Impact of Impregnated Catalysts on Coal Particle Disintegration at

Feb 2, 1994 - apparatus, equipped with sapphire windows and video recording equipment, permitted observation of 100-wm particles at temperatures as ...
0 downloads 0 Views 2MB Size
Energy & Fuels 1994,8, 1049-1054

1049

Impact of Impregnated Catalysts on Coal Particle Disintegration at Elevated Temperatures S. J. Abedi, L. Dukhedin-Lalla, and J. M. Shaw* Department of Chemical Engineering and Applied Chemistry, University of Toronto, Toronto, M5S 1A4, Canada Received February 2, 1994. Revised Manuscript Received May 12, 1994@

The impact of impregnated metal sulfate catalysts on the distintegration of coal particles in the presence of a solvent at elevated temperatures was explored with a novel view cell. The apparatus, equipped with sapphire windows and video recording equipment, permitted observation of 100;um particles at temperatures as high as 700 K and pressures as great as 200 bar. Coal particles were impregnated with iron and cobalt sulfates. The spatial distribution of these catalysts within coal particles and the impact of iron sulfate on coal particle disintegration at elevated temperatures were investigated. As the focus of this work was to identify disintegration mechanisms, disintegration rates per se were not measured. The impact of process variables was assessed through changes in particle morphology and fragmentation over a fixed reaction period. Particle disintegration was found to be unaffected by the presence of impregnated catalyst.

Introduction The rapid disintegration of coal particles in fluids at elevated temperatures is a crucial first step in coprocessing and coal liquefaction reaction environments and is a matter of concern for applications where coal particles are to be used as a disposable catalyst support. Despite its importance, little is known about this phenomenon and few direct measurements have been reported in the open literature. Pozzobon et a1.l conducted some preliminary experimepts with a capillary view cell, in the absence of catalytic effects, and proposed a physicochemical model for particle disintegration. In rough outline, their model proposes that small reagent molecules diffuse into the submicropores of coal particles where they react rapidly with coal constituents to produce molecular species of intermediate density. These latter species are hrther proposed to be too large to diffuse out of the pores and the resulting pressure rise within the pores causes the particles t o fracture. The model is supported by experimental evidence. For example, Pozzobon et a1.l observed micron- and submicron-sized fragments drifting away from parent particles when suitable reagents were present in the reaction environment. The proposed mechanism is also consistent with extant coal dissolution kinetic studies, where particle size is found to be a variable of secondary importance.2 Coal particles disintegrate physically at short reaction times with minimal reaction. This model was extended by Abedi et al.3a,bto include the impact of agitation and other process variables. The authors showed that presaturation of particles with an

* Author to whom correspondence should be addressed.

Abstract published in Advance ACS Abstracts, July 1, 1994. (1)Pozzobon, R. A,; Shaw, J. M.; Yushun, S. Can. J. Chem. Eng. lgSo,68,340-345. (2) Whitehurst, D. D.; Farcasiu, M.; Mitchell, T. 0. Coal Liquefaction; Academic Press: New York, 1080. (3) (a) Abedi, S. J.; Shaw, J. M. Interim Report for DSS Contract 6455.23440-0-9167,1992. (b)Abedi, S. J.; Rahimi, P. M.; Shaw, J. M. Prepr.-Am. Chem. SOC.,Diu. Pet. Chem. 1993,38, 346-350. @

Table 1. Coal Analysis Blind Canyon rec proximate analysis % moisture % ash % volatile % fixed carbon ultimate analysis % carbon % hydrogen % nitrogen % sulfur % oxygen

dry

daf

Battle River rec

dry

daf

5.20 26.01 13.17 13.89 10.53 14.23 39.23 41.38 48.05 27.06 36.57 42.64 42.40 44.73 51.95 36.40 49.20 57.36 66.15 69.78 81.04 47.13 63.70 74.27 4.80 5.06 5.88 3.25 4.39 5.12 1.21 1.28 1.49 0.95 1.28 1.50 0.45 0.47 0.55 0.40 0.54 0.63 9.02 9.52 11.06 11.73 15.86 18.48

