74
Energy & Fuels 1991,5, 74-78
preted this as a decrease in TOF with increasing ring density. On the basis of this finding, they concluded that two independent rate processes contribute to carbon gasification: (1) direct colliiion of O2from the gas phase with the edge (reactive) sites, and (2) reaction of the edge carbon atoms with migrating oxygen atoms which were originally chemisorbed on the basal plane. (In a later review paper, Yang32 acknowledged that it is also possible that the surface oxygen was adsorbed on impurities and/or defects and not on the basal plane.) In subsequent analysis of their data, they showed, however, that the "surface rate constant" (i-e.,TOF for the second rate process) does not vary significantly for different ring densities. It is possible that the apparent contradiction (assuming that the TOF for the first process is independent of ring density, as implied by the authors) can be reconciled by reinterpreting their results in terms of RSA decrease, and not necessarily in terms of TOF decrease (see eq 3). Indeed, Yang and Wong31 also reported that the amount of chemisorbed oxygen (after a 10-min reaction) measured in the argon flush experiments (from vacancy enlargement data), which we interpret as measurements of RSA, decreased with increasing ring density. In contrast to the above-mentioned interpretation of the C-O2 reaction mechanism, Yang and Yang33concluded that in the C-C02 reaction all gasification events are caused by direct collisions with the edge (reactive) sites. If migration of oxygen atoms on the carbon surface during (32)Yang, R. T.Chem. Phy.9. Carbon 1994,19,163-210. (33)Yang, K.L.;Yang, R. T. AIChE J. 1985,31,1313-1319.
gasification in C02 is negligible, our TK experiments should indeed titrate the true concentration of reactive sites, as argued and demonstrated in Figure 8. If the chemisorption/diffusion mechanism is in fact operative in the C-O2 reaction, the implication of the result shown in Figure 12 is that the contribution to RSA from the C-O complex formed as a result of the migration of chemisorbed (atomic) oxygen from unreactive to reactive sites was not sufficient to significantly affect the correlation between R and RSA. Transient kinetics studies using isotopes should be instrumental in clarifying this issue. Conclusions The analysis of char gasification in both carbon dioxide and oxygen using the transient kinetics approach gives a direct measurement of the reactive surface area of chars. The use of this technique has also provided a quantitative understanding of char reactivity variations with conversion. Turnover frequencies for char gasification have been determined. Their comparison for different chars in the C-C02 reaction suggests that char gasification may be a structure-sensitive reaction. Acknowledgment. Initial support of this work by the Minerals Research Institute and the Cooperative Program in Coal Research at Penn State University is gratefully acknowledged. Further support was provided by the Gas Research Institute, Contract No. 5086-260-1419. The coal samples were obtained from the Penn State/DOE Coal Sample Base. Registry No. C,7440-44-0; COz, 124-38-9.
Depolymerization of Coals Promoted by Zinc Halides near 100 "C Manjula M. Ibrahim and Mohindar S. Seehra* Department of Physics, West Virginia University, Morgantown, West Virginia 26506 Received May 25, 1990. Revised Manuscript Received August 15, 1990 Thermogravimetry (TG) and in situ electron spin resonance (ESR) spectroscopy of four coals mixed with up to equal amounts of ZnCI2, ZnBr2, and Zn12 are reported for temperatures from 25 to 500 "C. TG measurements show that, between 50 and 175 "C, zinc halides promote nearly 30% mass loss due to release of volatiles and, in ESR spectroscopy, a corresponding increase in the free-radical density is observed in flowing N2 gas. For one case, viz., ZnC12,the effect of different loadings has been measured and the nature of volatiles has been investigated by gas chromatography/mass spectrometry (GC/MS). These observations demonstrate that significant depolymerization of coals promoted by zinc halides occurs at temperatures between 100 and 150 "C and, at 200 "C, the presence of benzene derivatives signifies cracking. These results are important because the earlier studies of coal liquefaction with zinc halides were limited to temperatures 1300 "C. Introduction One of the thrust areas of research in the direct liquefactions of coals is to find a suitable catalyst or a combination of catalysts to effect conversion under milder conditions of temperatures and pressures. Lewis acid1p2and in particular zinc halides3have been used as test catalysts
* Author to whom correspondence should be addressed. 0887-0624/91/2505-0074$02.50/0
for lowering the operating conditions and to increase product selectivity. Struck and Zielke3 reviewed the status of zinc chloride as a catalyst for coal liquefaction in bench scale experiments. These experiments showed that, when used in 1:l coal to zinc chloride ratio, it yielded superior (1)Anderson, L.L.;Miin, T. C. Fuel h o c . Technol. 1986,90,165-174. (2) Nomura, M.; Sakashita, H.; Miyaka, M.; Kikkawa, S. Fuel 1983, 62,73-77.
