Pyrolysis of Coal Model Compounds. Thermal Behavior of Benzyl

Feb 28, 1995 - The cleavage reaction possibly proceeds by initial ¿pso-hydrogen addition, which appears to be enhanced by addition of DHA. (b) The ma...
0 downloads 0 Views 733KB Size
Download PDF
Energy & Fuels l996,9,849-854

849

Pyrolysis of Coal Model Compounds. Thermal Behavior of Benzyl-Substituted Polyaromatic Compounds Satoru Murata, Masayuki Nakamura, Masahiro Miura, and Masakatsu Nomura* Department of Applied Chemistry, Faculty of Engineering, Osaka University, Suita, Osaka 565, Japan Received February 28, 1995. Revised Manuscript Received May 22, 1995@

Pyrolytic behavior of benzyl-substituted naphthalenes, phenanthrenes, anthracenes, and fluorenes as coal model compounds has been investigated. The reaction was carried out a t 430 "C in a sealed tube in the presence or absence of 9,lO-dihydroanthracene (DHA) as a hydrogen source. The results obtained are as follows. (a) a-Substituted type substrates 1,5-dibenzylnaphthalene, 9-benzylphenanthrene, and 9-benzylanthracene selectively give toluene as the cleavage product along with the corresponding polyaromatic fragments accompanied by formation of oligomeric products. The cleavage reaction possibly proceeds by initial ipso-hydrogen addition, which appears to be enhanced by addition of DHA. (b) The major reaction of P-substituted type compounds 2,6-dibenzylnaphthalene,2- and 3-benzylphenanthrenes, and 2-benzylanthracene is oligomerization which may involve radical intermediates formed by hydrogen abstraction from the substrates and from the oligomers benzene may be selectively liberated as the volatile product, with the exception of the reaction of 2-benzylanthracene. In the presence of DHA, toluene can also be formed together with benzene, as for the a-type compounds. (c) The thermal behavior of 1-,2-, and 4-benzylfluorenes is rather similar to that of the ,&type compounds.

Introduction Of relevance to coal pyrolysis and liquefaction, there have been numerous studies on thermal behavior of coal model compounds in the presence or absence of hydrogen s0urces.l The results may also provide useful information in undertaking the efficient utilization of coal as fuel and chemicals by the processes. One of the model compound types extensively subjected to pyrolysis is diary1 methane,2-6 since the methylene bond is considered to be an important linkage connecting aromatic fragments in ~ 0 8 1as , ~well as ether, dimethylene, and polymethylene. The possible modes of cleavage of the relatively strong bond may involve ipso-hydrogen attack by free hydrogen addition, reverse radical disproportionation (RRD), and radical hydrogen transfer (RHT) mechanisms.1-6 In unsymmetrical diarylmethanes, there exist two possible cleavage directions. Futamura et al. described that 9-benzylphenanthrene undergoes hydrogenolysis in the presence of tetralin to give phenanthrene together with t01uene.~ Pyrolysis of 1-(4-(2-phenylethyl)benzyl)naphthalenein the presence of 9,lO-dihydroanthracene was reported to produce predominantly naphthalene by Farcaciu et al.5 In such benzyl-substituted polyaromatic compounds, the Abstract published in Advance ACS Abstracts, July 1, 1995. (1) Poutsma, M. L.Energy Fuels 1990,4, 114. ( 2 ) Malhotra, R.; McMillen, D. F. Energy Fuels 1993, 7, 227. McMillen, D. F.; Malhotra, R.; Tse, D. S.Energy Fuels 1991,5,179. Malhotra, R.;McMillen, D. F. Energy Fuels 1990,4, 184. McMillen, D.F.; Malhotra, R.; Hum, G. H.; Chang, S.-J. Energy Fuels 1987,1 , 193. McMillen, D.F.;Malhotra, R.; Chang, S.-J.;Ogier, W. C.; Nigenda, S. E.; Fleming R. H. Fuel 1987,66, 1611. (3)Allen, D.T.;Gavalas, G. R. Fuel 1984,63, 586. (4) Futamura, S.;Koyanagi, S.; Kamiya, Y. Fuel 1988,67, 1436. ( 5 ) Farcasiu, M.;Smith, C. Energy Fuels 1991,5,83. (6)Franz, J. A.; Ferris, K. F.; Camaioni, D. M.; Autrey, S. T.Energy Fuels 1994,8, 1016. ( 7 )Heredy, L.A.; Kostyo, A. E.; Neuworth, M. B. Fuel 1964,43, 414.Heredy, L.A.; Kostyo, A. E.; Neuworth, M. B. Fuel 1965,44,125. @

