Pyrolysis of coal at high temperatures - Energy & Fuels (ACS

Elmer B. Ledesma, Chun-Zhu Li, Peter F. Nelson, and John C. Mackie. Energy & Fuels 1998 12 (3), ... Ben D. Holt and Teofilo A. Abrajano. Analytical Ch...
1 downloads 0 Views 1MB Size
Energy & Fuels 1988,2,391-400

391

Pyrolysis of Coal at High Temperatures? Peter F. Nelson,***Ian W. Smith,* Ralph J. Tyler,t and John C. Mackies CSIRO Division of Coal Technology, P.O.Box 136, North Ryde, New South Wales 2113, Australia, and Department of Physical Chemistry, University of Sydney, New South Wales 2006, Australia Received December 7, 1987. Revised Manuscript Received March 15, 1988

Pyrolysis of coal and of tar produced by the rapid pyrolysis of coal has been studied in small fluidized-bed reactors and in a shock tube. Kinetic parameters determined for the formation of light hydrocarbon gases from the rapid pyrolysis of coal suggest that rate-limiting coupled processes, such as internal mass transfer, have an influence on the formation rates. However, when this limitation is removed, by studying the gas-phase cracking of tar free from the influences of the original coal or char, only the light olefins that apparently arise from a single type of functional group give activation energies indicative of bond-breaking reactions. Other products, which appear to have several different chemical sources, exhibit low activation energies of formation (110-140 kJ/mol). Results of a detailed study of tar composition as a function of temperature in the fluid-bed reactor were consistent with the proposal that long-chain polymethylene groups are the source of the olefins and that secondary cracking reactions of the tar are an important source of simple aromatics, polycyclic aromatic hydrocarbons, and carbon oxides. Results are also presented to show the effect of preparation conditions on the combustion kinetics of chars produced by rapid pyrolysis. At combustion temperatures of 700 "C, the reactivities of the chars show an inverse relationship to their preparation temperatures, but at higher combustion temperatures, these differences largely disappear.

Introduction Pyrolysis at high heating rates (>lo3 K/s) is often the initial step in the utilization of coal by combustion, gasification, or liquefaction. This involves the thermal decomposition of the coal's organic structure and the release of volatile products, which may account for up to 70% weight loss of the coal. Knowledge of the behavior of the many volatile species liberated during pyrolysis is important, since their composition, rate of release, and secondary reactions will have an important influence on such practical considerations as ignition, rate of combustion, and trace gaseous and particulate emissions. The rate of coal pyrolysis, variously described by measurements of total weight loss or of product formation as a function of time, has been extensively studied.'-4 However no consensus on the rate of coal pyrolysis has emerged from these studies. Variations in rates for weight loss determined by various workers are large enough to have a significant impact on models developed for coal combustion in practical systems. Most recent studies of rapid pyrolysis of coal have been limited to heating rates less than lo5 K/s and used coal particles tens of micrometers in size and larger,'-* conditions under which significant reactions may occur during initial heating of the particles. Consequently it has been difficult to determine rates of evolution of individual volatile species and their secondary gas-phase reactions. In the present work coal pyrolysis was investigated by using a shock tube with heating rates of lo7 K/s, which enabled the devolatilization kinetics to be effectively decoupled from particle heat-up effects. The kinetics of the formation of light hydrocarbon gases and carbon oxides were determined. 'Presented a t the Symposium on Coal Pyrolysis: Mechanisms and Modeling, 194th National Meeting of the American Chemical Society, New Orleans, LA, August 31-September 4, 1987. CSIRO Division of Coal Technology. University of Sydney.

