSDs, which have different S_ts before the carbonization, converge into around 5 MPa at T_c = 420°C, and then diverge at higher T_c to a range from 8.8 MPa (T225) to 16.9 MPa (T275). Thus, the effect of T_t on S_t diminishes in the early stage of carbonization at T_c up to 420°C, but grows at T_c = 500–1000°C, where the H/C and O/C atomic ratios of T275 coke decrease from 0.43 to 0.02 and 0.11 to 0.03, respectively (Fig. S3). The apparent density and S_t of the four cokes from T225–T300 were all within narrow ranges of 0.85– 0.87 g/cm³ and 4.8–5.1 MPa, respectively at T_c = 420°C. It was also found that the relative mass of coke at T_c = 1000°C to that at T_c = 420°C was within a very narrow range of 0.80–0.81regardless of T_t. S_t of coke at T_c = 1000°C was, nonetheless, influenced greatly by T_t. The effect of T_t on S_t of coke at T_c = 1000°C was thus attributed to the increment of S_t of coke during the carbonization at 500–1000°C. Thus, understanding the T_t effect on the coke S_t at T_c = 1000°C requires explanation of that on the increment of the coke S_t at T_c = 500–1000°C. Such explanation is, however, difficult within the range of coke properties that have so far been measured.

Fig. 13.

Changes in S_t with T_c in the carbonization of briquettes from cutter-milled TSD. The solid and dashed lines are drawn just for easier identification of the individual trends of S_t.

3.6. Further Discussion on Coke Properties

The discussion on the coke S_t, developed in the previous section, encounters difficulty in considering the mechanism of occurrence of the strength only based on the bulk properties of coke. It is believed that the three-dimensional structure of coke and its change along the carbonization are necessary for approaching the mechanism of the occurrence and development of coke strength. Figure 14 compares the fracture surfaces of T225 and T300 after carbonization at T_c = 420°C. Both surfaces were very similar to those of the corresponding briquettes before the carbonization. The fracture surface of T225 has a morphology that can be illustrated schematically in the pictures (a) and (b). The compressed honeycomb structures indicate flexibility and plasticity of T225 in the hot pressing. Similar morphologies were also observed on fracture surfaces of T250 and T275. On the other hand, that of T300 seems to consist of plates (as walls of honeycomb) that were probably formed by breakage of the honeycomb and stacking of plates during the hot pressing, as shown in the picture (c). This is consistent of the loss of plasticity in the torrefaction with T_t = 300–325°C, and also the lower S_t of T300 and T325 briquettes that those of T225–T275 briquettes.

Fig. 14.

Fracture surfaces of T225 and T300 after carbonization at T_c = 420°C.

The above-mentioned morphologies were inherited to cokes with T_c = 1000°C, as shown in Fig. 15 and also in Fig. S4. The compressed honeycomb structure is obvious on the fracture surface of the T225 coke, while such structure is hardly seen on that of the T300 coke. In a qualitative sense, the surfaces of T250 and T275 appear to be similar to those of T225 and T300, respectively. It is difficult to draw a main reason why the T275 coke has the greatest S_t, but a hypothesis occurs from the inheritance of the honeycomb structure. The honeycomb walls of T225 are flexible under the pressure in the hot pressing, and therefore compressed without their breakage. The compressed honeycomb structure is maintained during the carbonization, contributing to the coke strength. But meanwhile, such a structure hinders bonding of walls to those of other honeycombs during the carbonization because the walls are fixed in the honeycomb structure. On the other hand, the contribution of the honeycomb structure to the T300 coke strength is much smaller than that of T225. At the expense of this, plate-shaped walls are allowed to contact and to others, which leads to their bonding and occurrence of the coke strength. Thus, the breakage of the compressed honeycomb structure results in more frequent bonding of plates and an increase in the coke strength. In this hypothesis, the breakage of the honeycomb structure (during the hot pressing) causes both negative and positive effects on the coke strength. According to the effect of T_t on S_t of coke, the maximum S_t of the coke from T275 is qualitatively explained as follows. The breakage of honeycomb structure of T275 occurs more frequently than those of T225–T250, while bonding of walls occurs more frequently during the hot pressing. The S_t of T275 briquette is therefore as high as those of T225 and T250 briquettes. Thus, the honeycomb structure of T275 is ‘modified’ three-dimensionally with slight or no loss of S_t in the hot pressing. Such bonding of plates is less frequent for T300 and T325, due to more loss of plasticity than T275. In the subsequent carbonization (at T_c > 500°C) of T275, plate-to-plate bonding occurred obeying the mechanism similar to sintering, and its frequency is more than that in the carbonization of the T225 and T250 briquettes. S_t of T275 coke therefore increases with T_c more steeply than the T225 and T250 cokes. T300 undergoes such plate-to-plate bonding during the carbonization with a similar frequency to T275, and then its S_t increases in a similar manner.