1.5-DBN; 1,s-dibenzyl 2.6-DBN; 15-dibenzyl

2-BA; 2-benzyl 9-BA; 9-benzyl

2-BP 2-benzyl 3-BP: 3-benzyl 9.BP 9-benzyl

I-BF I-benzyl 2-BF 2-benzyl 4 - B F 4-benzyl

Figure 1. Substrates subjected to pyrolysis.

position of the substituent could also affect the reaction path. However, the influence has been little examined, while with a specific substrate, 1,2'-dinaphthylmethane, McMillen, Malhotra, and their co-workers have clearly demonstrated that the product ratio of l-methylnaphthalene to the 2-isomer depends on the hydrogen donor compounds added. In light of these results, and as part of our study of pyrolysis of coal and its model molecules,s~gwe have investigated pyrolytic behavior of benzyl-substituted naphthalenes, phenanthrenes, anthracenes, and fluorenes at different positions of these polyaromatic nuclei shown in Figure 1 with or without addition of 9,lOdihydroanthracene as a hydrogen source. The results are herein described.

Results and Discussion Pyrolysis of Dibenzylnaphthalenes. The reaction of 1,5-dibenzylnaphthalene (1,5-DBN) and the 2,6(8) Nomura, M; Matsubayashi, K.; Miyake, M. Chem. Lett. 1990, 1563. Hama, H. Murata, S.; Nomura, M. J . Jpn. Inst. Energy 1994, 73, 177.

(9) Murata, S.; Mori, T.;Murakami, A. Nomura, M. Energy Fuels 1996,9,119.

0887-0624/95/2509-0849$09.00/00 1995 American Chemical Society

Murata et al.

860 Energy & Fuels, Vol. 9, No. 5, 1995 Table 1. Pyrolysis of 1,5- and 2,6-Dibenzylnaphthalenes

(DBN)= conproducts (%o)cd DHAb version substrate (equiv) (%Y Be Tf BNg BMNh BPT' 14:86 15.9 5.6 35.7 28.5 1.3 1,B-DBN none 1.6 9:91 9.1 94.4 48.4 21.8 1,5-DBN 5 9:91 28.0 7.3 67.7 38.1 1.2 1,5-DBN 10 1.7 1OO:O tr 2,6-DBN none 11.8 18.2 tr 41:59 15.8 27.4 38.9 26.1 3.6 2,6-DBN 5 38:62 20.7 19.9 33.0 27.3 3.7 2,6-DBN 10 The reaction was carried out at 430 "C for 1 h. 9,lODihydroanthracene. Determined by GC analysis. Based on substrate consumed. e Benzene. f Toluene. g Benzylnaphthalene. Benzylmethylnaphthalene. Benzenekoluene ratio.

Scheme 1 path a

Y

Table 2. Pyrolysis of 2-, 3-, and 9-Benzylphenanthrenes (BP) and 2- and 9-Benzylanthracenes(BAP conP MP DHAb version or or substrate (equiv) (%F Be Tf Ag PAh BPTI 2-BP none 15.3 6.2 td tr tr 1OO:O 4753 2-BP 10 17.8 7.3 8.4 k 9.9 4.4 1OO:O 3-BP none 8.2 4.6 tr 17:83 10 9.4 4.4 21.2 3-BP 10:90 9-BP none 31.5 4.3 40.0 47.0 tr 9-BP 10 33.3 3.7 79.1 5:95 2-BA none 53.0 0.4 19.4 13.9 1.5 2:98 2-BA 10 69.4 1.4 38.4 4:96 9-BA none 98.3 0.6 49.5 66.9 tr 1:99 9-BA 10 98.6 0.6 98.9 1:99 The reaction was carried out at 430 "C for 1 h. * 9,lODihydroanthracene. Determined by GC analysis. Based on substrate consumed. e Benzene. f Toluene. g P, phenanthrene, A, anthracene. * MP, methylphenanthrene, MA, methylanthracene. Benzenehluene ratio. J Trace amount. Not determined.

