Calcium-Ion Exchange of Coal for Retention of Sulfur During Combustion Citation Chang, Karl Kho-Chung (1985) Calcium-Ion Exchange of Coal for Retention of Sulfur During Combustion. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/X4W7-D106. https://resolver.caltech.edu/CaltechETD:etd-08092006-145027 Abstract The kinetics of calcium-ion exchange of coal and the combustion of calcium-exchanged coals in a laminar flow furnace were investigated. Particle diffusion was found to be the rate-determining step for the process of ion exchange. A diffusion model gave good agreement with observed rates of ion exchange. Diffusion of ions in PSOC 680, a bituminous coal, was observed to be much slower than in PSOC 623, a lignite. Retention of sulfur during combustion of calcium-exchanged coals was observed to increase with increasing particle residence and oxygen concentration in the combustion gas. Chemical reaction was found to be the rate-determining step in the capture of SO 2 by CaO in coal ash. The results of this study were interpreted in a proposed mechanism for the release and capture of SO 2 . Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Chemical Engineering) Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemical Engineering Thesis Availability: Public (worldwide access) Research Advisor(s): Gavalas, George R. (advisor) Corcoran, William Harrison (co-advisor) Thesis Committee: Gavalas, George R. (chair) Corcoran, William Harrison Flagan, Richard C. Cass, Glen Rowan Shair, Fredrick H. Defense Date: 7 January 1985 Funders: Funding Agency Grant Number Caltech UNSPECIFIED Department of Energy (DOE) UNSPECIFIED Record Number: CaltechETD:etd-08092006-145027 Persistent URL: https://resolver.caltech.edu/CaltechETD:etd-08092006-145027 DOI: 10.7907/X4W7-D106 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 3074 Collection: CaltechTHESIS Deposited By: Imported from ETD-db Deposited On: 23 Aug 2006 Last Modified: 19 Nov 2025 19:06 Thesis Files Preview PDF (Chang_kkc_1985.pdf) - Final Version See Usage Policy. 3MB Repository Staff Only: item control page

CALCIUM-ION EXCHANGE OF COAL FOR RETENTION OF SULFUR DURING COMBUSTION Thesis By Karl Kho-Chung Chang In Partial Fulfilment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 1985 (Submitted January 7, 1985) ii This thesis is dedicated to my parents, who always knew that I was headed for graduate school. © 1985 Karl Kho-Chung Chang All Rights Reserved ACKNOWLEDGMENTS It is with my deepest appreciation that I acknowledge my thesis advisor, Pro- fessor George R.*Gavalas, for his thoughtful suggestions and guidance throughout the second half of this project. He provided direction to my research, while allowing me the latitude to be creative. I would like to express my gratitude to my former advisor, the late Professor William H. Corcoran, who patiently guided me through difficult times. His ideas on professional and per- sonal conduct will always be an inspiration to me. I would also like to ack- nowledge Professor Richard C. Flagan for his invaluable advice and contribu- tions on the second half of this project. The financial support of the California Institute of Technology and the Department of Energy is gratefully acknowledged. I wish to express my warm thanks to Donna Johnson and Kathy Lewis for helping me prepare this thesis at moment's notice. I thoroughly enjoyed the open and friendly atmosphere created here at Cal- tech by my fellow graduate students. In particular, I thank Murray, Puvin, Teri and Phil. I would like to thank Hank Blauvelt and Alex Ortiz, whose years of companion- ship made life at Caltech more bearable. Special thanks go to Kwang-I Yu, who