672
Energy & Fuels 1996, 10, 672-678
Studies on the Chemical Structural Change during Carbonization Process† Koh Kidena, Satoru Murata, and Masakatsu Nomura* Department of Applied Chemistry, Faculty of Engineering, Osaka University, 2-1 Yamada-oka, Suita, Osaka 565, Japan Received June 13, 1995. Revised Manuscript Received March 7, 1996X
In order to investigate a relationship between chemical structure of coking coals and their thermoplasticity during their carbonizationn, evaluation of both hydrogen transfer reaction and heat treatment coupled with SEM observation was conducted along with measurement of SPE/ MAS 13C NMR of virgin coals. The hydrogen transfer reaction from coal to hydrogen acceptor was carried out at 420 °C for 5 min. In the case using anthracene as acceptor, 0.6-1.1 mg of H2 was transferred from 1 g of daf coal to the acceptor. A correlation between the weight of hydrogen transferred and % carbon of each coal (coal rank) showed a similar tendency to that between Gieseler fluidity and coal rank. This result suggests that the quantities of donatable hydrogen could be correlated strongly with the development of plasticity. To obtain the insight into the amounts of functional groups involved in releasing hydrogen, solid state 13C NMR of sample coals was measured, the results indicating the presence of somewhat correlation between the concentrations of bridge methylene groups linking two aromatic moieties and maximum fluidities. Heat treatment of the coals up to their softening temperature, resolidification temperature, and 1000 °C was also conducted, the combination of crystallite parameters of the resulting chars and their SEM observation suggesting that lamellar structures of coal became disordered upon heating and then turned to the ordered structures as the heating proceeds. On the basis of the above results, chemical structural changes during carbonization process are discussed.
Introduction Carbonization is one of the most important processes in coal utilization, especially in Japan. This industry has been developing very sophisticated measurements for evaluation of coal: Gieseler fluidity and vitrinite reflectance of coal are generally used as indices for blending coals to make excellent metallurgical cokes. Gieseler fluidity seems to be a good parameter to represent thermoplastic properties of coal at high temperature, while vitrinite reflectance is considered to be a parameter related to the orientation of aromatic moieties in coal. We, however, are considering that these two parameters seem to be largely based on the empirical science and lack information about chemical reactivity of coal at a molecular level. The mechanism of plasticity had been thought to be very complicated, many studies having been conducted from various points of view. For example, Spiro et al.1,2 constructed a space-filling model for each coal chemical structure proposed by Given,3 Wiser,4 Solomon,5 and Heredy6 on the basis of the concept of three-dimensional coal molecules and explained the development of plasticity using their coal model concepts. Fortin et al.7,8 † Some of this work was presented at the 8th International Conference on Coal Science held in Oviedo, Spain, on September, 1995. X Abstract published in Advance ACS Abstracts, April 15, 1996. (1) Spiro, C. L. Fuel 1981, 60, 1121. (2) Spiro, C. L.; Kosky, P. G. Fuel 1982, 61, 1080. (3) Given, T. H. Fuel 1960, 39, 147. (4) Wiser, W. H. ACS Symp. Ser. 1978, 71, 29. (5) Solomon, P. New Approaches in Coal Chemistry, Am. Chem. Soc. Regional Meeting, Pittsburgh, PA, Nov. 1980. (6) Heredy, L. A.; Wender, I. Prepr. Pap.sAm. Chem. Soc., Div. Fuel Chem. 1980, 25, 38.