CH:

T

DBN

lpathb

BN

0 B

BMN

isomer (2,6-DBN)was at first conducted in a sealed tube at 430 "C for 1 h without addition of 9,lO-dihydroanthracene (DHA)and the resulting mixture was analyzed by GC and GC-MS. The conversions of 1,5- and 2,6DBN were 16 and 12%, respectively (Table 1). The observed higher conversion of 1,5-DBN parallels with the general trend that a 1-substituted naphthalene compound is thermodynamically less stable than the corresponding 2-substituted isomer. From 1,B-DBNwas produced a mixture of toluene and benzene in a ratio of 86:14 together with 1-benzylnaphthalene (1-BN) and 1-benzyl-5-methylnaphthalene(1,B-BMN). This suggest that path a in Scheme 1 preferably occurs compared with path b. The yields of toluene and 2-BN were around 30%, suggesting that the hydrogen source of the cleavage reaction is the substrate itself and hence oligomeric products, which are derived from a free radical, possibly a benzylic radical, were also produced. In contrast to the reaction of 1,5-DBN, 2,6-DBN selectively gave benzene in a yield of 18%,only a negligible amount of toluene being detected. However, the corresponding fragment product, 2,6-BMN was very low (less than 2%). These results may suggest that the major reaction path is different from that of 1,5-DBN. In the case of 2,6-DBN, oligomerization of the substrate may be the major reaction and benzene is formed after the process. It should be noted that 1-BN and 2-BN are stable under the conditions employed. This could be due to the fact that a large part of BN (bp ca. 350 "C) exists in the vapor phase, whereas that of DBN does as liquid under the reaction conditions employed. Addition of 5 or 10 equiv of DHA to the reaction considerably increased both the conversions of 1,5- and 2,6-DBN, indicating that hydrogen transfer from DHA to DBN effectively takes place. In the reaction of 2,6DBN, toluene was preferably formed along with 2-BN, in contrast to that without DHA. Thus, DHA appears to enhance path a in Scheme 1,while homo- and crosscoupling reactions of radicals formed from 2,6-DBN and DHA could also take place, since both the yields of toluene and 2-BN were around 30%.

Pyrolysis of Benzylphenanthrenes and Anthracenes. The resulk for the reaction of benzylphenanthrenes (BP) and benzylanthracenes (BA)are presented in Table 2. In the reaction of 9-BP without DHA, toluene (40%) and phenanthrene (47%) were formed along with benzene (4.3%), the conversion of the substrate being 31%. By contrast, 2-BP gave benzene (6.2%) as a sole measurable product. Similarly, a negligible amount of toluene was formed in the reaction of 3-BP, while phenanthrene (9.9%)was detected in a comparable amount with 3-methylphenanthrene (3-MP, 4.4%). These low product yields suggest that the reaction of 2- and 3-BP produces a large amount of high molecular weight products. The conversions of 2- and 3-BP were 15 and 8.2% which are smaller than that of 9-BP. These results may indicate that the reaction of 9-BP parallels with that of 1,5-DBN and the behavior of 2- and 3-BP is similar to 2,6-DBN. Addition of 10 equiv of DHA to the reaction of 9-BP enhanced toluene formation, which is similar to the reaction of 1,B-DBN. In the case of 2- and 3-BP, the formation of toluene was induced, as for the reaction of 2,6-DBN. However, the increase in the substrate conversion of each BP was relatively small compared with that of the corresponding naphthalene derivatives. The reaction of 9-BA without DHA gave a mixture of toluene (50%)and anthracene (67%)along with benzene (0.6%), as for the reaction of 9-BP. It was somewhat surprising that 2-BA also gave a similar product mixture, although the yield of each product was relatively low. This is in contrast t o the reaction of 2-BP where benzene was the single measurable product. Addition of DHA to the reaction of 2- and 9-BA also enhanced the yield of toluene. The conversion of 2- and 9-BA were considerably higher than 2- and 9-BP, respectively, suggesting that the anthracene derivatives are relatively more reactive than the corresponding phenanthrene compo~mds.~ Pyrolysis of Benzylfluorenes. The results for the reaction of 1-,2-, and 4-benzylfluorenes (1-,2-, and 4-BF) are summarized in Table 3. The reaction of 1-and 4-BF without DHA gave benzene and toluene in comparable amounts, while the yields of the corresponding fragments, fluorene, and methylfluorene were very low, suggesting formation of large amounts of oligomeric products. Interestingly, 2-BF selectively gave benzene,

Energy & Fuels, Vol. 9,No. 5, 1995 851

Pyrolysis of Coal Model Compounds Table 3. Pyrolysis of 1-, 2-, and 4-Benzylfluorenes (BF)" conproducts (%Fd DHAb version substrate (equiv) (%)c Be Tf Fg MFh B/l" 44:56 0.9 9.0 13.3 17.0 4.8 1-BF none 20:80 32.5 4.2 17.3 13.6 K 1-BF 10 tr 1OO:O tr 2-BF none 7.8 21.5 trj 56:44 10.2 8.1 14.3 10 24.2 2-BF tr 69:31 10.9 4.9 tr 4-BF none 7.2 58:42 15.6 6.3 4.5 4.0 4-BF 10 ~~

a The reaction was carried out a t 430 "C for 1 h. 9,lODihydroanthracene. Determined by GC analysis. Based on substrate consumed. e Benzene. f Toluene. Fluorene. Methylfluorene. i Benzene/toluene ratio. J Trace amount. Not determined.

