Studies on Curie-Point Pyrolysis of Coal Model Compounds - Energy

Masakatsu Nomura, Levent Artok, Satoru Murata, Akira Yamamoto, Hiroshi Hama, Hong Gao, and Koh Kidena. Energy & Fuels 1998 12 (3), 512-523...
0 downloads 0 Views 766KB Size
Energy & Fuels 1995,9, 119-125

119

Studies on Curie-Point Pyrolysis of Coal Model Compounds Satoru Murata, Toyokazu Mori, Akira Murakami, and Masakatsu Nomura" Department of Applied Chemistry, Faculty of Engineering, Osaka University, Suita, Osaka 565, Japan

Kazuo Nakamura Fundamental Research Laboratories, Osaka Gas Co., Ltd., 19-9, 6-Chome, Torishima, Konohana-ku, Osaka 554, Japan Received June 23, 1994. Revised Manuscript Received September 29, 1994@

Curie-point pyrolysis/GC and /GC/MS analyses were carried out for 14 coal model aromatic compounds with longer alkyl side chains or polymethylene bridges. It was found that (i)pyrolytic products reflected quantitatively constituents of original model compounds (though thermal decomposition in these systems is complicated with several secondary reactions), (ii)a series of 1-alkenes was obtained as major aliphatic products from the pyrolysis of the model compounds having longer alkyl side chain, while Curie-point pyrolysis of some coals or their extracts afforded a series of n-alkanes as major products along with minor amounts of 1-alkenes, and (iii)model compounds such as more condensed aromatic compounds or the aromatic compounds having oxygen functional groups afforded n-alkanes preferentially along with 1-alkenes. On the basis of the results obtained in this paper, we tried to correlate Curie-point pyrolytic products of extracts from Illinois No. 6 coal with its supposed organic chemical structure.

Introduction

Scheme 1 X(CH2)16CH3

It is well-known that coal organic materials (COM) contain a significant amount of aliphatic carbons. For example, Argonne Premium Coal samples were reported to contain 10-40% of aliphatic carbons by referring their 13CNMR spectra.' However, a detailed information about structure and distribution of aliphatic functional groups is not sufficient. To clarify the chemical structures and distribution of aliphatic functional groups in COM, several studies such as pyrolytic oxidation of COM with R u O ~ , ~ - ~ ~ or liquefaction under deuterium have been carried 0 ~ t . l ~

* Abstract published in Advance ACS Abstracts, November 1,1994.

(1)Franz., J. ~- A.: Garcia, R.; Linehan, J. C.; Love, G. D.; Snape, C. E. Energy Fuels 1992, 6, 598. (2) Calkins, W. H.; Hagaman, E.; Zeldes, H.Fuel 1984, 63, 1113. Calkins, W. H.; Tyler, R. J. Fuel 1984,63,1119. Calkins, W. H.Fuel 1984, 63, 1125. (3) Snape, C. E.; Ladner, W. R.; Bartle, K D. Fuel 1985,64, 1394. (4) Stock, L. M. ACC.Chem. Res. 1989,22,427. Stock, L. M.; Wang, S.-H. Energy Fuels 1989, 4, 336. ( 5 ) Nomura, M.; Ida, T.; Miyake, M.; Kikukawa, T.; Shimono, T. Chem. Lett. 1989, 645. Mataubayashi, K.; Nomura, M.; Miyake, M. Chem. Lett. 1990, 291. Nomura, M.; Mataubayashi, K; Miyake, M. Chem. Lett. 1990, 1563. (6) Hama. H.:Matsubavashi. R: Murata. S.: Nomura. M. J. JDn. ~

~

~

~~

~~

Znst.knergy'199h, 72,467,"Hama,H.; MuraG, SI;Nomura, M.; J. jpn. Inst. Energy 1994, 73, 177. (7) Stock, L. M.; Tse, K-T. Fuel 1983,62,974. Stock, L. M.; Wang, S.-H. Fuel 1985, 64, 1713. Stock, L. M.; Wang, 5.-H. Fuel 1986, 65, 1552. Stock, L. M.; Wang, S.-H. Fuel 1987,66,921. Stock, L. M.; Wang, S.-H. Energy Fuels 1989, 3, 533. (8)Mallya, N.; Zingaro, R. A. Fuel 1984, 63, 423. Ilsley, W. H.; Zingaro, R. A.; Zoeller, J. H.,Jr. Fuel 1986, 65, 1216. (9) Boucher, R. J.; Standen, G.; Eglinton, G.; Fuel 1991, 70, 695. Standen, G.; Boucher, R. J.;Eglinton, G.; Hansen, G.; Eglinton, T. I.; Larter, S. R. Fuel 1992, 71, 31. (10) Blanc, P.; Valisolalao, J.; Albrecht, P.; Kohut, J. P.; Muller, J. F.; Duchene, J. M. Energy Fuels 1991,5, 875. (ll)Mojelsky, T. W.; Ignasiak, T. M.; Frakman, 2.;McIntyre, D. D.; Low, E. M.; Montgomery, D. S.; Strausz, 0. P. Energy Fuels 1992, 6,83. Strausz, 0. P.; Mojelsky, T. W.; Lown, E. M. Fuel 1992,71,1355.