0887-0624/96/2510-0672$12.00/0
represented the difference of plasticity with coal rank from the approach to TEM observation of microtexture of coals. On the other hand, there are some studies to understand the plasticity of coal based on its chemical structure and reactivity.9-17 In these studies, two major theories have been proposed so far for chemical concept of thermal plasticity: one is the γ-compound theory and another is hydrogen transfer theory. The former was supported by Ouchi et al.13 on the basis of the results of extraction of Japanese Akabira coal with quinoline or pyridine, namely, the true pyridine-soluble material, which is significantly more abundant than the conventional extraction yield based on Soxhlet extraction, can be extracted when the coal was treated at 400 °C, and the extract has no hydrogen donor property. So, he stated that the plastic property of Akabira coal was explained by the presence of large amount of such soluble material with relatively small molecular weight, i.e., γ-compound. As to the latter one, Neavel9 had stated the importance of transferable hydrogen in coal for the development of plasticity and thought that the (7) Fortin, F.; Rouzaud, J. N. Fuel 1993, 72, 245. (8) Fortin, F.; Rouzaud, J. N. Fuel 1994, 73, 795. (9) Neavel, R. C. Coal Science I; Academic Press: London, 1982; Chapter 1. (10) Painter, P. C.; Rhoads, C. A.; Senftle, J. T.; Coleman, M. M.; Davis, A. Fuel 1983, 62, 1387. (11) Yokono, T.; Obara, T.; Iyama, S.; Yamada, J.; Sanada, Y. Nenryo Kyokaishi 1984, 63, 239. (12) Grint, A.; Mehani, S.; Trewhella, M.; Crook, M. J. Fuel 1985, 64, 1355. (13) Ouchi, K.; Itoh, H.; Itoh, S.; Makabe, M. Fuel 1989, 68, 735. (14) Clemens, A. H.; Matheson, T. W. Fuel 1987, 66, 1009. (15) Clemens, A. H.; Matheson, T. W.; Sakurovs, R. Fuel 1989, 68, 1162. (16) Clemens, A. H.; Matheson, T. W. Fuel 1992, 71, 193. (17) Clemens, A. H.; Matheson, T. W. Fuel 1995, 74, 57.
© 1996 American Chemical Society
Chemical Structural Change during Carbonization
Energy & Fuels, Vol. 10, No. 3, 1996 673
Table 1. Ultimate and Proximate Analyses and Gieseler Parameters of the Sample Coals proximate analyses (wt %, db)
Gieseler plastometerc
ultimate analyses (wt %, daf)
coala
ash
VM
FC
C
H
N
LS GN PM WW WB KP
9.5 9.8 7.3 13.8 8.0 3.8
23.5 23.4 34.3 34.2 32.9 43.4
67.0 66.8 58.4 52.0 59.1 52.8
88.3 88.1 85.7 84.7 82.7 81.2
4.6 5.1 5.5 5.9 4.5 5.9
1.5 1.9 1.7 1.8 2.2 1.3
S
Ob
log MF
ST
MFT
RT
0.3 0.6 1.0 0.6 0.6 0.4
5.3 4.3 6.1 7.0 10.0 11.2
2.30 2.99 3.81 2.47 0.95 0.60
420 397 387 391 412 390
464 456 438 433 432 414
490 498 476 460 446 452
a LS, Lusca; GN, Goonyella; PM, Pittstone-M; WW, Workworth; WB, Witbank; KP, K-Prima. b By difference. c MF, maximum fluidity (ddpm); ST, softening temperature (°C); MFT, maximum fluidity temperature (°C); RT, resolidification temperature (°C).
hydrogen could stabilize radicals generated from thermal bond cleavage reaction during heating of coal. Yokono and Sanada had developed the method to evaluate amounts of transferable hydrogen in coal by using the reaction of coal with anthracene at 400 °C.11 Clemens et al. had reported that oxidation of coal led to loss of coal plasticity and addition of a solvent-soluble fraction of coal and/or polyaromatic hydrocarbons to the oxidized coal resulted in restoring coal plasticity.14-17 In the present study we carried out detailed analyses of hydrogen transfer reaction from coal to several polyaromatic compounds like naphthacene and anthracene along with pyrene and investigated what kind of parts of coal organic materials (COM) play an actual role as hydrogen donor sites based on SPE/MAS 13C NMR measurements of the coals. We also observed the structural changes of coal on heat treatment by X-ray diffraction (XRD) and scanning electron micrography (SEM). Experimental Section Samples. Six coking coals were provided by the Iron and Steel Institute of Japan, these being pulverized (-100 mesh) and dried at 100 °C in vacuo prior to use. Ultimate and proximate analyses and the fluidity characteristics obtained from the Iron and Steel Institute of Japan are summarized in Table 1. Reagents employed in the present study were commercially available and purified by