& 90.4

no toluene being detected. Addition of DHA to each reaction increased the ratio of toluene to benzene as expected.

Figure 2. Heats of formation of ipso-hydrogen adducts of 1,5dibenzylnaphthalene (1,5-DBN) and 2,6-dibenzylnaphthalene (2,6-DBN).

Reaction Scheme The initial step of the reaction of 1,B-DBNto toluene and 1-BN and to benzene and 1,5-BMN(Scheme 1)may involve ipso-hydrogen addition. The hydrogen may be provided (a) from another substrate or DHA by a bimolecular process accompanied by formation of a benzylic radical formed by hydrogen abstraction of the substrate or DHA radical (9-hydroanthracenyl radical) (RRD mechanism) and/or (b) from DHA radical accompanied by formation of anthracene (RHT mechanism). A recent theoretical study has suggested that the RRD mechanism might be preferable.1-6 However, we do not argue the precedence of the two reaction mechanisms, since detailed kinetic investigation has not been carried out. By considering the reaction conditions and the substrate employed, free hydrogen does not seem to participate significantly in the reaction, since hydrogen adducts to the substrates may be enough stable to form free hydrogen atom. Hydrogen transfer t o a non-ipso position could also lead to the products.6 Its participation, if any, seems to be small, since no traces of dihydrogenated products could be detected even by careful analysis of the product mixture in the early stage of the reaction. Of the two possible sites of ipso-hydrogen addition, the results for the reaction indicate that the addition to the naphthalene ring predominantly occurs and that to the benzene ring is much less favorable process. This may be interpreted in terms of the relative stability of the two hydrogen adduct intermediates. Semiempirical MO calculation (by MOPAC-AM 1 method, see Experimental Section) of these intermediates has also suggested that the hydrogen adduct at the naphthalene ring is relatively more stable as intuitively expected based on the resonance stabilization (Figure 2). As noted above, the reaction may also produce high molecular weight products derived from the substrate radical, while their formation is apparently suppressed in the presence of DHA. The reactions of 9-BP and 9-BA may proceed, as does that of 1,5-DBA. The selective formation of toluene from these substrates is in harmony with the results reported p r e v i ~ u s l y . ~ , ~ The MO calculation of two possible hydrogen adducts to 2,6-DBN has also suggested that the adduct a t the naphthalene ring is relatively more stable (Figure 2). Therefore, from the substrate toluene along with 2-BN could also formed. In the presence of DHA, it was

-

Dimer (a)

I i

2

20. OO

'

'

500

1000

1500

2000 mlz

500

lo00

1500

2000

i

m/z

Figure 3. FD-MS spectrum of the reaction mixture of (a) 1,5dibenzylnaphthalene (1,5-DBN) and (b) 2,6-dibenzylnaphthalene (2,6-DBN).

indeed the preferable product. However, benzene was predominantly produced as the single major product in the absence of DHA. Thus, the major reaction path appears to be different from that of 1,5-DBN. It may be conceivable that in the case of 2,6-DBN alone, oligomerization of the substrate seems to be the major reaction and benzene may be formed from the oligomeric products as considered precedently . Consequently, the product mixture from the reaction of 1,5- and 2,6-DBN without DHA was analyzed by FDMS (Figure 3). Both the spectra show peaks in the m / z region more than 1200, suggesting that oligomeric products at least up to tetramers were formed. Among the peaks, those at m / z 614 and 920 are confidently assignable to the corresponding dimers and trimers. The existence of the other considerable peaks suggests that the oligomeric products underwent further reaction. The significant peak at m I z 524 may be due to a compound formed by cleavage of the dimer with liberation of toluene. The observed higher yield of toluene compared with 1-BN in the reaction of 1,5-DBN may be due to the contribution of such a secondary reaction. The fact that only a trace amount of toluene was detected in the reaction of 2,6-DBN,however, suggests that in this case,

Murata et al.