10; X=-CH2l b ; X--OCH,1c; X=-C(=O)I d ; X=-OC(=O).

2.;X-Ys-CH22b; X=-OCH,-, Y=-CH,O2c; X=Y=-C(=O)24; X=-OC(=O)-, Y=-C(=O)O.

30;X=-CH,-

M;X=-CH,4b; X=-OCH,-

3b; X=-OCHr 3c; X.-C(.O)3d; x=-OC(.O)-

These results suggested that aliphatic carbons are present in COM as a free form such as n-alkanes or terpanes trapped in COM, alkyl side chains attached to aromatic moieties, and polymethylene bridges between two aromatic moieties. In recent studies, we have been investigating coal chemical structure by the use of CP/MAS 13C NMR coupled with Curiepoint pyrolysis GC-MS techniques.6,6 Curie-point pyrolysis of the coal or its extract afforded a series of n-alkanes and 1-alkenes, alkylbenzenes, alkylphenols, alkylnaphthalenes, and so on. On the basis of these data we proposed a unit chemical structural model for Japanese bituminous Akabira ~ 0 a l . lHowever, ~ pyrolytic behavior of organic compounds in COM remains still unclear. Therefore, in order to get deeper understanding of pyrolytic behavior of organic compounds of COM, we prepared 14 coal model compounds having aliphatic functional groups via several functional groups and examined their pyrolytic behavior using Curie-point pyrolysis (Scheme 1). (12) Murata, S.; U-esaka, K-I.; Ino-ue, H.; Nomura, M. Energy Fuels,

in press.

0887-0624/95/2509-0119$09.00/00 1995 American Chemical Society

Murata et al.