recrystallization or distillation before use. Evaluation of Hydrogen Donor/Acceptor Ability of Coal. A sealed tube (Pyrex, 6 mm inner diameter × 100 mm long) containing a 1:1 mixture (weight ratio, totally 100 or 200 mg) of coal and polyaromatic hydrocarbon was placed in an electric furnace preheated at 420 °C and then kept for 5 min. The temperature inside the sealed tube raised up to desired temperature within 2 min, the heating rate being about 200 K/min. After 5 min passed, the sealed tube was taken out and the products inside the tube were recovered with dichloromethane. Qualitative and quantitative analyses of the products were undertaken by a Shimadzu QP-2000A GC/MS and a Shimadzu GC-14APFSC gas chromatograph with CBP-1 column (0.25 mm diameter × 25 m long), respectively. Semiempirical MO Calculations. All MO calculations were carried out on an Apple Macintosh computer by using a semiempirical molecular orbital calculation program, CAChe (Computer Aided Chemistry) MOPAC 94, which was purchased from CAChe Scientific Inc. The values of heat of formation for the polyaromatic hydrocarbons were determined by solving the Schro¨dinger equation using the AM1 semiempirical Hamiltonian. The difference of the standard heat of formation [∆(∆Hf)] was calculated as follows: ∆(∆Hf) ) ∆Hf (dihydrogenated compound) -∆Hf (aromatic compound), where ∆Hf is the standard heat of formation of each compound. CP/MAS and SPE/MAS 13C NMR Measurements. CP and SPE 13C NMR spectra were recorded on a Chemmagnetic CMX-300 with MAS method (10 kHz). For the measurements,
Figure 1. 13C NMR spectrum for PM coal by SPE/MAS method and deconvoluted Gaussian curves. about 150 mg of coal was packed to a vessel (diameter 5 mm × 8 mm long). The parameters employed were as follows: 200 s pulse delay, 45° pulse width, and 406-437 scan number. Deconvolution of the spectra was conducted on an Apple Macintosh computer with a commercial NMR data processing software, MacAlice (Ver. 2.0, JEOL DATUM). The resulting spectra were divided into 12 Gaussian curves. According to the chemical shifts of model compounds reported by Hayamizu et al.,18 these could be assigned as the following 11 types which are believed to be present in COM: carbonyl (CdO: peak top, around 187 ppm), aromatic carbon connected to oxygen (Ar-O: 167 and 153 ppm), aromatic carbon connected to other carbon (Ar-C: 140 ppm), aromatic bridgehead and/or aromatic tertiary carbon (Ar-C,H: 126 ppm; Ar-H: 113 ppm), aliphatic carbon connected to oxygen (-OCH2O-: 93 ppm; -O-CH2-, -OCH3: 56 ppm), aliphatic carbon of bridge methylene type (CH2′: 40 ppm), aliphatic carbon of methylene type in an alkyl chain (CH2: 31 ppm), methyl group at R-position to aromatic ring (R-CH3: 20 ppm), and methyl group at terminal position of an alkyl chain (t-CH3: 13 ppm). Each deconvoluted curve has an appropriate width to fit to original spectrum. As an example, Figure 1 shows the 13C NMR spectrum for PM coal and deconvoluted curve. Carbon distribution was calculated from the area intensity ratio of each deconvoluted peak. Heat Treatment of Coals and Their X-ray Diffraction Measurement. Heat treatment of coal was conducted by using a tubular electric furnace, Isuzu DKRO-14K, under a nitrogen stream (100 mL/min). Coal was heated up to 300 °C at a heating rate of 5 K/min and heated again up to its softening temperature or resolidification temperature (these were determined by Gieseler plastometry, shown in Table 1) at a rate of 3 K/min. After being kept at each temperature for 30 min, each sample was cooled to room temperature slowly (about 2 K/min). Treatment of coal at 1000 °C was also conducted by heating up to that temperature at the rate of 30 K/min, the sample then being quenched. The resulting semicokes were grounded roughly, then being subjected to X-ray diffraction measurement. The measurement conditions employed here were as follows: Cu KR radiation (λ ) 1.5405 Å), scan range of angle 12-50°, scan rate 10 deg/min. The X-ray apparatus was MAC SCIENCE Co. M18XHF-SRA. The broad peak appeared at around 26° of 2θ in the diffractograms. This (18) Yamamoto, O.; Hayamizu, K.; Yanagisawa, M.; Yabe, A.; Sugimoto, Y. J. Jpn. Inst. Energy 1994, 73, 267.