*

0887-0624/88/2502-0391$01.50/0

For many coals, tars represent the major initial species released during p y r ~ l y s i s . They ~ ~ ~ have been suggested as an important source of soot during coal c o m b ~ s t i o n ' ~ ~ and have potential as models of coal structure. The yields and nature of these tars depend not only on coal type but also on pyrolysis conditions, including particle heating rate, reactor residence time, nature of the gaseous atmosphere, and pressure.lv4 Tar release occurs at relatively low temperatures, and in many reaction systems, the tar has ample opportunity to undergo secondary vapor-phase reactions resulting in the formation of hydrocarbon gases, soot, and modified tar. In the present study, the kinetics of the formation of light hydrocarbon gases and CO from secondary reactions of the tar were determined by linking a fluid-bed pyrolyzer, operating at a relatively low temperature for tar production, to the shock tube. The compositions of the tars produced from pyrolysis of a subbitminous coal in a fluid-bed pyrolyzer were also determined in detail. Previous work on tar released during rapid heating experiments has either treated the tar as a single entity and investigated the kinetics of its formation as a function of experimental parameters such as final temperature, heating rate, or p r e s s u ~ e ~ or , ~determined ,~J~ (1)Howard, J. B.in Chemistry of Coal Utilization, Secondary Supplementary Volume; Elliott, M. A., Ed.; Wiley: New York, 1981;pp 665-784. (2)Gavalas, G. R. Coal Pyrolysis; Elsevier: Amsterdam, 1982. (3)Solomon, P. R.; Hamblen, D. G. Prog. Energy Combust. Sci. 1983, 9,323-361. (4) van Heek, K. H. Chem.-Ing.-Tech. 1983,55,777-784. (5)Tyler, R. J. Fuel 1979,58,680-686. (6)Tyler, R. J. Fuel 1980,59,218-226. (7)Seeker, W. R.;Samuelsen, G. S.; Heap, M. P.; Trolinger, J. D. In Proceedings of the 18th Symposium (International)on Combustion; The Combustion Institute: Pittsburgh, PA, 1981;pp 1213-1226. (8)McLean, W. J.; Hardesty, D. R.; Pohl, J. H. In Proceedings of the 18th Symposium (International) on Combustion; The Combustion Institute: Pittsburgh, PA, 1981;pp 1239-1248. (9)Freihaut, J. D.;Zabielski, M. F.; Seery, D. J. In Proceedings of the 19th Symposium (International) on Combustion; The Combustion Institute: Pittsburgh, PA, 1982;pp 1159-1167.

0 1988 American Chemical Society

Nelson et al.

392 Energy & Fuels, Vol. 2, No. 4, 1988 Table I. Experimental Conditions in Shock Tube and Fluidized-Bed Reactors shock tube fluidized bed -106 + 90 particle size, pm . 0 1

400

600

800

1000

Temperature I'C 1

Figure 12. Yields of three- and four-ringaromatics from pyrolysis of Millmerran coal in fluid-bed pyroly~er.'~ source of COz that manifests itself at low temperatures. Carboxylic groups are a possible candidate since they are known to decompose at low temperature to produce COP. Methoxy groups, which may be present in a low-rank coal such as this, will decompose at temperatures above 650 "C to produce C0.48 From studies of the thermal cracking of phenol described above, decomposition of hydroxyl groups at temperatures above 700 "C is a source of the CO produced. However, there are probably other fQnctionalgroups such as furans that contribute to the CO yield at high temperatures, and some CO derives from oxygen retained in the char. The results for CO and COz obtained for this subbituminous coal are similar to those obtained previously for a brown coal (Figure 5)" but differ significantly from those for slow heating of brown c0a150*51 where significant yields of water were observed. Possible reasons for this difference have been discussed previo~sly)~ and a change in the decomposition mechanism of the coal was concluded to be the most likely explanation. This phenomenon is obviously worthy of further investigation as it implies that a Bignificant proportion of the coal can be effectively gasified simply by rapid heating. (d) Polycyclic Aromatic Hydrocarbons. The predominant components of the tar at the highest temperatures studied were unsubstituted polycyclic aromatic hydrocarbons (PAH) with up to five rings. Mass balance calculations at these temperatures showed that at least 50% of the tar was G€ volatile, consistent with an increase in volatility due to cracking of aliphatic species and a decrease in polarity due to loss of oxygen functionality (particularly phenolic groups, see previous section). Results for three- and four-ring aromatic compounds are presented in Figure 12. Most of the aromatics, particularly the four-ring systems, increased simificantly in yield at temperatures above 800 "C. An exception was fluorene, which decreased slowly above 850 "C. However, most species increased in yield and, since the total yield of tar was decreasing, the polycyclic aromatics became increasingly important compo(48)Mackie, J. C.; Doolan, K. R.; Nelson, P. F. J. Phys. Chem., in nrDPCl ~ - - ~ ~ . (49)Cliff, D.I.; Doolan, K. R.; Mackie, J. C.; Tyler, R. J. Fuel 1984, 63,394-400. (50)Schafer, H. N. S. Fuel 1979,58,667-672. (51)Schafer, H.N. S. Fuel 1979,58,673-679.