0 1991 American Chemical Society
Depolymerization of Coals
Energy & Fuels, Vol. 5, No. 1, 1991 75
Table I. Weight 70 of Selected Constituents of the Four Coals Used in These ExDerimentso other volatile matter C H vitrinite exinite macerals coal (drv basis) (man (maD (dmmfl (dmmfl (dmmf) 81.54 5.72 83.9 6.3 9.8 C 36.03
~~
30 mm
k
4
~
~
D E F
~~
25.64 30.68 38.65
83.66 4.95 87.74 5.13 84.36 5.55
22.0 41.0 60.0
10.4 12.6 12.0
67.6 46.4 28.0
1 :?
OFor more complete analysis, see ref 8. Coal C: Manchester Seam, Kentucky. Coal D: Peach Orchard Seam, Kentucky. Coal E Elswick Seam, Kentucky. Coal F: Leatherwood Seam, Kentucky.
product distribution and selectivity. These experiments were done at 420 "C and 17.3 MPa hydrogen pressure. There are a number of other studies available on the catalytic activity of zinc halides in coal liquefaction and in various model compound^.^^ However, these studies were done either at high pressures of hydrogen or at temperatures above 300 "C. A disadvantage of reacting coal at elevated temperatures (besides additional costs) is that thermal cracking makes substantial contributions to the reaction and there is a loss of full catalytic control and with it the gains in selectivity it can offer. On the other hand, at lower temperatures, a good balance of hydrogenation and cracking can be achieved to get better products? Hence it is important to find whether zinc halides have any depolymerization activity at lower temperatures. In this paper, we report clear evidence of depolymerization of coals with zinc halides at the mild temperatures of 100-150 "C. This evidence of low-temperature depolymerization comes from thermogravimetric (TG) and in situ electron spin resonance (ESR) experiments and from the analysis of the volatiles by gas chromatography/mass spectrometry (GC/MS). With TG, we observe a nearly 30% mass loss between 50 and 175 OC when coals are mixed with zinc halides (ZnClz,ZnBrz, ZnIJ and, in ESR spectroscopy, a corresponding increase in the free-radical density is observed in flowing N, gas. In the case of one coal mixed with ZnClz,analysis of the volatiles evolved at the temperatures of 100,150, and 200 "C carried out using GC/MS shows the presence of both aromatic and aliphatic compounds in addition to gases such as HzO, COz, and NH3. Results of these experiments are presented below.
Experimental Section In Table I, we list the four Kentucky coals used in our experiments dong with some of the important information on their analysis. These are the same coals used in our recent work on the comparison between liquefaction yields and TG/ESR parameter@ and a more complete listing of their proximate, ultimate and petrographic analysis is given in ref 8. The three halides, viz., ZnC12, ZnBr,, and ZnI,, were obtained from Alfa Products Inc. and were used as obtained. The coals and halides, taken in the appropriate ratios, were dry-mixed in a mortar in a glovebox just prior to the experiments in order to avoid excessive exposure to air. T G measurements were carried out in a Mettler system (Model TA3000) using 10 mg of the samples and a heating rate of 10 "C/min and N2gas flow rate of 100 cm3/min. For in situ ESR experiments up to 600 "C, an X-band (-9 GHz) reflection-type spectrometer with a TElo2mode cavity in which heat is transferred to the sample by flowing N2gas was (3)Struck, R. T.;Zielke, C. W. Fuel 1981, 69,796-800. (4)Mobley, D.P.;Bell, A. T. Fuel 1979,58, 661-666. (5)Yokono,. T.: . Ivama. _ . S.:. Sanada.. Y.:. Makino, K. Fuel 1985.. 64.. 1014-1016. (6)Salim, S.S.;Bell, A. T. Fuel 1982, 61,745-754. (7) Sulimma, A.;Leonhardt, P.; van Heek, K.H.; Juentgen, H. Fuel 1986,65, 1457-1461. (8) Ibrahim, M. M.; Seehra, M. S.; Keogh, R. Fuel Process. Technol. 1990,25, 215-226.