Fig. 15.

Fracture surfaces of cokes from T225–T300 for T_c = 1000°C.

The above-mentioned hypothesis seems to explain the effect of T_t on S_t of coke at T_c > 500°C, but not the convergence of S_t at T_c = 420–500°C (see Fig. 13). Then, examination of the hypothesis by more-detailed structural analyses of coke and briquette is necessary for understanding of the mechanism of coke strength occurrence over the entire range of torrefaction to carbonization in the future work.

4. Conclusions

The authors studied the sequence of torrefaction, pulverization, and binderless briquetting by the hot-press molding and carbonization, and have demonstrated its applicability to preparation of formed coke with S_t of 8–32 MPa from the chipped cedar. Such large S_t cannot be achieved without torrefaction. It has also been revealed that S_t is maximized with T_t = 275°C. S_t is also increased by increasing the degree of pulverization. The torrefaction not only improves the pulverizability greatly but also induces the self-pulverization.

The inherent honeycomb structure of SD is compressed or broken forming plate-shaped honeycomb walls during the briquetting. The compressed honeycomb and broken honeycomb are more abundant in the torrefied cedar at lower and higher T_t, respectively, due to the loss of flexibility of honeycomb walls by increasing T_t. T_t over 275°C induces a decrease in S_t of briquette, due to collapsing of honeycombs and insufficient bonding of walls of broken honeycombs.

The effect of T_t on S_t of coke with T_t = 1000°C is attributed mainly to increment of S_t at T_c = 500–1000°C, where S_t increases by 1.8–3.6 times. S_t with T_c = 420°C is almost independent of T_t. Different T_t also results in a different relationship between S_t of coke and its apparent density. The higher apparent density does not necessarily lead to greater S_t.

Acknowledgment

A major part of this study was carried out as that of Grant-in-Aid for Scientific Research (A) (Subject Number: 21H04632), which has been financially supported by Japan Society for the Promotion of Science. The other part of this study was financially supported by The Iron and Steel Institute of Japan, and The Japan Science and Technology Agency (for a JST Mirai Program; Grant No. JPMJMI20E6). The authors are grateful to Cooperative Research Program of Network Joint Research Center for Materials and Devices that has been supported by the Ministry of Education, Culture, Sports, Science and Technology, Japan. An author of this paper, Aditya Wibawa, was supported by a Kyushu University Kyushu University Program for Leading Graduate Schools: Advanced Graduate School Program in Global Strategy for Green Asia for his financial support.

Supporting Information

This material contains pyrolysis characteristics of the original and torrefied cedars, elemental composition of coke and SEM images of fracture surface of coke.

This material is available on the Journal website at https://doi.org/10.2355/isijinternational.ISIJINT-2022-013.