674
Energy & Fuels, Vol. 10, No. 3, 1996
Kidena et al. Table 2. Difference of Heat of Formation by MOPAC-AM1 Method naphthacene anthracene pyrene
∆Hf(A)a
∆Hf(H2A)a
∆(∆Hf)b
86.9 62.9 67.3
57.7 38.3 49.4
-29.2 -24.6 -17.9
a A, aromatic compound; H A, dihydrogenated aromatic com2 pound. b ∆Hf(H2A) - ∆Hf(A).
Figure 2. Evaluated hydrogen donor ability (HDA) of the sample coal for the reaction with naphthacene (9) and with anthracene (b). could be divided into two Gaussian peaks which were assigned as γ-band and (002) band by curve fitting method, respectively.19,20
Results and Discussion Evaluation of Hydrogen Transferability of the Sample Coal. In order to evaluate a degree of hydrogen transfer reaction, two parameters, hydrogen donor ability (HDA) and hydrogen acceptor ability (HAA), were defined according to the method reported by Yokono et al.21,22 The reaction of the coals with several hydrogen acceptable polyaromatic hydrocarbons such as anthracene (ANT), pyrene (PYR), and naphthacene (NAC) was conducted at 420 °C for 5 min in a sealed tube. Under the conditions, the additive reagents are expected to act as hydrogen acceptor. The reaction of six coals with ANT gave 9,10-dihydroanthracene (DHA) and 1,2,3,4-tetrahydroanthracene (THA) as major hydrogenated products and recovery of additives (sum of yields of ANT, DHA, and THA) was relatively high (>90%). Heat treatment of ANT without coal was also carried out, this affording no detectable amount of hydrogenated products along with the quantitative recovery of ANT. These results indicated that neither disproportionation nor polymerization of ANT occurs in this system. Hydrogen donor ability (HDA) of coal was estimated according to the following equation:
HDA (mg H2/g coal) ) (wt of THA × 4/182 + wt of DHA × 2/180) × 1000/wt of daf coal The parameter, HDA, increased with increase of carbon contents of the coal and then turned to decrease with a maximum value at PM coal (Figure 2). In the reaction of coals with pyrene, the amount of hydrogen transferred from coal to pyrene (PYR) was very small ( ANT (24.6) > PYR (17.9). This order agrees well with the amount of hydrogen transferred from coal to the acceptors. In the case using 9-methylanthracene (9MA) as hydrogen acceptor, ANT was produced (in 2-10 wt % yield) along with the normal hydrogenated compounds like 9,10-dihydro-9-methylanthracene (DHMA). These results might suggest that transferable hydrogen in COM plays a role for not only stabilization of the radicals generated by bond fission but also acceleration of aryl-alkyl bond cleavage by ipso-hydrogen attack. A plot of maximum fluidity of the sample coals toward their rank (% of carbon) is shown in Figure 3. From Figures 2 and 3, both amounts of hydrogen transferred from the coals to anthracene and Gieseler fluidity of the coals against coal rank show a very similar tendency, this suggesting that the specific transferable hydrogen in coal plays a very important role in the plastic stage of coal as suggested by Neavel9 and Clemens.16,17 The reaction of the six coals with DHA, one of the typical hydrogen donatable hydrocarbons, was also conducted, where only the dehydrogenation of DHA to ANT was expected; however, the production of THA was also observed. When the heat treatment of DHA without coal was conducted, the extent of dehydrogenation to ANT and/or disproportionation of DHA to ANT and THA was far smaller than for the case with coal. Although its chemistry is not clear, hydrogen acceptor ability of coal was estimated according to the following equation with the assumption that THA was produced by the coal-catalyzed disproportionation of DHA:
HAA (mg H2/g coal) ) (wt of ANT × 2/178 wt of THA × 2/182) × 1000/wt of daf coal
Chemical Structural Change during Carbonization
Figure 4. Evaluated hydrogen acceptor ability (HAA) of the sample coal for the reaction with 9,10-dihydroanthracene.