852 Energy & Fuels, Vol. 9, No. 5, 1995 100,

'B'

Dimer. Benzene A(114.5)

B(115.5)

E(118.6)

D(119.6)

C

I

Trimer

mlz

Figure 5. FD-MS spectrum of the reaction mixture of 2-benzylphenanthrene (2-BP). G(112.3)

F (121.9)

Figure 4. Possible mechanism for formation of benzene in the pyrolysis of 2,6-dibenzylnaphthalene (2,6-DBN).Value in parentheses indicates heat of formation (kcallmol) calculated by the MOPAC-AM1 method. compound(s)having m l z 524 could not be formed from its dimer and come from higher oligomeric products, although the details are not clear. In the spectrum of the product mixture form 2,6-DBN, a considerable signal at m l z 538 which corresponds to compound(s) dimer-benzene is also observed. It should be noted that the corresponding peak in the spectrum that from 1,5DBN was relatively small. Thus, formation of the compound(s) having mlz 538 may provide the clue for the origin of the selective formation of benzene from 2,6DBN. Based on the above results, a possible route which may lead to the formation of benzene from 2,6-DBN is illustrated in Figure 4. The heats of formation of the postulated radical intermediates were also calculated by the semiempirical MO method. To save calculation time the intermediates were represented by somewhat simplified compounds; a 2,6-DBN molecule and its dimerization pair were replaced by 2-BN and naphthalene, respectively. Reaction of a benzylic radical formed from 2-BN reacts with naphthalene (pair A) gives radical B and the subsequent hydrogen loss (possibly by reaction with another radical species in the medium) affords a dinaphthylphenylmethane C which corresponds to a 2,6-DBN dimer. ipso-Hydrogen addition to C regenerates radical B or gives two other possible radicals D and F. Radicals B,D, and F may afford pairs A, E,and G, respectively. The MO calculation suggests that G is the most stable pair, while the most stable hydrogen adduct radical is B. The observed selective formation of benzene from 2,6-DBN would imply equilibrium between C and E as well as C and A. While interconversion among radicals B,D, and F could occur without formation of C, it does not seem to be significant, since the process had to involve a relatively high energy barrier.6 In order t o confirm which product, i.e., benzene or naphthalene, is favorably formed from the dimer equivalent C, (1,2'-dinaphthyl)phenylmethanewas prepared and pyrolyzed under the same conditions as those applied to 2,6-DBN. It was observed that the compound was consumed in 17% conversion to produce benzene selectively;the yields of benzene and naphthalene were 9.0 and 1.3%,respectively. This suggests that benzene may be selectively liberated from 2,6-DBN dimer and also from higher oligomeric products. It is noted that, in the reaction of the triarylmethane, formation of 1,2'-

146.8

147.6

Figure 6. Possible mechanism for formation of benzene in the pyrolysis of 2-benzylphenanthrene (2-BP). The numbers indicate heat of formation (kcallmol) calculated by the MOPAC-AM1 method.

dinaphthylmethane was not observed. This may imply that the corresponding unsubstituted dinaphthylmethyl radical tends to form oligomerized products. In the presence of DHA, ipso-hydrogen addition to 2,6DBN, as for the reaction of 1,5-DBN,may also occur to give the product pair of toluene and 2-BN. Thus, it may be considered that, in the reaction of 2,6-DBN alone, dimerization is the energetically favorable process than the cleavage reaction via ipso-hydrogen addition at the early stage. It should be noted that analysis of the gas phase of the reaction mixture revealed evolution of hydrogen, although its quantification was not made. The FD-MS spectrum of the product mixture from 2-BP showed a characteristic signal at m I Z 458 (dimer76) which may corresponds a diphenanthrylmethane as well as that at 444 (dimer-901, this being similar to that for 2,6-DBN (Figure 5). The semiempirical MO calculation could also support that the reaction of 2-BP without DHA proceeds by a similar way t o that of 2,6-DBN (Figure 6). In the reaction of 3-BP, formation of phenanthrene and 3-methylphenanthrene was also observed, while only a trace amount of toluene could be detected. These products could also be formed by secondary reaction of oligomeric materials. The reaction of 2-BA without DHA unexpectedly gave the product pair of toluene and anthracene, while the yields were rather low and hence a relatively large amount of oligomeric products is considered to be formed. This would imply that (a) oligomerization is the major reaction as for the case of 2-BP and 2,6-DBN, but benzene liberation as in Figures 4 and 6 is an energetically less favorable process and (b) in this reaction, ipso-hydrogen addition can occur. There is a possibility that 9,lO-dihydroanthracene derivatives are formed during the pyrolysis and they act as hydrogen donors as well as the substrate itself. The substrates, 1,5-DBN, 9-BP, and 9-BA, may be considered to be a-benzyl-substituted type compounds,