was always around when I needed advice and counseling. | also gratefully thank Ellen Johnson, who helped with my experiments and helped to keep me sane during the rigors of my last years at Caltech. Most of all I want to thank my parents for their love and support for making all this possible. iv Abstract The kinetics of calcium-ion exchange of coal and the combustion of calcium- exchanged coals in d laminar flow furnace were investigated. Particle diffusion was found to be the rate-determining step for the process of ion exchange. A diffusion model gave good agreement with observed rates of ion exchange. Diffusion of ions in PSOC 680, a bituminous coal, was observed to be much slower than in PSOC 623, a lignite. Retention of sulfur during combustion of calcium-exchanged coals was observed to increase with increasing particle residence and oxygen concentra- tion in the combustion gas. Chemical reaction was found to be the rate- determining step in the capture of SO2 by CaO in coal ash. The results of this study were interpreted in a proposed mechanism for the release and capture of S02. Table of Contents Page Acknowledgments ----++scceesc cee e cece cere ec eeereretace see eee eee iii Abstract ......... ee eee ee eee tee eee ee een ee eee eee ee ee eens iv Table of Contents 22... ccc cece ccc cee cece eee neces cee eees v List of Tables 2... ccc cece ce cece eee e ee eeteeeteneeee viii List of Figures ..- eee e eee cece reece cence eee eees rrr x Chapter 1. INTRODUCTION ........ eee eee cee cece eet ee eee tee neces 1 References 02. ccc cece cc cee cee eee eee tee ete eee ee eee ee eeeeeeees 4 Chapter 2. CALCIUM-ION EXCHANGE OF COAL 2.2... .. ccc ccc eee ee cece e eee 5 2.1 Introduction ..... 0... cece ccc nee ce eee eet e ee ee eee e eres eeeees 5 2.2 Experimental ....... cc cece ce cece cece ete e rete eee ee cence 6 2.2.1 Coal Preparation sew ee mee we ee te ee ee ee eee eee e ee eee ee eee ees 6 g.2.c Experimental Apparatus and Procedure .........c cece ceees 8 2.2.2.1 Coal Oxidation .... ccc ccc cee eee tent teeee 8 2.2.2.2 lon Exchange of Coal --- +--+ sees eee e rece eect eee eaees 8 z.2.3 Analysis of lon-Exchange Capacity of Coal ..............00005 13 2.2.3.1 Carboxyl Group Determination .--+.--ssseeeeee ce eeee 13 2.2.3.2 Total Acid Group Determination ............0200e eee 13 23 ResultS ..o. ccc ccc eee e eee e ee ee ee te eee eeneuees 14 24 DiscusSion ....... ccc eee eee ec eee eee eee eee teen eee eenee 25 2.4.1 Reaction Rate of Ion Exchange ----++++eseee sce ssereceeeeces 26 24.2 Film Diffusion ...... 20... eee ccc ec eect we ce tee etneesaee 31 vi ‘Page 2.4.3 Particle Diffusion...... eee e ecw eee e ence rece eseceeeees 32 2.4.4 Effect of Partial Oxidation ........ cece cece cere cece eeeeens 38 Z.5 Conclusions .i.... cece cece cc cece cee e eee e eee eee e ee neeee 43 References 2.0... . cece ccc eect cee te tee eee eee e ere eeeeeees 44 Chapter 3. COMBUSTION OF CALCIUM-EXCHANGED COAIS ........5.2--06- 45 83.1 Introduction .... 2. eee cece ce cece ce eee teen eee e eee t eee eeeeees 45 3.2 Experimental ........ cece ccc eee ce eee te teen eee ee eee eenceces 47 3.2.1 Coal Preparation ........ cece cece eee eee cece eee n eee eeees 47 3.2.2 Experimental Apparatus and Procedure ...-++++-+-++++eeeers 48 3.2.3 Chemical Analysis of Coal ......... 0... ce eee cece n cence neces 53 3.2.3.1 Sulfur Determination ...--.eeee cece eee ee eee eeeeees 53 3.2.3.2 Calcium Determination ....... 0. cece ee eee ee ee eee eee 54 33 Results ..- cc ccc e ccc ce ee eee cee ee eee ee eee eee e teen eee eeeens 56 B.4 Discussion 1.0... sce e eee eee e ence tence eee e ence nec eeeeees 58 3.4.1 Effect of Residence Time .....