Outlet lor
50 mm
Pytex tube
Sample M 3 mm
Figure 1. Schematic diagram (not to scale) of the flow chamber used in the ESR experiments. The joint a t A allows the Pyrex ESR tube to be adjustable in height below A.
'\\ .L .L
K
(ot 20'
0
\
F+ ZnI,
'4: \
'
'
100
'
'
200
' 300
'
' 400
'
'
500
e )O
T("C) Figure 2. Remaining weight of the sample plotted as percentage against temperature. Coal F and the zinc halides are mixed in the 1:1 ratio. Correction for the weight loss due t o the halides is applied. used. Details of this system and procedures have been reported earlier. *11 The free-radical density N is calculated by double integration of the derivative ESR line using an on-line computer and comparing it with a standard.1° All our earlier experiment&ll were carried out on samples which were vacuum sealed in ESR tubes. To cany out experiments in a flowing gas, the flow chamber shown in Figure 1was used. ESR measurements were done from room temperature to 450 "C in the presence of nitrogen or hydrogen gas a t a flow rate of 10 cm3/min. The nature of volatiles was analyzed by GC/MS experiments (Finnigan Model 4021) using helium as carrier gas and a CSQ gas (9)Seehra, M.S.;Ghosh, B.; Mullins, S. E. Fuel 1986,65, 1315-1316. (10)Seehra, M. S.; Ghosh, B. J. Anal. Appl. Pyrolysis 1988, 13, 209-220. (11) Seehra, M. S.; Ghosh, B.; Zondlo, J. W.; Mintz, E. A. Fuel Process. Technol. 1988, 18, 279-286.
Ibrahim and Seehra
76 Energy & Fuels, Vol. 5, No. 1, 1991 100
-8 v
-
90-
E .-
80-
3 0,
.-c 70 .-C
2
a 0
100
200
300
400
500
600
T("C) Figure 3. Remaining weight as percentage for coals C, D, E, and F of Table I plotted against temperature when mixed with ZnClz in the 1:l ratio. column. Coal C mixed with equal amount of ZnClz was heated a t the rate of 10 OC/min in situ in a pyroprobe. The sample was held at each required temperature, viz. 100,150, and 200 OC, for about 10 min and the evolved products were analyzed.