nents of the tar at the highest temperatures studied. It is thus possible under the heating conditions of the fluid-bed reactor to form complex PAH species purely from pyrolysis of Millmerran coal but only after the aliphatic and oxygen-containing functional groups have been reduced by cracking reactions. The aromatics produced are very similar to those produced from the combustion of with acenaphthylene, phenanthrene, anthracene, fluoranthene, benz[e]acenaphthylene, ppene, and chrysene major components in both cases. The source of the polycyclic aromatics is not completely clear. Pyrolysis of simple aromatic species such as toluene can give rise to complex mixtures of polycyclic species.53@ On the other hand, flames of acetylene and benzene,55 ethylene and acetylenets benzene,s7and butadiene* have been shown to produce polycyclic aromatic species. Indeed the relative concentrations of polycyclic aromatics observed in a near-sooting low-pressure benzene-oxygen-argon flame5' are remarkably similar to those observed in the present experiments. In the flame a mechanism involving addition reactions of aromatic radicals (predominantly benzyl and phenyl) to unsaturated aliphatics such as acetylenic species and stabilization of the adduct by the formation of six-membered rings is postulated to occur.57 Such reactions are used to account for the large amounts of phenylacetylene and indene observed in the flame. Phenylacetylene, styrene, and indene are also identified as important intermediates in PAH formation in the ethylene and acetylene flames.56 Yields of these species produced in the present experiments are shown in Figure 13. Significant yields were obtained at temperatures greater than 700 "C; however, a maximum was observed at 850 "C, and yields decreased at higher temperatures. Thus pyrolysis and combustion of a wide range of starting materials including aliphatic hydrocarbons may give rise to complex mixtures of PAH. In a recent review of the formation of PAH and soot in combustion, Homann59has shown that the relative concentrations of the thermodynamically stable PAH are similar irrespective of fuel type and that the initial steps in the formation can (52)Lee, M. L.;Prado, G. P.; Howard, J. B.; Hites, R. A. Biomed. Mess Spectrom. 1977,4,182-186. (53)Badger, G.M. Bog. Phys. Org. Chem. 1965,3,1-40. (54)Smith, R. D.J. Phys. Chem. 1979,83,1553-1563. (55)Homann, K.H.; Wagner, H. G. In Proceedings of the 11th Symposium (International) on Combustion; The Combustion Institute: Pittsburgh, PA, 1967;pp 371-379. (56)Crittenden, B. D.;Lohg, R. Combust. Flame 1973,30,359-368. (57)Bittner, J. D.;Howard, J. B. In Proceedings of the 18th Symposium (Intermtioml) on Combustion; The Combustion Institute: Pittsburgh, PA 1981;pp 1105-1116. (58)Cole, J. A.; Bittner, J. D.; Longwell, J. P.; Howard, J. B. Combust. Flame 1984,56,51-70. (59)Homann, K.H. In Proceedings of the 20th Symposium (International) on Combustion; The Combustion Institute: Pittsburgh, PA, 1984; pp 857-870.

Pyrolysis of Coal at High Temperatures

Energy & Fuels, Vol. 2, No. 4, 1988 399

I

I

, Y\,

I

I qbI

"-'I\

10-2 O'

'Ob0

3000 Wovcnum bcrr

ZQOO

(ooo

1cm-l)

Figure 14. Diffuse reflectance IR spectrum of Millmerran tar produced at 908 "C. Light ends (bp e200 "C) were removed.