120 Energy & Fuels, Vol. 9, No. 1, 1995

Experimental Section Preparation of Model Compounds. Octadecylbenzene (la),2-octadecylnaphthalene (3a),and 6-octadecyltetralin (4a) were obtained from reduction of the corresponding ketones IC, 3c, and 4c (33 mmol), with NzH4.HzO (132 mmol) and potassium hydroxide (132 mmol) in triethylene glycol (20 mL) at 120 "C for 2 h. After evaporation of hydrazin and water, the products were purified with silica gel column chromatography using hexane as eluant. Aryl octadecyl ethers, lb, 3b, and 4b, were prepared from 1-bromoctadecane (30 mmol) and sodium salt of phenol, 2-naphthol, or 6-hydroxytetralin (100 mmol) in dry DMF (100 mL) at room temperature and purified with silica gel column chromatography using hexane-benzene as eluant. Aryl heptadecyl ketones, IC and 3c, were prepared from octanoyl chloride (30 mmol) and benzene (50 mL) or naphthalene (30 mmol in a 50 mL dry nitrobenzene) in the presence of aluminum trichloride (50 mmol) at 50 "C for 16 h and purified by silica gel column chromatography using hexanebenzene as eluant. Octanoic acid aryl esters, Id and 3d, were prepared from octanoyl chloride (50 mmol) and phenol or 2-naphthol (50 mmol) in distilled ether (100 mL) in the presence of triethylamine (50 mmol) with magnetically stirring at room temperature for one night. These products were purified by recrystallization from hexane. Difunctional compounds 2a-d were synthesized by the similar method to that for the monosubstituted compounds from 1,lO-dibromodecane or 1,lO-decane dioyl chloride and benzene or phenol. Each compound was purified by recrystallization and the purity was judged t o be 295% by GC and NMR analyses. Octadecylbenzene(la)was a solid; mp 32.5-33.5 "C; lHNMR 6 0.88 (t, J = 6.8 Hz, 3 H), 1.25-1.34 (m, 30 H), 1.57-1.64 (m, 2 H), 2.59 (t, J = 7.8 Hz, 2 H), 7.16 (d, J = 11.0 Hz, 3 H), 7.26 (dd, J = 11.0 and 3.6 Hz, 2 H);13C NMR 6 14.1, 22.7, 29.1,29.6,31.6,31.9,36.0,125.4, 127.9,128.3, 142.9; MS, m l e 330 (M+). Octadecyl phenyl ether (lb) was a solid; mp 46.5-47.5 "C; IH NMR 6 0.88 (t,J = 6.6 Hz, 3 H), 1.26-1.30 (m, 28 H), 1.411.48 (m, 2 H), 1.77 (qn, J = 6.9 Hz, 2 H), 3.94 (t,J = 6.9 Hz, 2 H), 6.88-6.93 (t, J = 7.8 Hz, 3 H), 7.23-7.28 (m, 2 H); 13C NMR 6 14.1, 22.7, 26.2, 29.4, 29.8, 31.9, 67.9, 114.5, 120.4, 129.5, 159.1; MS, m l e 346 (M+). 1-Phenyloctadecan-1-one (IC)was a solid; mp 61-62 "C; 'H NMR 6 0.88 (t, J = 6.6 Hz, 3 H), 1.26-1.37 (m, 28 H), 1.63 (qn, J = 7.4Hz, 2 H), 2.95 (t, J = 7.4Hz, 2 H),7.45 (t, J = 7.3 Hz, 2 H), 7.54 (t, J = 7.3 Hz, 3 H), 7.95 (d, J = 7.3 Hz, 2 H); 13CNMR 6 14.1,22,7,24.4,29.6,30.0,32.0,38.6,128.1, 128.5, 132.8, 137.1, 200.6, MS mle 344 (M+). Phenyl octadecanoate (Id) was a solid; mp 50-51 "C; 'H NMR 6 0.88 (t, J = 6.6 Hz, 3 H), 1.26-1.42 (m, 28 H), 1.74 (qn, J = 7.5 Hz, 2 H), 2.54 (t, J = 7.5 Hz, 2 HI, 7.07(d, J = 7.5 Hz, 2 H), 7.19-7.24 (dd, J = 9.8 and 7.5 Hz, 1H), 7.36 (t,J = 7.5 Hz,2 H); 13CNMR 6 14.2,22.6,24.9,29.2,29.4,29.6,29.9, 31.9,34.4, 121.5,125.6,129.3, 150.8,172.2; MS, m l e 360 (M+). 1,lO-Diphenyldecane (2a) was a n oil; lH NMR 6 1.26-1.30 (m, 12 H), 1.60 (qn, J = 7.7 Hz, 4 H), 2.59 (t,J = 7.7 Hz, 4 H), 7.14-7.20 (m, 6 H), 7.26 (dd, J = 12.0 and 7.6 Hz, 4 H); 13C NMR 6 29.4, 29.7, 31.6, 36.0, 125.7, 128.3, 128.5, 142.9; MS, m l e 294 (M+). 1,lO-Diphenoxydecane (2b) was a solid; mp 84-86 "C; 'H NMR 6 1.22-1.32 (m, 8 H), 1.41-1.46 (m, 4 H), 1.73-1.80 (m, 4 H), 3.92 (t, J = 7.8 Hz, 4 H), 6.87-6.93 (m, 6 HI, 7.24-7.28 (m, 4H); 13CNMR6 26.0,29.1,29.3, 67.9, 114.5, 120.4, 129.4, 159.1; MS m l e 326 (M+). (13)Nomura, M.; Muratani, T.; Tajima,Y.; Murata, S. Fuel Process. Technol., submitted for publication. (14) Nomura, M.; Matsubayashi, IC;Ida, T.; Murata, S. Fuel Process. Technol. 1992,31, 169.

1,lO-Diphenyldecan-1,lO-dione (2c) was a solid; mp 93.594.5 "C;1H NMR 6 1.36-1.40 (m, 8 H), 1.70-1.75 (m, 4 H), 2.95 (t, J = 7.4 Hz, 4 H), 7.44 (dt, J = 1.6 and 7.4 Hz, 4 H), 7.54 (tt, J = 7.4 and 1.5 Hz, 2 H), 7.94-96 (m, 4 H); 13C NMR 6 24.7, 29.3, 38.7, 128.0, 128.8, 133.0, 137.3, 200.8; MS m l e 322 (M+). Diphenyl 1,lO-decanedioate (2d)was a solid; mp 64.5-65.5 "C; lH NMR 6 1.39-1.44 (m, 8 H), 1.72-1.79 (m, 4 H), 2.54 (t, J = 7.6 Hz, 4 H), 7.07 (dt, J = 7.8 and 1.2 Hz, 4 H), 7.20 (dt, J = 7.8 and 0.8 Hz, 2 H), 7.33-7.38 (m, 4 H); 13CNMR 6 24.9, 29.2,34.5,121.8,125.9,129.4,150.8,172.2; MS m l e 354 (M+). 2-Octadecylnaphthalene (3a) was a solid; mp 46-47 "C; lH NMR 6 0.88 (t, J = 7.0 Hz, 3 H), 1.25-1.33 (m, 30 H), 1.69 (qn, J = 7.5 Hz,2 H), 2.75 (t, J = 7.5 Hz, 2 H), 7.32 (dd, J = 8.6 and 1.4 Hz, 1H), 7.36-7.45 (m, 2 H), 7.59 ( 8 , 1HI, 7.737.79 (m, 3 H); 13C NMR 6 14.1, 22.7, 29.3, 29.6, 31.4, 32.0, 36.2, 125.0, 125.8, 126.3, 127.4, 127.5, 127.6, 127.7, 131.9, 133.7, 140.5; MS m l e 380 (M+). 2-Naphthyl octadecyl ether (3b) was a solid; mp 62-62.5 "C; IH NMR 6 0.88 (t, J = 6.8 Hz, 3 H), 1.26-1.37 (m, 28 H), 1.46-1.52 (m, 2 H), 1.3 (qn, J = 6.9 Hz, 2 HI, 4.05 (t,J = 6.9 Hz, 2 H), 7.11-7.15 (m, 2 H), 7.30 (dt, J = 1.4 and 7.5 Hz, 1 H), 7.41 (dt, J = 1.4 and 7.5 Hz, 1H), 7.69-7.75 (m, 3 H); 13C NMR6 14.1,22.7,26.1,29.4,29.5,29.7,31.9,68.0,106.5,119.0, 123.4, 126.2, 126.7, 127.6, 128.9, 129.3, 134.6, 157.1; MS mle 396 (M+).