Experimental Results a. TG Experiments. The changes in the weights of zinc halides, and coal F mixed with zinc halides in the ratio 1:l as a function of temperature and under the flow of nitrogen gas, are shown in Figure 2. ZnC12and ZnBr2are quite stable for temperatures up to 500 "C whereas Zn12 shows a rapid weight loss above 350 "C probably because of thermal decomposition. The small losses observed for ZnC12and ZnBr2 and for Zn12 below 350 "C are probably due to the release of adsorbed water. These losses are taken into consideration when the weight loss of coal in the presence of a halide is calculated. In Figure 2, we also show the weight loss of a coal F (after correcting for the weight loss of the halide) when mixed with ZnC12, ZnBr2, and Zn12 in the ratio of 1:l by weight. In the presence of the halides, coals begin to lose weight beginning near 50 "C and by 150 "C with Zn12and 175 "C with ZnC1, and ZnBr2 between 20 and 30% of the weight is lost. This weight loss is nearly equal to the percentage of volatile matter in these coals (see Table I). To demonstrate that the phenomenon of weight loss in coal F below 200 "C in the presence of halides is not restricted to one coal only, we carried out experiments with the remaining three coals of Table I also. The results obtained by using ZnClz shown in Figure 3 demonstrate that all four coals show similar behavior. Although the weight loss and the temperature at which the first stage of the weight loss is completed varies slightly from coal to coal, the observations are qualitatively similar in all coals. The weight loss occurring above 350 "C is probably due to thermal cracking as it is also observed in coals without the presence of zinc halides. To measure the effect of halide loading on the weight loss, we show in Figure 4 the remaining weight of coal C as a function of temperature for different loadings of ZnC1,. (The percentage loading is defined as (weight of ZnC12/ weight of coal) X 100.) The weight loss of coal at 175 "C vs percentage loading is shown in Figure 5 and it shows that the weight loss is nonlinear, approaching saturation for 100% loading. In Figure 6 we have plotted the rate of mass loss normalized to the room temperature mass viz. (dm/dt)/m, vs temperature for coal C and for coal C mixed with the
60 -
50L
0
'
'
100
'
'
200
'
'
300
'
'
400
'
'
500
'
600
T("C) Figure 4. Remaining weight of coal C for different loadings of ZnCl,, plotted against temperature. 100% loading represents 1:1 mixing.
30.
0
20
60
40
80
100
120
% ZnCl,
Figure 5. Weight 108s of coal C at 175 OC vs % ZnCl,. 100% loading represents 1:l mixing of coal and ZnC12 The solid line is drawn as a guide for the eye.
I
t 0
100
200
300
400
500
600
T("C) Figure 6. The measured reaction rate (dm/dt)/mo for coal C and for coal C mixed with halides in the 1:l ratio, plotted against temperature. Note the peak in the reaction rate near 150 "C due to halides.
different halides at 100% loading. There is a peak in this reaction rate also at temperatures between 150 and 175 "C in the presence of the halides and there is no change in this quantity in this temperature range for coal alone.
Depolymerization of Coals
Energy & Fuels, Vol. 5, No. 1, 1991 77
C + ZnC12(N,)
C
-0 vO
100
200
300
400
500
600
T(*C) Figure 7. The free-radical density N (after correcting for the Curie law contribution) plotted against temperature for coal C and for coal C mixed with the halides in the 1:l ratio. The experiments were carried out in flowing Nz gas.
Recent measurements have shown8that this reaction rate bears a good correlation with percentage conversion to liquid products. b. ESR Experiments. ESR spectra of all four coals at room temperature consist of a broad and a narrow component, reflecting the different macerals present in these coals. Following the studies of Retcofsky et al.12and Silbernagel et al,13we infer that the broad component is due to vitrinites and exinites and the narrow component is due to inertinities. In our earlier ESR studies on the pyrolysis of coa1s,8-11we have reported on the changes in the g values, line widths, and free-radical density N as a function of temperature. The major changes occur in N only (after correcting for the Curie law variation) and are well represented by the data on coal C in Figure 7 . Here there is an increase in N from 25 to -200 "C (stage l), a decrease in N as temperature increases to -400 "C (stage 21, and a rapid rise in N above 400 "C (stage 3). In some coals a fourth stage is observed at still higher temperatures in which N again decreases with increasing temperatures due to the repolymerization of free radicals.1° The increase in N above 400 "C (in stage 3)is due to thermal cracking as it is also evident in the TG experiments of Figure 4 (0% loading). Fowler et al.14 have recently reported similar observations in some British coals. The data of N vs T for coal C mixed with zinc halides at 100% loading is also shown in Figure 7. It is evident that, above about 150 "C, the magnitudes of N are considerably higher in the presence of halides, and the four stages mentioned above and discussed in more detail in earlier publications9J0J4are shifted to lower temperatures. The rapid drop in N above about 400 OC is due to broadening of the ESR line and no attempt was made to determine the lower limit of the signal, although the spectrometer is able to detect N = 10l6/g.If we compare the results of Figure 4 with results of Figure 7 and look at the differences between the observations on coal C (0% (12) Retcofsky, H. L.; Stark, J. M.; Friedel, R. A. Anal. Chem. 1968, 40, 1699-1704.