10-3

1

COAL PYROLYSIS

-

L-

H C GAS - - - - -CH&zH2 C 0 , C C p H2O

COAL

~ u 4 ~ c 3 H 6 t

pdymelhylen~~alkyl~romst~c~ groups phenol8 I furnnlc groups

TAR

i

co

__-________ t-PRIMARY-

---co,

-

SECONDARY-

-TERTIARY

L

400

103/1 IK-')

Figure 16. Combustion reactivity of subbituminouscoal char: (a) as a function of char preparation (i.e. pyrolysis) temperature (p, is the combustion reaction rate (g/cm2s));(b) by various techniques, showing effect of heating time. (R,is the combustion rate coefficient per unit external surface area (g/cm2satm)). Reactivities: (-) measured in a flow reactor;M(- - -) measured in a single particle reactor by ignition temperature measuremenGm (X) measured in single-particlereactor by particle temperature

\

f

800

600

Recycler

I

mea~urement.~~

H2

1000

Tempaature ('Ct

Figure 15. General model for coal pyrolysis.

involve addition of small aliphatic species, such as acetylene, to aromatic radicals. Given the similarity of product distributions observed in this study and in the flame studies, it seems likely that analogous reactions are contributing to PAH formation in coal pyrolysis. The IR spectra of tars produced at high temperatures provide additional evidence for this postulate. Figure 14 shows a diffuse reflectance spectrum of tar produced at 908 "C.The light ends boiling at temperatures below 200 "C have been removed, leaving only compounds with three or more condensed aromatic rings. The product shows strong aromatic absorption. However, it also shows a feature at 2200 cm-' that is not normally present in the IR spectra of co@ sal' or tars.13 This region of the spectrum is, indeed, characteristic of the stretching frequency of carbon-carbon triple bondss1 suggesting that the tars contain addition products of acetylenic species and aromatics. On the basis of these results and the results of the secondary cracking studies described above, it is possible to formulate a general model (Figure 15) for some of the chemical processes occurring in coal pyrolysis. It should be emphasized that this is based on work with a small number of low-rank coals and should only be applied to (60) Painter, P. C.; Starsinic, M.; Coleman, M. M. In Fourier Transform Infrared Spectroscopy; Fenaro,J. R., Basile, L. J., Eds.; Academic: New York, 1986; Vol. 4, pp 169-241. (61).Silverstein, R. M.; Bassler, G. C. Spectrometric Identification of Organtc Compounds, 2nd ed.; Wiley: New York, 1967.

other coal types with caution. Three regimes of pyrolysis are identified, and these occur for residence times of 0.5-1 s at the temperatures shown in Figure 15. Primary pyrolysis involves an initial formation of hydrocarbon gas, carbon oxides and water, tar, and char. The sources of the hydrocarbon gas and carbon oxides are labile functional groups in the coal such as methoxy and carboxylic groups. At temperatures above 600 "C, secondary reactions of the tar occur and are the most important source of additional hydrocarbon gas. The first groups to decompose are the long-chain polymethylenes, which yield light olefins, particularly ethylene and propylene. These reactions will probably be less important for higher rank coals. At temperatures above 600 OC, yields of simple aromatics increase as the polymethylene disappears. The alkyl- (mainly methyl-) substituted aromatics decompose a t temperatures above 700 OC to give CHI and the parent aromatic, and phenols decompose to give CO and hydrocarbon gas. Finally a t temperatures above 800 OC a tertiary stage is reached in which the products of the secondary reactions themselves react. PAH precursors such as acetylene, phenylacetylene, styrene, and indene are formed, and these lead eventually to PAH and soot. The char will also be a source of the CO and H2 produced at high temperatures due to cross-linking reactions. Char Reactivity. The combustion reactivity of the three chars determined in the flow reactor (production temperatures 540,600, and 800 "C respectively) is given in Figure 16a. At a combustion temperature of -700 "C, the reactivities of the chars show an inverse relationship to their preparation temperature: the lower the preparation temperature the higher the reactivity.62 At 1000 OC (62) Young, B. C. In Proceedings of the International Conference on Coal Science; Verlag Gluckauf Essen, FRG, 1981; pp 260-265.