1-(2-Naphthyl)octadecan-2-0ne (3c) was a solid; mp 64.565.5 "C;'H NMR 6 0.88 (t,J = 6.8 Hz, 3 H), 1.25-1.44 (m, 28 H), 1.79 (qn, J = 7.0 Hz, 2 H), 3.08 (t,J = 7.0 Hz, 2 H), 7.527.60 (m, 2 H), 7.87 (t, J = 7.8 Hz, 2 H), 7.95 (d, J = 7.8 Hz, 1 H), 8.03 (dd, J = 7.8 and 1.6 Hz, 1H), 8.46 (s, 1H); 13CNMR 6 14.1, 22.7, 24.6, 29.4, 29.6, 31.9, 38.7, 124.0, 126.7, 127.7, 128.27, 128.34, 129.5, 129.6, 132.6, 134.4, 135.5, 200.5; MS m l e 394 (M+). 2-Naphthyl octadecanoate (3d) was a solid; mp 72-73 "C;

1HNMR60.88(t,J=6.8Hz,3H),1.26-1.46(m,28H),1.79 (qn, J = 7.5 Hz, 2 H), 2.60 (t,J = 7.5 Hz, 2 H), 7.22-7.24 (m, 1H), 7.42-7.49 (m, 2 H), 7.54 (d, J = 2.4 Hz, 1H), 7.77-7.85 (m, 3 H); I3C NMR 6 14.1, 22.7, 25.0, 29.1, 29.3, 29.4, 29.6, 29.9,31.9,34.5, 118.5,121.2, 125.6, 126.5,127.6, 127.7, 129.3, 131.4, 133.8, 148.4, 172.5; MS m l e 410 (M+). 6-Octadecyltetralin (4a) was an oil; lH NMR 6 0.88 (t,J = 6.8 Hz, 3 H), 1.25-1.30 (m, 30 H), 1.52-1.61 (m, 2 H), 1.761.80 (m, 4 H), 2.52 (t, J = 8.2 Hz, 2 H), 2.73-2.78 (m, 4 HI, 6.87-6.91 (m, 2 H), 6.96 (d, J = 8.0 Hz, 1HI; 13C NMR 6 14.4, 23.0, 23.6, 29.3, 29.7, 29.9, 32.0, 32.2, 35.9, 125.9, 129.26, 129.30, 134.5, 137.1, 140.3; MS m l e 384 (M+). Octadecyl 6-tetralyl ether (4b) was a solid; mp 36.5-37.5 "C; lH NMR 6 0.88 (t, J = 6.8 Hz, 3 H), 1.26-1.30 (m, 30 H), 1.41-1.43 (m, 2 H), 1.75-1.78 (m, 4 H), 2.69-2.73 (m, 4 H), 3.90 (t, J = 6.6 Hz,2 H), 6.59 (d, J = 2.4 Hz, 1H), 6.65 (dd, J = 8.5 and 2.2 Hz, 1 H), 6.95 (d, J = 8.5 Hz, 1 H); 13C NMR 6 14.1, 22.7, 23.2, 23.5, 26.1, 28.6, 29.2, 29.6, 31.9, 68.0, 112.3, 114.5, 129.0, 129.8, 138.1, 157.0; MS m l e 400 (M+). Curie-Point Pyrolysis of Model Compounds. Curiepoint pyrolysis gas chromatographic (Py-GC), and mass spectrometric (Py-GC-MS) analyses were carried out by using a Japan Analytical Industry JHP-3 type Curie-point pyrolyzer equipped with a Shimadzu GC-14BPFSC (CBP-1 capillary column, inner diameter 0.25 mm x length 25 m) and the same type of pyrolyzer equipped with a JEOL JMS-DX-303 GC-MS, respectively. Acquisition and analysis of MS data were carried out on a JEOL JMA-DA-5100 data station. Each model compound (ca. 0.5 mg) was pyrolyzed at 670 "C for 3 s at a heating rate of 2500 K/s under a nitrogen stream. Products remained on pyrocell and pyrofoil were defined as tar and coke, respectively (see Figure 1). The weight of these fractions was measured by a microbalance. The weight of volatiles, which were introduced into GC, was calculated using following equation; weight of volatiles = weight of sample - weight of tar - weight of coke.