The observed values of the amount of hydrogen from DHA to coal are shown in Figure 4, this suggesting that low-rank coal tends to consume more amount of DHA. Generally, low-rank coal is believed to have more amount of weaker bonds such as Ar-CH2-CH2-Ar′ or Ar-CH2-O-Ar′, DHA playing some role to stabilize the radical species generated from the cleavage of these weaker bonds. Therefore, the amount of hydrogen transferred from DHA to coal might indicate the reactivity of coal.23 CP/MAS and SPE/MAS 13C NMR Measurements of the Sample Coals. The results of hydrogen transfer reaction indicated that coal has some specific hydrogens transferring to several hydrogen acceptors. The candidate for specific hydrogens may be the one located at naphthenic rings such as the center ring of 9,10dihydroanthracene and aliphatic substituents linked with aromatic rings such as the methylene position of diarylmethane or 1,2-diarylethane. Clemens et al. had pointed out that ethylene bridges between two aromatic moieties (e.g., 1,2-diphenylethane) and hydroaromatic sites play an important role in coal plastic stage.16 So, we tried to evaluate amounts of these kinds of functional group in coal on the basis of the data from 13C NMR spectroscopy. Generally, in order to get 13C NMR spectra with the higher signal to noise ratio in a short period, CP/MAS method is very convenient.24 However, many researchers had reported that the CP method does not give us quantitative information concerning carbon distribution because this method is easily affected by the presence of paramagnetic species in COM. In their studies, the use of the single pulse excitation (SPE) method or Bloch decay was recommended for more quantitative 13C NMR analysis of coal. Indeed, we had conducted 13C NMR measurement using both CP/MAS and SPE/MAS methods, the carbon aromaticity (fa) from the latter method being 10-20% higher than those from the former method. So, we decided to employ the more time consuming but more quantitative SPE/MAS NMR ac(23) As to the reaction of coal with ANT, one reviewer pointed out the possibility that the apparent low HDA of low-rank coal is due to undergo back-donation from DHA to coal. Clemens et al. (ref. 16) also stated the possibility of this reaction. Clemens et al. carried out the reaction of coal with ANT for 0.5 to 1 h (heating the samples from 300 to 400 or 445 °C at heating rate of 3 K/min), while our reaction time was 5 min (including the period to raise temperature). So, we are thinking that such secondary reactions (back-donation, condensation, or polymerization) could be suppressed greatly under our conditions. (24) Pines, A.; Gibby, M. G.; Waugh, J. S. J. Chem. Phys. 1973, 59, 569.
Energy & Fuels, Vol. 10, No. 3, 1996 675
Figure 5. Distribution of 11 kinds of carbon of the sample coals determined by SPE MAS 13C NMR measurement.
Figure 6. Calculated active methylene carbons for the sample coals.
Figure 7. Crystallite parameters (a) Lc and (b) layer spacing, by X-ray diffraction measurement of the sample coals treated at softening temperature (ST), resolidification temperature (RT), and 1000 °C.