Energy & Fuels, Vol. 9, No. 5, 1995 863

Pyrolysis of Coal Model Compounds Table 4. Difference of Heats of Formation between Substrates and Their ipeo-HydrogenAdducts Estimated by Semiempirical Calculation Using the MOPAC-AM1 Method

AA&

AAHf

substrateQ 1,B-DBN 2,6-DBN 2-BP 3-BP 9-BP

(kcal/molIb 0.4 7.0 4.0 4.4 2.4

substratea 2-BA 9-BA 1-BF 2-BF 4-BF

(kcal/mol)b 0.7 -0.7 8.4 5.2 5.8

a See Figure 1. AHf (ipso-hydrogen adduct at polyaromatic ring) - AHf (the corresponding substrate).

whereas 2,6-DBN, 2-BP, 3-BP, and 2-BA may be classified to p-benzyl substituted compounds. The results of the pyrolysis of these compounds are summarized as follows: (a)the a-substituted type substrates selectively give toluene as the cleavage product along with the corresponding polyaromatic fragments accompanied by formation of oligomeric products. The cleavage reaction possibly proceeds by initial ipso-hydrogen addition, which can be enhanced by addition of the hydrogen source, DHA. (b) The major reaction of the p-substituted type compounds is oligomerization involving radical intermediates formed by hydrogen abstraction from the substrates and benzene may be selectively liberated from the oligomers, with the exception of the reaction of 2-BA. In the presence of DHA, toluene can also be formed, as for the a-type substrates. Although benzylfluorenes cannot be classified to the a- and p-types, the behavior of 2-BF seems to be similar to that of the 0-type compounds. 1-BF and 4-BF also gave benzene in considerable amounts and the yields of the detectable products were relatively low. Thus, the major thermal reaction may also be rather similar to the ,&type compounds. Although from 1-BFand 4-BF toluene was produced considerable amounts, the yield of the corresponding fragment, fluorene, was relatively too small to be rationalized by the initial ipso-hydrogen addition mechanism. It was confirmed that fluorene is stable under the reaction conditions. Thus, in the reaction of 1-BF and 4-BF, toluene could be formed from oligomeric products. However, the reason for the different behavior between 2-BF and 1- and 4-BF is not clear. To obtain an insight into the different behavior between the a- and p-type compounds, the reaction energetics of ipso-hydrogen addition onto each substrate were calculated by MOPAC-AM1 method. The energy changes (AAHf) by the addition to the polyaromatic rings of the substrates are summarized in Table 4. The data suggest that the reactions of the a-type compounds are relatively smaller energy change processes, whereas those for j3-type substrates are relatively more endothermic with the exception of that for 2-BA. The energy changes for the hydrogen additions t o 1-,2-, and 4-BF are also large. Therefore, it may be considered that one of the major origins for the observed different pyrolytic results depending on the substrates examined is the ease of ipso-hydrogen addition.

Experimental Section lH NMR spectra were obtained with a JEOL JMN-GSX400 spectrometer (400 MHz) for CDC13 solutions. GC-MS spectra were obtained with a Shimadzu QP-2000 spectrometer.