,.*-

GC probe

coil-:\ Energy &Fuels, Vol. 9, No. 1, 1995 121

Curie-Point Pyrolysis of Coal Model Compounds

*--

FID

.*.

*a-

...‘

-

Carrier gas Pyrocell (N2)

-. -.-. 1.

Pyrofoil

,.-*

*.

..--. --..

--.

n

to GC

u

Figure 1. Schematic diagrams of Curie-point pyrolyzer. Semiempirical MO Calculations. All MO calculations were carried out on a Titan 750V workstation (Kubota Pacific Computer Co.) by using semiempirical molecular orbital calculation program, MOPAC (version 5.0).16 Bond dissociaR R was calculated tion energy of the reaction (AE)R-R’ by using the following equation: AE = AHfo(R’) AHfo(R”)AHfo(R-R’), where AHf, is an energy for standard heat of formation. For each calculation, AM1 method was used and, especially for radicals, AM1-UHF (unrestricted Hartree-Fock) method was used. Titan version of this program was purchased from Simulation Technology Inc.

-

+

+

pyrolytic products in volatile fraction

tar fraction

coke fraction

I /

1 18

lb IC

ld

20

Results and Discussion Pyrolysis of Model Compounds 1 and 2. In our

2b

recent studies, Curie-point pyrolysis of some coals or their extracts was found to afford a series of n-alkanes having more than 30 carbon atoms along with the corresponding l - a l k e n e ~ . ~ However, >~ their origins are still unclear. On the basis of the results reported previously, the possible origins of the aliphatic products are thought to be alkanes or terpanes trapped in COM, alkyl side chains attached to aromatic moieties in COM, and polymethylene bridges between two aromatic moieties. Subsequently, we carried out pyrolysis of model compounds having aliphatic functional groups. We selected aromatic hydrocarbon connecting a longer alkyl side chain and two aromatic hydrocarbons having a polymethylene bridge between them as coal model compounds where we adopted an alkyl-aryl bond, an ether type bond, a ketone type bond, and an ester type bond as a connecting group between aliphatic and aromatic moieties: the latter two functional groups, ketones and esters, are found to be rich in lower rank coals (Scheme 1). Curie-point pyrolysis of benzene derivatives with a longer alkyl side chain (la-d) and a polymethylene bridge (2a-d) was carried out at 670 “C for 3 s under a nitrogen stream. Yields of volatile, tar, and coke are summarized in Figure 2. Yield of coke, which is a major product in pyrolysis of macromolecules such as coal or its extract, was very low (less than 10%). From this

2d

(15) D e w a r , M. J. S.; Zoebisch, E.G.; H e a l y , E.F.;S t e w a r t , J.J. P. J.Am. Chem. SOC.1986,107,3902.

the substrate in volatile fraction

2c

0

20

40

60

80

100

Yield (wt%)

Figure 2. Yield of pyrolytic products from la-d and 2a-d.

figure, it was found that compounds 2a-d afforded less amounts of tar and coke fractions than that from compounds la-d, which might be due t o their low molecular weight. FD-MS analysis of tar fraction from l a revealed that this fraction was composed of the substrate and small amount of its dimerized product (Figure 3). Pyrograms for octadecylbenzene(la)and 1,lO-diphenyldecane (2a) are shown in Figure 4, these suggesting that very complicated reactions occurred. Major products from l a were a series of 1-alkenes and alkylbenzenes and from 2a alkylbenzenes were mainly obtained along with small amount of 1-alkenes. These tendencies were also shown in pyrolysis of other model compounds, lb-d and 2b-d. Major products and their yields (peak area) are shown in Table 1. These results suggested that major origin of aliphatic compounds from pyrolysis of coal is alkyl side chain of COM. Relative yields of alkenes from la-d are shown in Figure 5 (yield of hexene = 1.00, mole ratio). The distribution of C S - C ~alkenes ~ from each compound was similar t o each other, these results suggesting that alkenes are produced from a common intermediate such as higher alkyl radicals.