aromatichydroaromatic solvent coupled with mild agitation greatly accelerated the rate of particle disintegration through the provision of a complementary breakage mechanism. Presaturation of a coal particle with a hydroaromatic solvent and hydrogen facilitates the penetration of these species through to the particle core via diffusion through pores or directly through the coal matrix.* If the particle is then heated, these species react with the coal matrix in much the same way as they do at the outer surface and the particle structure is weakened. As some macerals are more readily penetrated with solvents or are more reactive than others, coal particles quickly become an agglomeration of randomly sized subparticles held together by thin fluid films. Under mild agitation, these films are ruptured and the much finer subparticles which result then disintegrate according to the basic mechanism. These mechanisms are summarized in Figure 1. It is well-known that iron- and cobalt-containing compounds act as catalysts during hydrogenation reactions with coal in the presence of a solvent5 and contact between coal and catalyst can be realized in three ways: by combining coal particles with finely divided (4) Larsen, J. W.; Lee, D. Fuel 1983,62, 1351-1354. (5) Garg, D.; Givens, E. N. Znd. Eng. Chem. Process Des. Dev. 1982, 21, 113-117.

0887-0624/94/2508-1049$04.5~/0 0 1994 American Chemical Society

1050 Energy & Fueli, Vol. 8, No.5,1994

Abedi et al.

r with or without Agitation

Figure 1. Mechanisms for coal particle disintegration.

300

v

0

I

I

I

100

200

300

time. mln

Figure 2. Reactor heating profile.

catalyst (2-3 p m in diameter), by impregnating coal particles, or by dispersing catalyst at a molecular level using chemical vapor d e p o ~ i t i o n ~or - ~another technique. The impregnation and molecular dispersion methods are more attractive than the particulate addition method.6 However, it is unclear which is the most effective catalyst addition method as findings conflict6-* and the role played by catalysts in the particle disintegration process in particular is unclear. This matter is addressed in the present work, where the impact of impregnated metal sulfate catalysts on the mechanism or rate of particle disintegration is explored.

Experimental Section Dried and weighed coal particles were treated with an iron or cobalt sulfate solution. Three different metal sulfate concentrations, 2.5,5.0, and 10.0 wt %, were employed. Once treated, the coal samples were dried in a desiccator under vacuum for 48 h. The treated coal particles were then carefully

Figure 3. EDX scan for a 5-mm Battle River coal particle treated with excess FeS04-7HzO solution and decanted. Note the catalyst crust (a) 2.5 wt % solution; (b)wt % solution. fractured with a gentle hammer blow to reveal inner surfaces that were not directly exposed to the solution. Analysis of these surfaces using a Hitachi S-570 scanning electron microscope equipped with a LINK AN-10000 energy-dispersive

Energy & Fuels, Vol. 8, No.5,1994 1061

Coal Particle Disintegration

no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

cod Blind Canyon Blind Canyon Battle River Battle River Blind Canyon Battle River Battle River Battle River Blind Canyon Blind Canyon Battle River Blind Canyon Blind Canyon Battle River Battle River Battle River. Blind Canyon Blind Canyon Battle River Battle River

Table 2. Summary of Experimental Program solvent catalyst presatum 3.4517.79 no 3.45f7.88 no 3.45f7.03 no 3.45f7.58 no 3.45f7.58 no 3.45ff.58 no 3.4517.24 Yes 5.5019.65 no 5.50111.03 no 6.90111.72 no 6.90111.72 no 6.90111.03 Yes 5.5018.27 no 6.9017.58 no 6.9017.58 Yes 6.90110.2 Yes 6.9018.96 Yes 6.9018.23 Yes 5.5018.50 no 6.9018.09 no

press. initlmax (MPa)

Figure 4. (a)Fe scan for an interior surface of an Battle River coal particle ( 1 mm) treated with excess 10.0 wt % FeS04*7H20 solution ( 1 3 0 0 ~magnification). (b) A similar Fe scan for an untreated Battle River coal particle (1300x magnification). X-ray analyzer revealed the extent of catalyst impregnation into the coal particles, and maps of iron and cobalt distributions in coal particles were obtained for both treated and untreated particles. Chemical analyses of the Blind Canyon (6) Garg, D.;Givens, E. N.Fuel Pmcess. Technol. 1983,7,50-70. (7) Suzuki, T.;Yamado, 0.;Fujita, K. Fuel 1984.63, 1706-1709. (8)Mitra, J. R.;Chowdhury, P. B.;Mukhetjee, D. K. Fuel P m s s . Techml. 1904,8,282-291. (9)Tsodikov,M. V.; Maksimov, Y.V.; Ellert, 0. G.; Perederii, M. A;Grozhan, M. M.Khim. Tuer. Topl. 1989,23,96-101.