FD-MS spectra were obtained with a JEOL JMS-DX-303 spectrometer. GC analysis was carried out on a Shimadzu GC8APF gas chromatograph with a silicone OV-17 column (i.d. 2.6 mm x 1.5 m) and a Shimadzu GC-14A with a CBP-1 capillary column (i.d. 0.5 mm x 25 m). Substrate Preparation. The benzyl-substituted polyaromatic compounds were prepared from the corresponding benzoyl ketones, 1,5-1° and 2,6-dibenzoylnaphthalenes,2-," 3-,12 and 9-benzoylphenanthrenes,132-14 and 9-benzoylanthracenes,14 and 2-benzoylfluorene15 and 1-16 and 4-benzoylflu~renones.~~ They were reduced by using triethylsilane in trifluoroacetic acid according t o the reported method.18 In the case of the anthracene derivatives, 2- and 9-benzyl-9,lOdihydroanthracenes were formed by the reduction. Thus, the dihydro compounds were dehydrogenated by using a P d C catalyst in refluxing 1-hexanol for 16 h to produce the desired materials. The methods used for the preparation of the ketones were reported previously with the exception that for 2,6-dibenzoylnaphthalene.2,6-Dibenzoylnaphthalenewas obtained by chlorination of 2,6-naphthalenedicarboxylicacid with thionyl chloride followed by treatment with benzene in the presence of aluminum chloride. (1,2'-Dinaphthy1)phenylmethane was prepared by the reaction of 2-benzoylnaphthalene with 1-naphthylmagnesium bromide in ether followed by treatment with triethylsilane. The purity of each substrate judged by GC analysis is '99%. The melting point and EIMS and lH NMR data of the substrates were as follows. 1,5-Dibenzylnaphthalene:mp 147-148 "C (lit.19 147-148 "C). MS m l z 308 (M+, 94.9), 217 (loo), 202 (40.01, 91 (27.8). lH NMR 6 4.45 (s, 4 H), 7.18-7.28 (m, 12 H), 7.38 (dd, 2 H, J = 6.8, 8.3 Hz), 7.93 (d, 2 H, J = 8.8 Hz). 2,6-Dibenzylnaphthalene:mp 126-127 "C (lit.20123 "C). MS m l z 308 (M+, loo), 217 (90.2), 202 (33.7), 91 (41.8). 'H NMR 6 4.11 (s, 4 H), 7.19-7.30 (m, 12 H), 7.56 (s, 2 H), 7.70 (d, 2 H, J = 8.3 Hz). 2-Benzylphenanthrene: mp 106-107 "C (lit.12107-108 "C). MS m l z 268 (M+, loo), 252 (20.71, 191 (12.21, 189 (15.8), 133 (15.0), 91 (16.3). lH NMR 6 4.25 (s, 2 H), 7.19-7.32 (m, 5 H), 7.42 (dd, 1 H, J = 8.3, 2.0 Hz), 7.55-7.64 (m, 2 H), 7.67 (t, 2 H, J = 9.3 Hz), 7.80 (d, 1 H, J = 7.8 Hz), 7.86 (dd, 1 H, J = 1.5, 7.8 Hz), 8.50 (s, 1 H), 8.63 (d, 1 H, J = 8.3 Hz). 3-Benzylphenanthrene: mp 77-78 "C (lit.1279-80 "C). MS 268 (M+, loo), 252 (18.6), 191 (12.8), 189 (17.4), 133 (10.9). 'H NMR 6 4.19 (s, 2 H), 7.22-7.31 (m, 5 H), 7.47-7.72 (m, 6 H), 7.86 (dd, 1H, J = 7.8, 1.0 Hz), 8.60 (d, 1H, J = 8.3 Hz),8.63 (d, 1 H, J = 8.3 Hz). 9-Benzylphenanthrene: mp 153-154 "C (1it.l2153-154 "C). MS m l z 268 (M+, loo), 252 (21.71, 191 (13.91, 189 (12.01, 133 (10.3), 126 (14.1). lH NMR 6 4.48 (9, 2 H), 7.18-7.29 (m, 5 H), 7.52-7.64 (m, 5 H), 7.81 (dd, 1 H, J = 7.3, 1.0 Hz), 8.03 (dd, 1 H, J = 8.3, l.OHz), 8.66 (d, 1 H, J = 8.3Hz), 8.72 (d, 1 H, J = 8.3Hz). 2-Benzylanthracene: mp 125-126 "C. MS 268 (M+, 1001, 252 (17.4), 191 (10.3), 133 (19.6). 'H NMR 6 4.17 ( 8 , 2 HI, 7.21-7.33 (m, 6 H), 7.42-7.44 (m, 2 H), 7.76 (s, 1 H), 7.907.98 (m, 3 H), 8.33-8.37 (m, 2 H). (10) Clar, E.; Lovat, M. M.; Simpson, W. Tetrahedron 1974,30,3293. (11)Cook, J. W.; Preston, R. W. G. J. Chem. SOC.1944, 553. (12)Bachmann,W. E., J. Am. Chem. SOC.1935, 57, 555. (13) Bachmann, W. E., J. Am. Chem. Soc. 1934,56, 1363. (14)Tamaki,T. Bull. Chem. SOC.Jpn. 1978,51, 1145. (15)Bachmann W. E.; Xaveria, S. M.; Barton, I. H. M. J.Org. Chem. 1938, 3, 300. (16)Fieser, L. F.; Seligman,A. M. J.Am. Chem.Soc., 1935,57,2174. (17) Nightingale, D.: Heiner, H. E.: French, H. E. J.Am. Chem. Soc. 1950, 72,-1875: (18)West. C. T.: Donnellv. -. _ .S. J.:. Kooistra D. A.:. Doyle, - . M. P. J.O w Chem., 197S,38,2675. (19) Chambers, R. R. Jr.; Collins; C. J.; Maxwell, B. E. J. Org. Chem. 1985,50,4960. (20) Dziewonski, K.; Wodelski, S. Roczniki Chem. 1932, 12, 366; Chem. Abstr. 1933,27, 2145.