122 Energy & Fuels, Vol. 9, No. 1, 1995

I

Murata et al. Table 1. Main Products from the Curie-Point Pyrolysis of la-da la lb IC Id hexene 3.48b 5.39 3.84 3.51 heptene 1.54 2.64 1.75 1.53 octene 1.01 1.65 1.19 1.07 nonene 1.74 1.03 1.17 1.04 decene 2.38 1.29 1.49 1.53 undecene 1.78 1.03 1.16 1.06 dodecene 0.87 1.33 1.06 1.08 tridecene 1.25 1.02 1.01 0.92 tetradecene 1.45 0.90 1.20 1.21 pentadecene 2.19 1.02 1.00 0.91 hexadecene 0.74 0.96 4.40 1.72 2.14 heptadecene 0.60 2.56 0.48 octadecene 4.99 0.53

I

1

\

benzene toluene ethylbenzene styrene propylbenzene Ph-CsHs phenol

CHsICH,l,,Q

i

total 0 658

165

0

1.97 0.42 0.46 0.88 1.64

29.02

44.43

4.54 0.74 0.36 1.32 0.10 0.13

0.86 1.29 0.75 2.17 0.18 0.17 15.24

29.03

36.71

10.72

a Each product was pyrolyzed at 670 "C for 3 s under a nitrogen stream. Based on the area of total compounds excluding the area of the substrate.

i

/

0.46 3.56 1.01 6.95 0.44

500

1000

1.2

W

l a Ph-C,.H,,

mi2

Figure 3. FD-MSspectra of l a (top) and the tar obtained from pyrolysis of l a (bottom); (e)indicate the divalent ion of l a (M2+).

..

67

5

10

15

20

Carbon number

Figure 5. Relative yield of 1-alkenes from pyrolysis of la-d (yield of 1-hexene = 1.00, mole ratio).

0

5

io

15 20 Retention time (men)

25

30

Figure 4. Curie-point pyrograms for l a (top) and 2a (bottom) at 670 "C for 3 s under Nz; the Arabic numbers indicate the carbon numbers of 1-alkenes.

The ratio of total aliphatic carbons to total aromatic carbons in the pyrolytic products from la was calculated to be 3.11, this value agreeing well with that of original la (3.01,where the molecular formula of the gaseous products was assumed to be (CH21,. This is a most significant point of this reaction when this pyrolytic

technique is applied to analysis of the components of COM. These results also suggest that, although this pyrolysis proceeds in a complicated fashion along with several secondary reactions and formation of tar and coke fraction, the pyrolytic products obtained could reflect the chemical structure of original COM, if COM is structurally uniform. However, COM is now believed to be heterogeneous so it is important to treat carefully the pyrolytic results of COM. In this respect, the latest information that lighter fraction of COM is very close in constituents to heavier fraction of COM is very interesting. Differences in Pyrolytic Products between from Coal and from the Model Compounds. Pyrolysis of the model compounds afforded a series of 1-alkenes, while, as we had already reported, pyrolysis of coal or its extracts afforded a series of n-alkanes as main products along with minor amount of l - a l k e n e ~ . This ~,~ is a major difference between pyrolysis of coal and that of coal model compounds. We thought three possible reasons for these differences: (i) the presence of hydroaromatic compounds in coal, which might donate hydrogen atoms to alkyl radicals to give n-alkanes, (ii) the difference of degree of condensation of aromatic

Curie-Point Pyrolysis of Coal Model Compounds

Energy & Fuels, Vol. 9,No.1, 1995 123 Table 2. Main Producta from the Curie-Point Pyrolysis of 3a-da 3a

f

67 8

0

9 10 l T a 1 4

5

10

hexene heptene octene nonene decene undecene dodecene tridecene tetradecene pentadecene hexadecen e heptadecene octadecene

17

15 20 Retention time (min)

25

30

Figure 6. Curie-point programs for 3a (top) and 4a (bottom) at 670 "C for 3 s under Nz; the Arabic numbers indicate the carbon numbers of 1-alkenes.

moieties, which could change the pyrolytic behavior of coal or model compounds, (iii) the presence of free n-alkanes in COM, which could become the origin of n-alkanes evolved during pyrolysis of coal. In order to obtain deep insight into these possibilities, we carried out the following two experiments: pyrolysis of the model compounds having more condensed aromatic ring such as naphthalene or hydroaromatic ring such as tetralin and pyrolysis of free n-alkanes. Pyrolyses of naphthalene (3a-d) and tetralin (4ab) derivatives were carried out. Figure 6 shows the pyrograms of 3a and 4a. Derivatives of naphthalene and tetralin and a series of aliphatic compounds were found in these pyrograms, these results suggesting that pyrolysis of these compounds proceeded in a similar fashion to those of la-d (Tables 2 and 3). However, a series of n-alkanes was shown in the pyrograms (Figure 6). The order of the ratio of alkaneslalkenes (average of C1o-C15 aliphatic compounds, mole ratio) increased as follows: -0 (la-d) 0.05 (4a) 0.06 (3c) -= 0.07 (4b) -= 0.11 (3a) 0.14 (3b,d). These results suggest that the more condensed aromatics and oxygen functional groups are some reasons of an evolution of n-alkanes in pyrolysis of coal or its extracts. Pyrolysis of n-hexacosane,which is a model compound of free alkanes trapped in COM, gave a series of 1-alkenes, indicating that free alkanes trapped in COM could be pyrolyzed to 1-alkenes having less carbon number under these conditions. These results indicate that Curie-point pyrolysis of both the alkylaromatics and free n-alkanes afforded 1-alkenes along with varying amounts of n-alkanes, so we could not determine the origin of the alknaes formed during pyrolysis of coal. Calkins et al. had reported that the origin of the lower alkenes produced during pyrolysis of coal is (CH21, group^;^ however, they had not