(13) Silbernagel,B. G.; Gebhard, L. A.; Dyrkacz, G. R.; Bloomquist, C. A. A. Fuel 1986,65,55&585. (14) Fowler, T. G.; Bartle, K. D.; Kandiyoti, R. Fuel 1987, 66, 1407-1412.
100
-i\/
+ ZnCI,(H2) 0-4 200
300
400
500
600
T("C) Figure 8. Comparison of the effect of flowing Nzgas and flowing
Hzgas on the free-radical density at different temperatures for coal C and coal C mixed with ZnC1, in the 1:l ratio. Table 11. Volatiles Obtained from the Pyrolysis of Coal C Loaded with 100% ZnC1," 100 "C 150 O C 200 O C carbon monoxide ammonia benzene carbon dioxide furan l,l,2,3-tetramethylcyclopropane 2-propanone nonane water 2-methylbutane methylbenzene ammonia 1-chloropentane ethylbenzene methane 2-pentene chloromethane butanol methanol 2-butanone ethylene oxide
"The sample was heated in a pyroprobe and the volatiles evolved at 100, 150, and 200 O C were identified by using GC/MS.
loading) and coal C with 100% ZnC12 loading, then a consistent picture emerges. In the presence of ZnC12,the mass loss between 100 and 200 "C is accompanied by an increase in N , between 200 and 300 "C, a negligible mass change is accompanied by a decrease in N, between 300 and 400 "C a further mass loss is accompanied by a rapid increase in N, and above 400 "C, the rate of mass loss decrease is accompanied by a decrease in the free-radical density N . From the point of view of this work, the most crucial observation is an increase N (relative to coal alone) between 100 and 200 "C since this coincides with the rapid mass loss observed here for the first time in the same temperature range. The earlier reported liquefaction experiments with zinc halides as catalysts1$were carried out at temperatures above 300 OC. Later we describe some initial experiments using GC/MS which clearly demonstrate that, between 100 and 200 "C, depolymerization/ cracking of coal C loaded with 100% ZnC12is indeed taking place. The effect of flowing hydrogen gas on N is shown in Figure 8 where we have plotted N vs temperature for coal C in N2flow, coal C + ZnC12 (100% loading) in N2flow, and coal C + ZnC12(100%loading) in Hzflow. It is evident that the effect of H2is to quench some of the free radicals and lower the observed density N . c. GC/MS Experiments. The nature of volatiles was analyzed by using GC/MS experiments in which coal C loaded with 100% ZnC12was heated to 100,150,and 200 "C successively as described in the Experimental Section. In Figure 9 we show the time scan of the evolved products where the evolved products from coal C alone are also
Zbrahim and Seehra
78 Energy & Fuels, Vol. 5, No. 1, 1991 100
I
C(l00"C)
I
C(15O'C)
I
I
I
50i
h
100
v)
1
'= 4-
A
C
NH,
C
"
+ ZnCI,
CY
U6"SV"3
C6H,CH,CH,
loo^^ c( 100OC) + ZnCI,
CH3OH
50
C(2OO0C)
CH&I
CH2CH,0 NH3
0 4
8
12
TIME (m) Figure 9. GC traces of volatiles evolved on heating coal C mixed with ZnC12 to temperatures of 100, 150, and 200 OC. For comparison, data for coal C alone is also shown. Identification of volatiles was carried out by GC/MS. The intensities are normalized to 100% with the highest peak and the data for pure coal is normalized to that for coal + ZnClz.
shown for comparison. Each of the masa fractions was then analyzed by the mass spectrometer, and in Table 11, we list some of the identified products which include both aromatic and aliphatic compounds in addition to gases such as H20, CO,, and NH3. It is noted that the evolved fractions are quite complex, and we have used the available library of compounds for matching. However, because of the complexity of the fractions, the matching is not always unique and not all the components could be identified. It is for this reason that we have not attempted to quantify the evolved compounds in our listing in Table I1 at the present time, although we are quite certain of the identification of the compounds listed in Table 11.