References

• 1)   T.  Matsumura,  M.  Ichida,  T.  Nagasaka and  K.  Kato: ISIJ Int., 48 (2008), 572. https://doi.org/10.2355/isijinternational.48.572
• 2)   M. G.  Montiano,  E.  Díaz-Faes,  C.  Barriocanal and  R.  Alvarez: Fuel, 116 (2014), 175. https://doi.org/10.1016/j.fuel.2013.07.070
• 3)   M. A.  Diez,  R.  Alvarez and  M.  Fernández: Fuel, 96 (2012), 306. https://doi.org/10.1016/j.fuel.2011.12.065
• 4)   M.  Castro Díaz,  H.  Zhao,  S.  Kokonya,  A.  Dufour and  C. E.  Snape: Energy Fuels, 26 (2012), 1767. https://doi.org/10.1021/ef2018463
• 5)   Y.  Ueki,  Y.  Nunome,  R.  Yoshiie,  I.  Naruse,  Y.  Nishibata and  S.  Aizawa: ISIJ Int., 54 (2014), 2454. https://doi.org/10.2355/isijinternational.54.2454
• 6)   N. A.  Ryan,  S. A.  Miller,  S. J.  Skerlos and  D. R.  Cooper: Environ. Sci. Technol., 54 (2020), 14598. https://doi.org/10.1021/acs.est.0c04321
• 7)   S.  Kudo,  A.  Mori,  R.  Soejima,  F.  Murayama,  Karnowo,  S. Nomura, Y. Dohi, K. Norinaga and J.-i. Hayashi: ISIJ Int., 54 (2014), 2461. https://doi.org/10.2355/isijinternational.54.2461
• 8)   K.-v.  Yip,  H.  Wu and  D.-k.  Zhang: Energy Fuels, 21 (2007), 419. https://doi.org/10.1021/ef060396l
• 9)   M.  Matoba,  S.  Kudo,  A.  Mori,  K.  Norinaga,  K.  Uchida,  Y.  Dohi,  K.  Uebo and  J.-i.  Hayashi: ISIJ Int., 59 (2019), 1440. https://doi.org/10.2355/isijinternational.ISIJINT-2018-819
• 10)   K.  Uchida,  S.  Kudo,  A.  Mori,  U. P. M.  Ashik,  K.  Norinaga,  Y.  Dohi,  K.  Uebo and  J.-i.  Hayashi: ISIJ Int., 59 (2019), 1449. https://doi.org/10.2355/isijinternational.ISIJINT-2018-847
• 11)   A.  Mori,  S.  Kubo,  S.  Kudo,  K.  Norinaga,  T.  Kanai,  H.  Aoki and  J.-i.  Hayashi: Energy Fuels, 26 (2012), 296. https://doi.org/10.1021/ef201168g
• 12)   S.  Kudo,  A.  Mori,  G.  Hayashi,  T.  Yoshida,  N.  Okuyama,  K.  Norinaga and  J.-i.  Hayashi: Energy Fuels, 32 (2018), 4364. https://doi.org/10.1021/acs.energyfuels.7b03155
• 13)   Karnowo,  S. Kudo, A. Mori, Z. F. Zahara, K. Norinaga and J.-i. Hayashi: ISIJ Int., 55 (2015), 765. https://doi.org/10.2355/isijinternational.55.765
• 14)   K.  Lee,  D.  Lanning,  L.  Lin,  E.  Cakmak,  J. R.  Keiser and  J.  Qu: ACS Sustain. Chem. Eng., 9 (2021), 11652. https://doi.org/10.1021/acssuschemeng.1c02542
• 15)   S.  Yu,  J.  Park,  M.  Kim,  H.  Kim,  C.  Ryu,  Y.  Lee,  W.  Yang and  Y.-g.  Jeong: Energy Fuels, 33 (2019), 8632. https://doi.org/10.1021/acs.energyfuels.9b01086
• 16)   S.  Nam,  M. B.  Hillyer and  B. D.  Condon: Carbohydr. Polym., 228 (2020), 115374. https://doi.org/10.1016/j.carbpol.2019.115374
• 17)   D.  Vural,  J. C.  Smith and  L.  Petridis: Phys. Chem. Chem. Phys., 20 (2018), 20504. https://doi.org/10.1039/C8CP03144D
• 18)   Y.  Liu,  R.  Liu,  M.  Li,  F.  Yu and  C.  He: Carbohydr. Polym., 220 (2019), 141. https://doi.org/10.1016/j.carbpol.2019.05.060
• 19)   Y.  Yu,  D.  Liu and  H.  Wu: Energy Fuels, 26 (2012), 7331. https://doi.org/10.1021/ef3013097
• 20)   Y.  Ohmukai,  K.  Fujimoto,  I.  Hasegawa,  S.  Hayashi and  K.  Mae: J. Chem. Eng. Jpn., 41 (2008), 312. https://doi.org/10.1252/jcej.07we262
• 21)   T.  Liu,  A.  Mori,  R.  Arai,  S.  Asano,  S.  Kudo and  J.-i.  Hayashi: Energy Fuels, 34 (2020), 7183. https://doi.org/10.1021/acs.energyfuels.0c01164
• 22)   Q. V.  Bach,  W.-H.  Chen,  Y. S.  Chu and  Ø.  Skreiberg: Bioresour. Technol., 215 (2016), 239. https://doi.org/10.1016/j.biortech.2016.04.009
• 23)   W.-H.  Chen,  W.-Y.  Cheng,  K.-M.  Lu and  Y.-P.  Huang: Appl. Energy, 88 (2011), 3636. https://doi.org/10.1016/j.apenergy.2011.03.040
• 24)   S.  Kudo,  J.  Okada,  S.  Ikeda,  T.  Yoshida,  S.  Asano and  J.-i.  Hayashi: Energy Fuels, 33 (2019), 11253. https://doi.org/10.1021/acs.energyfuels.9b02939
• 25)   J. W.  Patrick and  A. E.  Stacey: Fuel, 51 (1972), 81. https://doi.org/10.1016/0016-2361(72)90044-0
• 26)   T.  Miyagawa and  I.  Fujishima: J. Fuel Soc. Jpn., 54 (1975), 983. https://doi.org/10.3775/jie.54.12_983