4.00b 1.84 1.18

1.21 1.47 1.20 1.05 1.32 0.98 1.32 1.45 2.82 0.09

Ar-H' &-Me Ar-Et ArCHeCH2 Ar-C3H7 h-CaH5 ArOH

0.39 6.32 1.68 11.41 1.59

total

41.32

3c

3d

4.50 1.90 1.27 1.34 1.75 1.49 1.15 1.05 1.56 1.65 0.93 1.87 13.11

3b

4.45 1.96 1.35 1.33 1.62 1.31 1.17 1.11 1.47 1.17 2.75 1.29

2.56 1.14 0.84 0.85 1.03 0.90 0.91 0.89 1.39 3.08 1.38

1.37 0.28 0.37 0.93

6.40 1.43 0.83 2.49

0.91 0.55 0.45 1.54

0.36 20.58 57.47

29.17 32.11

47.57

a Each product was pyrolyzed at 670 "C for 3 a under a nitrogen stream. Based on the area of total compounds excluding the area of the substrate. Ar = 2-naphthyl.

Table 3. Main Products from the Pyrolysis of 4a and 4ba 4a 4b hexene 3.36b 4.77 heptene 2.03 1.55 1.22 octene 0.98 nonene 0.97 1.29 decene 1.16 1.76 undecene 1.03 1.42 dodecene 0.78 1.04 0.94 tridecene 1.00 1.17 tetradecene 1.17 pentadecene 0.82 hexadecene 0.60 1.36 2.20 0.64 heptadecene 11.30 octadecene 0.23 naphthalene 2-naphthol tetralin 6-hydroxytetralin total

0.10

0.31 16.24

0.65 3.05 0.76 10.25 43.65

Each product was pyrolyzed at 670 "C for 3 s under a nitrogen stream. Based on the area of total compounds excluding the area of the substrate. a

determined whether these (CH2)n groups are in longer alkyl side chains attached to aromatic moieties or in free alkanes. Pyrolytic Mechanism. Mechanism for pyrolysis of these model compounds was investigated. Poustma et a1.I6 and Savage et a1.17 had investigated the pyrolysis mechanism and kinetics for many coal model compounds and the latter had also reported that pyrolysis of alkylaromatics (Ar-CnHzn+l) at relatively low temperature (around 400 "C) mainly affords methylarenes (Ar-CHs), n-alkenes (Cn-1H2n-2)~ vinylarenes (Ar-CH=CH2), and n-alkanes (Cn-2Hzn-2); however, under the conditions employed in the present study, the (16)Poutama, M. L. Energy Fuels 1990, 4 , 113. (17) Savage, P. E.; Korotney, D. J.Ind. Eng. Chen. 1990,29,499. Srmth, C. M.; Savage, P. E. AIChE J. 1991, 37, 1613; Energy Fuels 1991, 5, 146.

Murata et al.

124 Energy & Fuels, Vol. 9, No. 1, 1995

l a

semiempirical molecular orbital calculation program (MOPAC-AM1 method15),the resulting energies being shown in Scheme 2. Calculated values were lower than the values obtained experimentally, for example, bond PhCH2' R' and dissociation energies for PhCH2-R PhO-R PhO' R' were calculated 51 and 42 kcaV mol (R = C17H35), respectively, these values being 2833%lower than the values obtained experimentally such as 70 (R = C2H5 or n-C3H7) and 63 kcdmol (R = CH3 or C2H5).18 Generally, it is well-known that MO calculations of odd electron systems is very difficult, this being probably the source of errors. However, it was reported that AM1 method could reproduce the relative order of these values,15 so we used these calculated values. We would like to note that these values in Scheme 2 are not exact bond dissociation energies. First, the mechanism of pyrolysis of la is discussed. From pyrolysis of la,toluene, styrene, and 1-alkenes were obtained as major products. First stage of the reaction is thought to be a fission of the bond between a- and P-position to the phenyl ring, giving benzyl and 1-heptadecyl radicals (initiation reaction, eq 1) (see Chart 1). Benzyl radical extracts hydrogen to afford toluene (eq 2). The fate of the heptadecyl radical is very complicated: loss of hydrogen atom (eq 31, releasing ethylene (eq 4), or extraction of hydrogen atom (eq 5) to give 1-heptadecene, 1-pentadecyl radical, or n-heptadecane, respectively. In the propagation step, the radicals produced in situ attacks la to give radicals (eqs 6 and 81, which were converted to styrene, heptadecyl radical, o-phenylalkyl radicals, 1-alkenes, o-phenyl-l-

-

lb

1c

84

41

65

Id

3b

4b

In kcavmol.

pyrolysis temperature was higher, resulting in production of more complicated mixture. The first stage of the pyrolysis is thought to be initiated by cleavage of the weakest bond in the model in unimolecular fashion. Therefore, bond dissociation energies for each bond in the model compounds were very important. So we calculated these values by using

-

+

la 0

cy

0

0

- 0' +co

C17H35

C17H35

*C17H35

18

O"C17Hs

0""7H35

OX.R 1 b-d

wC17H35 111

R.