-. cece cee cece ee ee ree cece nees 66 3.4.2 Effect of Oxygen Concentration ......... eee cere eee ence nee 86 3.4.3 Effect of Calcium Content ........ ccc cee ce eee een eee eens 93 3.4.4 Effect of Furnace Temperature ...---..eeee cece eee nce enes 96 3.5 Conclusions .......... eee eee eee Lec eeeteueceueseeueuaveeuees 101 References 2... eee ccc cece eee eee eee ence nee c eee e cence enees 103 Chapter 4. CONCLUSIONS AND RECOMMENDATIONS 3 ...---+eeeeeececeees 105 4.2 Recommendations ...... APPENDIX. Furnace Calibration vii eee erm eee eerwnecoeeeneeseeoe eee seanoneseeeeves Cr Table Table 2.1, Table 2.2. Table 2.3. Table 2.4. Table 3.1. Table 3.2, Table 3.3. Table 3.4. Table Ai. Table A.2. viii List of Tables Elemental Analyses of PSOC 623 (Darco Lignite and PSOC 680 (Indiana No. 6) on a Dry Basis Provided by the PSU Coal Bank. Effect of Partial Oxidation on the Content of Acidic Groups in PSOC 680. Effect of Oxidation on Ca/S Content Obtainable by Calcium-Ion Exchange of PSOC 680. Calcium-Ion Exchange of Carboxylic-Acid Groups in Par- tially Oxidized PSOC 680 (120x120 Mesh). Sulfur and Calcium Analysis of Coal Samples Used in the Combustion Experiments. Sulfur Analysis of PSOC 6B0 and PSOC 623 on a Dry Basis Provided by PSU Coal Bank. Sulfur Balance for the Combustion of PSOC 680 (Ca/S = 3.4) at 1600°K in 40% Ox. First Order Reaction Rate Constants for the Retention of S02 During Combustion of Coals at 1600°K. Corrected Data for Retention of Sulfur During Combus- tion of PSOC 623A (Ca/S = 2.5) at a Furnace Tempera- ture of 1600°K. Corrected Data for Retention of Sulfur During Combus- tion of PSOC 623B (Ca/S = 0.8) at a Furnace Tempera- 20 39 42 57 57 60 79 110 110 Table A.3. Table A4. ix ture of 1600°K. Corrected Data for Retention of Sulfur During Combus- tion of PSOC 680A (Ca/S = 3.4) at a Furnace Tempera- ture of 1600°K. Corrected Data for Retention of Sulfur During Combus- tion of PSOC 6B0A (Ca/S = 3.4) at a Furnace Tempera- ture of 1400°K. 111 111 Figure Figure 2.1. Figure 2.2. Figure 2.3. Figure 2.4. Figure 2.5. Figure 2.6. Figure 2.7. Figure 2.8. Figure 2.9. Figure 2.10. Figure 2.11. list of Figures Fluidized-bed reactor. Cross sectional view of reactor (not to scale). Ion-exchange apparatus. Effect of particle size on the extent of calcium exchange of PSOC 623 at pH 8.3 and 2e°C in 0.1 N [Ca**], Effect of calcium concentration on the extent of cal- cium exchange of PSOC 623 (120x200 mesh) at pH 8.3, Effect of temperature on the extent of calcium exchange of PSOC 623 (120x200 mesh) at pH 8.3 and 0.1 N [Cat*]. Effect of pH on the extent of calcium exchange of PSOC 623 (120x200 mesh) at pH 22°C and 0.1 N [Ca**]. Effect of oxidation at 195°C on the extent of calcium exchange of PSOC 623 at pH 8.3 and 0.1 N [Ca**]. Effect of oxidation at 195°C on the rate of calcium exchange of PSOC 680 at pH 8.3 and 0.1 N [Ca**]. Effect of oxidation at 260°C on the extent of calcium exchange of PSOC 680 at pH 8.3 and 0.1 N [Ca**]. Effect of oxidation at 260°C on the extent of calcium E wo ® 10 11 16 17 18 19 2] 22 23 24 Figure 2.12. Figure 2.13. Figure 2.14. Figure 2.15. Figure 2.16. Figure 2.17. Figure 2.18. Figure 3.1, Figure 3.2. Figure 3.3. Figure 3.4. Figure 3.5. xi exchange of PSOC 680 at pH 8.3 and 0.1 N [Ca**]. Test of second order dependence of calcium exchange of PSOC 623 with respect to the concentra- tion of carboxylic-acid groups. Effect of temperature on the rate of calcium exchange of PSOC 623. Effect of particle size on the rate of calcium exchange of PSOC 623. Effect of calcium concentration on the rate of cal- cium exchange of PSOC 623. Test of particle-diffusional limitation on calcium exchange of PSOC 623. Effect of partial oxidation on the fractional conver- sion of carboxylic-acid groups in PSOC 680. Test of particle-diffusional limitation on calcium exchange of partially oxidized PSOC 6B0. Combustion furnace (not to scale). Placement of thermocouples (to scale). Coal feeder (not to scale). Flow system for combustion experiments. Effect of gas flowrate and oxygen concentration on sulfur retention during combustion of calcium-ion- exchanged PSOC 680 (Ca/S = 3.4) in a laminar flow furnace. 