cording to the method reported by Snape et al.25 The resulting spectra were separated into 11 kinds of carbon; the carbon distributions of each coal are shown in Figure 5. The aliphatic region (2500 K/s.31 Yields of semicokes were 73-86 % for pyrolysis at resolidification temperature, 37-61% for treatment at 1000 °C, and 72-81% for rapid pyrolysis at 764 °C. At resolidification temperature, the size of the particles in semicoke was developed more or less and larger particles could be seen than the case of raw coal. The ordered structure can be seen in the case of PM coal. Furthermore, when the heat treatment was conducted at 1000 °C, the unit of the structure was fairly developed, especially in the case of PM coal. This observation is parallel with the behavior of Lc value from the X-ray diffraction measurement. In all cases, pyrolytic residues showed well-ordered structures and well-developed pore systems in spite of the higher yields of semicokes, suggesting that rapid heat treatment might be effective for coke formation even if low-rank coal was used. Chemical Structural Changes during Carbonization Process. On the basis of the results obtained here, chemical structural changes during the carbonization process are discussed briefly. The relationship between the amount of hydrogen from coal to ANT and coal rank could indicate that the specific hydrogens in the COM play some roles in coal plastic stage. The role of these hydrogens might be to stabilize radicals generated by bond cleavage reaction or attack to ipso-position of alkyl substituents to assist the cleavage of some arylalkyl bonds as shown in Figure 9. This had been suggested by Malhotra and McMillen.32,33 The transferable hydrogen is believed to locate at naphthenic rings such as 9,10-dihydroanthracene. This assumption agrees well with the results that plots of concentration of active methylene carbons and maximum fluidity vs coal rank showed a similar tendency. Based on the results of this work and previous wok, we proposed the schematic model for the beginning (29) Hurt, R. H.; Davis, K. A.; Yang, N. Y. C.; Headley, T. J.; Mitchell, G. D. Fuel 1995, 74, 1297. (30) Sugawara, K.; Abe, K.; Sugawara, T.; Nishiyama, Y.; Sholes, M. A. Fuel 1995, 74, 1823. (31) Nomura, M.; Mori, T.; Murakami, A.; Murata, S.; Nakamura, K. Energy Fuels 1995, 9, 119. (32) McMillen, D. F.; Malhotra, R.; Tse, D. S. Energy Fuels 1991, 5, 179. (33) Malhotra, R.; McMillen, D. F. Energy Fuels 1993, 7, 227.
678
Energy & Fuels, Vol. 10, No. 3, 1996
Kidena et al.
Figure 10. Schematic representation for reaction model at the beginning stage of carbonization process.
stage of carbonization process (Figure 10). With heating up to around 400 °C, cleavage of weaker bond and hydrogen transfer occur almost simultaneously. This transferring hydrogen may be supplied from naphthenic rings and consumed for the stabilization of the thermally generated radicals as mentioned above. The aromatic compounds with naphthenic rings were thought to have nonplanar structure as shown in Figure 10, this being converted into a planer and more condensed structure via release of hydrogen, and the resulting polyaromatic hydrocarbons are thought to be a kind of precursor for the high-quality cokes. Therefore, hydroaromatic structure could act as not only hydrogen source but also highly stacked aromatic structure. The results of XRD measurement gave several pieces of information on the behavior of the aromatic clusters during heat treatment. At the softening temperature, the orientation of the aromatic sheets becomes disordered, and the lamellar portions develop as temperature raised. Therefore, during the plastic stage, bond cleavage reaction, hydrogen transfer reaction, and rearrangement of aromatic sheets occur, and the balance of such reactions seems to be important. The rapid pyrolysis leads to bond cleavage, hydrogen transfer, and molecular movement within a short period, so it is thought that the hydrogen would be forced to transfer in local area, so that hydrogen uptake takes place effectively and polymerization of COM (leading to disordered coke) may be suppressed. Conclusions We conducted a set of experiments for six coking coals. From the reaction of coal with three kinds of hydrogen-
accepting hydrocarbons, we found out that the amounts of donatable hydrogen, which could also be quantified by solid state 13C NMR measurement, correlate with Gieseler fluidity, and only specific hydrogens should be effective in the development of plasticity. These specific hydrogens in coal could be transferred to anthracene under the reaction conditions (420 °C, 5 min). Therefore, by the reaction with many coal model compounds, active sites for hydrogen transfer affecting coal plasticity might be clarified. On the basis of the XRD and SEM observation for the heat-treated coals (semicoke), lamellar structure was found to loosen at the level of softening temperature and no apparent cluster of COM could be seen except for the case of PM coal, which has the highest fluidity among the coals employed. While the timing among the hydrogen transfer and molecular rearrangement must be important in the development of plasticity, we had monitored them by a pair of investigations, namely, hydrogen transfer reaction and XRD coupled with SEM observation.
Acknowledgment. This work was partially supported by a Grant-in-aid provided by the Iron and Steel Institute of Japan, and the authors acknowledged the support of Toray Research Center Co. Ltd. for NMR measurement and Sumitomo Metal Industries Ltd. for useful discussion. EF9501096