RH

+

Curie-Point Pyrolysis of Coal Model Compounds .:*alkanes

03-alkenes

00

1000

Energy &Fuels, Vol. 9, No. 1, 1995 126 higher). Although aromatic compounds having several alkyl side chains ( 4 5 ) were found in the pyrogram, these compounds could not be identified until now. However, the results of oxidation reaction of COM with RuOd told us that the abundance of alkyl side chains or polymethylene bridges having more than three carbons was very 1ow,12 suggesting that these aromatic compounds might be polyalkylated one (preferentially polymethylated compounds). On the basis of the results of pyrolsysis of the model compounds and coal extracts, it was suggested that COM was mainly composed of compounds of aromatic hydrocarbons and oxygencontaining aromatic compounds with several alkyl side chains and bridges.

As to the aliphatic structure of COM, some

researcher^^-^ had reported that coal contains longer alkyl side chains attached to aromatic rings and/or ..

1000

1500

Scan number

2000

Figure 7. Total ion chromatograms for the Curie-point pyrolysate from AS fraction of Illinois No. 6 coal.

alkenes, and alkyl radicals (eqs 7 and 9). The alkyl radicals formed were thought to be degradated according t o the eqs 3-5. As to lb, IC,and Id, the first stage of the reaction is thought t o be bond fission of the weakest bonds as shown in Scheme 2 and eqs 10-14, to give phenoxy, phenyl, octadecyl, and heptadecyl radicals. Phenoxy and phenyl radicals abstract hydrogen from another radicals or substrates to give phenol and benzene, which were observed as major products in the pyrolysis of these compounds. The fate of these alkyl radicals were thought to be similar to those from l a (eq 15). In the propagation steps, similar degradation might take place (eqs 16 and 17). As to 3 and 4, pyrolysis proceeded in a similar fashion t o give similar products, but in higher yields. This is partly due to more weaker bonds in 3 and 4 (Scheme 2). Difunctional compounds 2a-d were pyrolyzed in a similar fashion to give benzene derivatives as major products. It is unclear why pyrolysis of the compounds having more condensed aromatic rings such as 3 or 4 afforded more amount of alkanes than that from 1. Consequently, we have to investigate the pyrolytic behavior of the model compounds having more condensed aromatic rings or polyfiunctionalized model compounds. Pyrolysis of COM. Figure 7 shows the typical pyrogram for coal extracts (acetone soluble fraction of Illinois No. 6 coal).6 In this pyrogram, more than 140 compounds were identified including, n-alkanes (C3C28), l-alkenes (c5-c23), alkylbenzenes (-C& styrene, alkylphenols (-C3), alkylnaphthalenes (-CE), dibenzofurans (-C3), and three ring compounds. On the basis of the results of the pyrolysis of model compounds, styrene was thought to be formed from the pyrolysis of the compounds with alkyl side chains (propyl group or (18)McMillen, D. F.; Golden, D. M. Annu. Rev. Phys. Chem. 1982, 33, 493.

linear alkanes which are located in the pores of the coals and are not chemically bound as described in the Introduction, however, we could not confirm the ratio of free n-alkanes t o alkyl side chains on the basis of the only pyrolytic results of coal itself and coal model compounds. To obtain this kind of information, we are now carrying out the further studies concerning more detailed analysis of the pyrolytic products of COM, quantitative and qualitative analysis of 13C NMR of COM, and deuterium incorporation during coal liquefaction.13

Conclusion The results obtained in this study are summarized below.

1. We have investigated Curie-point pyrolytic behavior of 14 coal model compounds and found that pyrolysis proceeds in a very complicated fashion along with several secondary reactions. However, the calculated value for total aliphatic carbons t o total aromatic carbons contained in the pyrolytic products was similar to that of the original model compounds, suggesting that pyrolytic products could reflect the structure of original compounds in a quantitative way. 2. It was found that pyrolysis of model compounds employed in this study afforded l-alkenes as major aliphatic products; on the other hand, pyrolysis of coal or its extract afforded n-alkanes along with minor amount of alkenes. These results suggest that structure of aromatic moieties such as degree of condensation and oxygen functional groups affected strongly the pyrolytic behavior of aliphatic moieties. 3. On the basis of the results of pyrolysis of the model compounds, pyrolysis was thought to proceed via radical mechanism initiated by fission of the weakest bonds in the model compounds in a unimolecular fashion.

4. On the basis of the data from pyrolysis of these model compounds and coal extracts, the features of chemical structure of COM was discussed, briefly. EF940120Q