Discussion As noted earlier, the earlier studies with zinc halides as catalysts were generally limited to temperatures above 300 "C. For example, Jolly et al.15 studied the evolution of H, and CH, in the presence of ZnCl, at temperatures 1350 "C and observed an increase in the amount of these evolved gases when loading of ZnCl, was increased from 1% to 25%. The measurements of Yokono et a1.16 using in situ ESR spectroscopy and a pressure of 10 MPa showed that, in the presence of nitrogen, ZnC12affects an increase (15) Jolly, R.; Charcosset, H.; Boudou, J. P.; Guet, J. M. Fuel Process. Technol. 1988.20. 51-60. (16) Yokono, T.; Iyama, S.; Sanada, Y.;Shimokawa, S.; Yamada, E. Fuel 1986,65, 1701-1704.
in N, whereas, in the presence of H2,N decreases. These observations were, however, made at temperatures above 300 "C whereas we report here similar results at lower temperatures. From a study of the model compounds in the presence of a donor solvent (tetralin) and hydrogen at high pressures, Vernon" concluded that both molecular hydrogen and a source of free radicals, generated either catalytically or thermally, are necessary for good conversion. Our results show that zinc halides are effective in promoting an increase in free radicals, and with the uptake of H,, the density of free radicals decreases. The mechanism of cleavage of linkages and cracking of aromatic structures by zinc halides has been investigated by Bell et al.3J8using model compounds. According to these studies, scission of bonds in model compounds by zinc halides eventually leads to the formation of a carbonium ion and a smaller hydrocarbon. The structure of coals is, however, quite complex, and as our results using GC/MS show, a variety of compounds are liberated in the low-temperature catalytic depolymerization reported here. A simple carbonium ion mechanism is not likely to produce an increase in the free-radical density as observed here as well as in the work of Yokono et a1.16 A more detailed analysis of the products of the reaction may eventually lead to a better understanding of the mechanism of the lowtemperature catalytic depolymerization of coals by zinc halides reported here. We have plans to undertake this work in the near future.
Concluding Remarks The results presented here clearly demonstrate that significant depolymerization of coals promoted by zinc halides is taking place in the range of 100-150 "C. Further, the presence of benzene and its derivatives at 200 "C clearly point to the cracking of the coal structure. The practical implications of these results are quite obvious if direct liquefaction experiments using zinc halides as catalysts could show significant conversions at these low temperatures. It is hoped that this work will stimulate such experiments because, at these lower temperatures, the corrosive nature of zinc halides3may not pose as severe a practical problem as encountered at higher temperatures. Acknowledgment. This work was supported in part by grants from the U.S. Department of Energy through the Consortium for Fossil Fuel Liquefaction Science. We thank R. A. Keogh, Center for Applied Energy Research, University of Kentucky, for providing the coal samples used in this study and Mr. R. R. Smith for assistance with the GC/MS experiments. Registry No. ZnClz, 7646-85-7; ZnBrz, 7699-45-8; Zn12, 10139-47-6; CO, 630-08-0; COP, 124-38-9;H20, 7732-18-5; NHS, 7664-41-7;CH,, 74-82-8; ClCH3, 74-87-3;OHCH3, 67-56-1; ethylene oxide, 75-21-8; furan, 110-00-9; 2-propanone, 67-64-1; 2-methylbutane, 78-78-4; 1-chloropentane, 543-59-9; 2-pentene, 109-68-2; 1-butanol, 71-36-3; 2-butanone, 78-93-3; benzene, 71-43-2; 1,1,2,3-tetramethylcyclopropane, 102653-33-8;methylbenzene, 108-88-3; ethylbenzene, 100-41-4. (17) Vernon, L. W. Fuel 1980,59, 102-106. (18) Frederick, T. J.; Bell, A. T. J. Catal. 1984,87,196-209,210-225, 226-237.