29 30 34 37 40 4] 49 50 51 52 59 Figure 3.8. Figure 3.7. Figure 3.8. Figure 3.9. Figure 3.10. Figure 3.11. Figure 3.12. Figure 3.13. Figure 3.14. xTi Effect of gas flowrate and oxygen concentration on sulfur retention during combustion of calcium-ion- exchanged PSOC 623 (Ca/S = 0.8) in a laminar flow furnace. Effect of gas flowrate and oxygen concentration on sulfur retention during combustion of calcium-ion- exchanged PSOC 623 (Ca/S = 2.5) in a laminar flow furnace, Effect of gas flowrate on sulfur retention during combustion of PSOC 680 (Ca/S = 3.4) at a furnace temperature of 1400°K. Calculated equilibrium 50, partial pressure for the +0, > CaS0,. reaction CaO + S02 + 5 Furnace temperature versus distance from coal injector. Theoretical calculation of particle temperature for the combustion of a bituminous coal at a furnace temperature of 1600°K. Illustration of particle temperature profile. Effect of partial residence time on sulfur retention for the combustion of PSOC 680 (Ca/S = 3.4) at vari- ous oxygen concentrations. Effect of partial residence time on sulfur retention for the combustion of PSOC 623 (Ca/S = 0.B) at vari- ous oxygen concentrations. 62 63 65 67 68 70 71 72 Figure 3.15. Figure 3.16. Figure 3.17. Figure 3.18. Figure 3.19. Figure 3.20. Figure 3.21. Figure 3.22. Figure 3.23. xTT7 Effect of partial residence time on sulfur retention for the combustion of PSOC 623 (Ca/S = 2.5) at vari- ous oxygen concentrations. First order dependence of sulfur capture with respect to SO, concentration during combustion of PSOC 623 (Ca/S = 3.4) at 1600°K. First order dependence of sulfur capture with respect to SOz concentration during combustion of PSOC 623A (Ca/S = 2.5) at 1600°K. First order dependence of sulfur capture with respect to SQ, concentration during combustion of PSOC 623B (Ca/S = 0.8) at 1600°K. Relative concentration (C/ C§o,) of 502 as a function of time for the diffusion of 50, through bulk gas as described in Eq. (3-39). The effect of oxygen on first order reaction rate con- stants. Calculated calcium vaporization during combustion of a coal with 2.5% Ca at a furnace temperature of 1600°K. The effect of oxygen on vaporization of calcium dur- ing combustion of pulverized coals at a furnace tem- perature of 1750°K (data of Quann). Effect of calcium content on sulfur retention for combustion of coal in 40% Q5o. 75 76 77 87 88 91 92 94 Figure 3.24. Figure 3.25. Figure 3.26. Figure 3.27. Xiv Effect of calcium content on sulfur retention for combustion of coal in 31% Og. Effect of furnace temperature on sulfur retention during combustion of PSOC 680 (Ca/S = 3.4). Effect of residence time on sulfur retention during combustion of PSOC 680 (Ca/S = 3.4) at a furnace temperature of 1400°K. Theoretical calculation of particle temperature for the combustion of a bituminous coal with a diameter of 50 wm. 97 99 100 CHAPTER 1 INTRODUCTION In the wake of world energy crises in the 1970's, coal has become an increas- ingly important source of energy. However, coal utilization gives rise to a number of ecological problems, among which are air pollution from particulates, nitrogen oxides and sulfur dioxide produced by combustion in utility