Murata et al.

864 Energy & Fuels, Vol. 9, No. 5, 1995 9-Benzylanthracene: mp 131-132 "C (lit.21 133 "C). MS m l z 268 (M+, 1001, 252 (19.81, 191 (32.8). 'H N M R 6 5.01 (s, 2 H), 7.10-7.21 (m, 5 HI, 7.43-7.57 (m, 4 H), 8.01-8.05 (m, 2 H), 8.19-8.23 (m, 2 H), 8.43 (s, 1 HI. 1-Benzylfluorene: mp 102-103 "C. MS m l z 256 (M+,30.11, 178 (16.2), 165 (loo), 91 (14.2). 'H NMR 6 3.73 ( s , 2 HI, 4.12 (s,2 H), 7.13 (d, 1H, J = 7.8 Hz), 7.16-7.21 (m, 3 HI, 7.247.29 (m, 3 HI, 7.33-7.37 (m, 2 HI, 7.49-7.51 (m, 1HI, 7.69 (d, 1 H, J = 7.3 Hz), 7.77 (d, 1 H, J = 7.8 Hz). 2-Benzylfluorene: mp 106-107 "C (lit.22106 "C). MS m / z 256 (M+,43.3), 165 (loo), 91 (11.0). 'H N M R 6 3.84 ( 8 , 2 H), 4.05 (s, 2 H), 7.18-7.36 (m, 9 HI, 7.50 (d, 1 H, J = 7.3 Hz), 7.68-7.75 (m, 2 HI. 4-Benzylfluorene: mp 75-76 "C. MS m l z 256 (M+, 35.81, 178 (20.91, 165 (loo), 91 (10.9). 'H N M R 6 3.94 (s, 2 HI, 4.50 (s,2 H), 7.06 (d, 1 H, J = 7.8 Hz), 7.17-7.29 (m, 8 H), 7.46 (d, 1 H, J = 7.3 Hz), 7.53-7.56 (m, 1 H), 7.76-7.78 (m, 1 H). (1,2'-Dinaphthy1)phenylmethane:mp 126-127 "C. MS m / z 344 (M+,loo), 265 (4351,215 (7721,202 (21.1), 131 (24.5). 'H NMR d 6.43 (s, 1 HI, 6.98 (d, 1 H, J = 7.3 Hz), 7.15-7.17 (m, 2 H), 7.23-7.45 (m, 10 H), 7.65-7.67 (m, 1H), 7.75-7.81 (m, 3 H), 7.85-7.87 (m, 1 H), 8.02 (d, 1 H, J = 8.3 Hz). Pyrolysis. The reaction was performed using a Pyrex glass tube (6 mm (i.d.) x ca. 100 mm). A benzyl-substituted polyaromatic compound ca. 20 mg (and an appropriate quantity of 9,lO-dihydroanthracene) was added to the tube and the (21) Cook, J. W. J. Chem. SOC.1926, 2160. (22) Dziewonski, K, Reicher, Z. Bull. Int. Acad. Polon. 1931, A 643; Chem. Abstr. 1933,27, 283.

content was melted by heating. After cooling, the tube was evacuated and flame sealed. The sample tube was then heated for 1 h by inserting it into a thermostated electric furnace which was preheated at 430 "C. After cooling, it was opened and ca. 2 mL of dichloromethane containing an appropriate internal standard for GC analysis was added to the product mixture. Product identification and quantification were made by GC-MS and GC analysis. Semiempirical MO Calculation. The calculation was carried out on a Titan 750V workstation (Kubota Pacific Computer Co.) by using the semiempirical molecular orbital calculation program, MOPAC-AM1 (version 5.0).23Generally, it is well-known that MO calculation of odd electron systems is especially very difficult, this being probably the source of errors. However, it has been reported that the AM1 method could reproduce the relative order of energy values. Thus, we employed this method for the calculation of the radical intermediates as well as the model substrates and the products.

Acknowledgment. We gratefully thank Osaka Gas Co. Ltd. for the computer facility and reviewers for their highly beneficial comments concerning the mechanistic aspects. EF950041U (23) Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P.

J. Am. Chem. SOC.1985,107, 3902.