solvent tetralin tetralin tetralin tetralin tetralin tetralin tetralin tetralin tetralin tetralin tetralin tetralin tetralin tetralin tetralin tetralin hexadecane hexadecane tetralin tetralin

gas

agitn no no no no no no no no no no no no Yes Yes Yes no no no no Yes

Figure 5. Battle River coal particle (1 mm) treated with excess 10.0 wt % CoSO.p7H20 solution: (a) Co scan for an interior surface; (b) SEM micrograph showing the interior surface. (bituminous) coal and Battle River (subbituminous) coal are shown in Table 1. All high-temperature experiments were performed in a 50an3sandwich-style cell manufactured by D. B. Robinson and Associates. The cell is constructed from 316 stainless steel and the cell body is machined into a stainless steel block. The sapphire windows at both ends are sealed and cushioned by graphite shmis and clamped in place by stainless steel cover plates. The studs that secure the whole assembly include spring washers which compensate for differential thermal expansion effects. The cell is also equipped with 10 ports.

1052 Energy & Fuels, Vol. 8, No.5,1994

Abedi et al.

_1 b

Figure 6. h r e a c t i o n surface morphology: (a)Blind Canyon coal; (b) Battle River coal. During this study, tht.ee were used for pressure and temperature monitoring, and a rupture disk connection, and two for fluid inlet and outlet. The maximum working pressure at 700

K is 200 bar. During a typical experiment, four or five 1000pm diameter impregnated coal particles and 20 mL of reagent grade tetdin or hexadecane were loaded into the cell, which was then depressurized with an aspirator to remove air. For experiments with agitation, gas bubbles with a 2.4 m m diameter were continually fed through the base of the cell (3 f 2 bubbled 8 ) and gas was bled from the top to maintain a target pressure. About 10 mL of tetralin vaporized during the course of each of these experiments. For experiments without agitation, the reactor was sealed and allowed to reach a target pressure as the cell was heated to 700 K. A typical heating curve is shown in Figure 2. Particle disintegration was recorded on VHS video tape at 9 times magnification. Photomicrographs of coal particles were also taken before and after each experiment. For a more detailed description of the apparatus and procedure see Abedi.l0

Results and Discussion Impregnated Catalyst Distribution. Battle River coal particles were soaked in three different iron sulfate heptahydrate solutions, 2.5, 5.0, and 10.0 wt %, for 3 days. The coal particles were then decanted and dried. The samples increased in mass by approximately 1, 1, and 1.5 wt %, respectively. All coal particles, regardless of treatment period or solution concentration, exhibited the same iron distribution under EDX. The bulk of the iron was found in a 50-200 pm thick band on the outer (10) Abedi, S. J. MA&. Thesis, University of Toronto, 1993.

Figure 70 Postreaction surface morphology, Blind,Canyon coal, hydrogen and tetralin, P = 3.45 MPa: (a).without catalyst, run 1;(b) with catalyst, run 5.

surface of the particles (Figure 3). Subsequent visual microscopy revealed that this iron was largely present as sulfate crystals. By removing the coal particles from the solution before decanting, it was possible to prevent crystal formation on the surface and to examine the spatial distribution of iron or cobalt which penetrated the coal particles without interference from the perimeter. Iron sulfate readily penetrates 1 mm diameter coal particles and appears to be uniformly distributed even on grids as small as 10 x 10 pm (Figure 4). Analysis at greater magnification is impeded by the topography of the fracture surfaces. These surfaces are not smooth. Thus, not all parts of a surface reflect X-rays evenly. Experiments performed with cobalt sulfate yielded comparable results. Cobalt sulfate is also readily distributed within coal particles and the distribution appears homogeneous on a 100 x 100pm grid as shown in Figure 5. The question of whether the observed fiacture surfaces are representative or are artifacts of the experiments is open. Every precaution was taken to ensure that surfaces selected for analysis were perpendicular to the exterior surface of the coal particles. However, it was not possible to ascertain whether or not the fiacture surfaces followed preexisting cracks and were therefore more accessible to the penetrating catalyst than a random interior surface. As a control experiment would be difficult to devise, the reported catalyst distribution results, though encouraging, must be considered tentative. Even so, catalyst is clearly distrib-

Coal Particle Disintegration

Energy & Fuels, Vol. 8,No.5,1994 1063

b

- .

Figure8. Postreactionsurface morphology, Battle River coal, hydrogen and tetralin, P = 3.45 MPa: (a) without catalyst, run 3; (b)with catalyst, run 6.