400

Energy & Fuels 1988,2, 400-405

these differences have largely disappeared. The question then arises as to the relative contributions to the observed reactivity made by the consumption of the volatile and solid components of the char. The burningrate data in Figure 16a were determined by using a flow reactor. The particles and hot gas entering the reactor were mixed some distance ahead of the positions in the reactor where burning rate measurements were made. There is some indicatione2 that the volatile matter is evolved rapidly early in the reactor: the burning-rate data are for the consumption of the solid char after the prior release of the volatiles. Support for this view is given by Rybak et al.,20who determined the reactivity of the 600 OC char from a measurement of particle ignition temperature and found that the ignition temperature was affected by the volatile content of the char. The more times the char was cycled through the heated ignition reactor (in the absence of oxygen), the higher was the ignition temperature ultimately

measured. Figure 16b shows that the reactivity of the char decreases with increasing heating time (increasing number of cycles through the reactor) in a manner qualitatively similar to the reactivity change with pyrolysis temperature shown in Figure 16a. After eight cycles the reactivity is close to that measured in the flow reactor and in a quite different reactor at the Sandia L a b ~ r a t o r i e s . ~ ~

Acknowledgments. This work was supported by the National Energy Research, Development and Demonstration Council and the Australian Research Grants Scheme. Redst4 NO.CHI, 7482-8; CzH4,74-85-1; CZH2,7486-2; C&, co, 630-08-0; coz, 124-38-9;C6H6, 71-43-2.

115-07-1;

(63) Mitchell, R. E.; McLean, W. J. In Proceedings of the 19th Symposium (International) on Combustion; The Combustion Institute: Pittsburgh, PA, 1982; pp 1113-1122.

Time-Resolved Pyrolysis Mass Spectrometry of Coal: A New Tool for Mechanistic and Kinetic Studied Tanmoy Chakravarty,t Willem Windig,$ George R. Hill, and Henk L. C. Meuzelaar* Biomaterials Profiling Center, University of Utah, 391 South Chipeta Way, Suite F, Research Park, Salt Lake City, Utah 84108

M. Rashid Khan Morgantown Energy Technology Center, US.Department

of Energy, P.O.Box 880,

Morgantown, West Virginia 26505 Received December 7, 1987. Revised Manuscript Received April 3, 1988

A micropyrolysis experiment in combination with multivariate analysis was used to describe the devolatilization behavior of a hvAb Pittsburgh No. 8 coal (obtained from Argonne National Laboratory). The experimental conditions were such as to minimize heat- and mass-transfer effects or secondary reactions. The method described here enables extraction of multiple chemical components from a single coal pyrolysis experiment, thus reducing the uncertainty due to varying reaction conditions in different experiments. Three different types of thermal behavior, namely "desorption", "depolymerization", and "thermal degradation" can be observed in time. This points to the need for a kinetic model with three different reaction order terms.

Introduction Most coal devolatilization studies to date have focused on the determination of reaction rates for reactions occurring under widely different conditions encountered in liquefaction, gasifcation, coking, or combustion processes. Published rates on more or less comparable coals may differ by several orders of magnitude, especially when *Author to whom correspondence should be addressed. Presented a t the Symposium on Coal Pyrolysis: Mechanisms and Modeling, 194th National Meeting of the American Chemical Society, New Orleans, LA, August 31-September 4, 1987. *Present address: Bechtel Corp., 3000 Post Oak Blvd., Houston, T X 77252. Present address: Eastman Kodak Co., Rochester, NY 14650.

0887-0624/88/2502-0400$01.50/0

obtained at high temperatures (>lo00 K) and/or high heating rates (102-106 K/s).'v2 A t the present state of the art in coal devolatilization research, more emphasis should perhaps be placed on elucidating the mechanisms of the chemical reactions underlying the observed phenomena. When thermal conversion reactions in coal are studied, it seems correct to concentrate first on the so-called "primary" reactions before attempting to elucidate the many possible secondary reaction pathways. This is especially true when viewed (1) Howard, J. B.; Peters, W. A.; Serio, M. A. EPRI Report No. AP1803; EPRI: Palo Alto, CA, April 1981. (2) Solomon, P. R.; Serio, M. A.; Carangelo, R. M.; Markham, J. R. Fuel 1986, 65, 182.

0 1988 American Chemical Society