Figure 9. Postreaction surface morphology, Battle River coal, hydrogen and tetralin, P = 6.9 MPa, agitation; (a) with catalyst, run 14; (b)without catalyst, run 20.

uted within the coal particles and can in principle affect particle disintegration. Impact of Impregnated Catalyst on Particle Disintegration. A matrix of high-temperature experiments was established which involved the following variables and levels: coal type (Blind Canyon, Battle River), gas type (hydrogen, nitrogen), solvent type (tetralin, hexadecane), catalyst (iron sulfate, none), initial gas pressure (3.45, 5.5, 6.90 m a ) , agitation (bubble injection, none), and presaturation of particles with solvent (yes, no); i.e., for some experiments the coal particles were treated overnight in solvent under pressure but a t room temperature. All other operating conditions were kept constant. A total of 20 experiments were performed as shown in Table 2. In general, it is worth noting that Battle River coal was more reactive than Blind Canyon coal under the reaction conditions employed based on the extent of particle fragmentation and perforation. Thus, coal type is of importance in determining the rate of particle disintegration. However, this variable is of less importance than other variables examined. The surface of unreacted Battle River and Blind Canyon coal particles are smooth and featureless as shown in Figure 6. Postreaction particles (Figures7-11 are smaller, exhibit rough surface morphologies, and are perforated (Figures 7b, 8a, l l a ) in some cases. These changes in particle morphology arise from the various disintegration mechanisms depicted in Figure 1 and appear to be unaffected by added catalysts. A comparison of the results depicted in Figures 6-11

clearly shows that added catalyst does not affect the disintegration process significantly. There is little difference between postreaction particle morphologies or particle sizes for both coal types for catalyzed and uncatalyzed trials run under otherwise similar condicoal type tions, for example, impregnated catalyst (Figures 7 and 8),and impregnated catalyst high- or low-pressure hydrogen (Figures 8 and 11). In addition, there is no combination of catalyst plus another variable which provides a synergistic change in particle disintegration mechanism or rate relative to uncatalyzed trials: e.g., results from experiments with catalyst agitation (Figure 9), catalyst solvent presaturation (Figure lo), catalyst hydrogen pressure (Figure 11). It has been reportedl1-l3 that the main contribution of catalysts during coal hydrogenation processes is the hydrogenation of solvent-soluble species. Our results are consistent with this finding, as the dispersed iron sulfate catalyst does not appear to play a role in coal particle disintegration. The experiments reported in this paper have a long heating period. Typical heat up times in large scale coal liquefaction or coprocessing units are closer to 6 min than 2 h. Thus, the impact of mass-transfer effeds on particle disintegration rate would be more pronounced and the particles would be more durable at elevated temperatures in large-scale processes than indicated by

+

+

+

+

+

(11)Rubem, R G.; Cmunauer, D. C.; Jewell, D. M.; Seshadri, K. S. h l 1977,56,17-24. (12) Rottendorf, H.;Wilson, M. A. Fuel 1980,59,175-180. (13) Derbyshire, F. J.; Whitehurst, D.0. F u l 1981,60,655-662.

1084 Energy & Fuels, Vol. 8, No.5, 1994

Abedi et al.

-

Figure 10. Postreaction surface morphology, Blind Canyon coal, hydrogen and hexadecane, P = 6.9 MPa, saturated: (a) with catalyst, run 17; (b)without catalyst, run 18.

our experiments. Even so, particle disintegration rates accelerate a t temperatures greater than 700 K, and particles merely act as a catalyst dispersant rather than as a support. This is particularly true for catalyst near external surfaces of particles, and other dispersant methods may prove equally effedive and less expensive.

Conclusions The nature of the distribution of added catalysts on and within coal particles is sensitive to the method of impregnation. Catalysts may reside predominantly within or on the surface of coal particles. Homogeneous

.” v.ys

* 7--

3*scr*-l’r

Figum 11. Yostreaction surface morphology, Battle River coal, hydrogen and tetralin: (a) with catalyst, P = 5.5 m a , run 8; (b)without catalyst, P = 5.9 MPa, run 19.

distribution of iron and cobalt catalysts within coal particles is readily achieved. However, the rate of particle disintegration is unaffected by the presence of impregnated iron catalysts. Coal particles are not durable a t elevated temperatures and act more readily as dispersants than as supports for catalysts.

Acknowledgment. We thank the Department of Metallurgy and Material Science a t the University of Toronto for the use of equipment and gratefully acknowledge financial assistance provided by CANMET through DSS contract 6438.23440-0-9167,and NSERC through the operating grant program.