boilers. Efforts to control sulfur emission during coal combustion have led to various commercial processes of post-combustion scrubbing. Because of the high cost and complex operation of these scrubbing processes, alternative methods of sul- fur removal have been under active investigation. The most successful of these, fluidized combustion in a bed of calcined limestone or dolomite, is already com- mercial] in industrial boilers and might soon become commercialized in utility boilers. At the relatively low temperatures of fluidized combustion, sulfur oxides combine with the calcium or magnesium oxide particles to form stable sulfates. Fluidized combustion has several other advantages as well, but its application is limited to new power plants. Other methods of sulfur control suitable for exist- ing pulverized coal-fired boilers rely on injecting the calcium sorbent with the coal and removing the spent sorbent with the ash. At the high temperatures of the pulverized coal flame, calcium sulfate is not thermodynamically stable. Therefore, sulfur oxide removal must exploit either rate limitations in the release of sulfur oxides or the capture of sulfur oxides in the cooler post-flame part of the furnace, where the equilibrium formation of sulfates becomes ther- modynamically favorable. Two modes of calcium sorbent injection have been investigated. The first involves the injection of ground limestone along with the coal either in a conven- tional furnace or in a furnace employing staged combustion (Case et al., 1982). This method has the merit of extreme simplicity and low cost but has yielded sulfur removal in the range of 40-60%, insufficient to meet existing environmen- tal standards. Evidently, the kinetics of sulfur oxide capture in the post-flame region of the furnace are not sufficiently rapid as observed by Borgwardt (1970), perhaps due to the slow diffusion within the partially sintered calcium oxide par- ticles (Pigford and Sliger, 1973; Hartman and Coughlin, 1976). Another method that has been investigated involves the calcium sorbent in closer association with the coal particles. One version is to impregnate the coal with a calcium solution and precipitate calcium carbonate right in the pores of particles. The other method is to add calcium by exchanging acidic groups in coal with calcium cations, using a suitable solution. This exchange may be represented by 2RH + CA**t + R°Cat*R™ + 2H* where RH is a carboxylic or phenolic group in coal. Lignites have sufficient acidic groups for accepting as much as 6-8% calcium by weight. Bituminious coals do not possess sufficient acidic groups but they can acquire the required ion exchange capacity by mild oxidation (~ 200°C) at the cost of losing some heating value. Adding calcium by ion exchange produces very fine, atomic-scale, dispersion in close association with the sulfur source resulting in more rapid kinetics of sulfur capture and reduced diffusional limitations due to deactiva- tion of CaO by solid product layers. Freund and Lyon (1982) studied the combustion of calcium-exchanged coal as a function of fuel equivalence ratio and found sulfur removal as high as 90% at sufficiently high equivalence ratios. At equivalence ratios used in conven- tional combustion the sulfur removal fell to 60% or lower. In view of this, they proposed this mode of calcium addition in connection with staged combustion. In the first fuel-rich stage sulfur would be captured in the form of sulfide. If this fuel-rich stage is followed by an oxygen-rich stage, the sulfide would be con- verted to sulfate provided the temperature remains sufficiently low. The present work examines both the kinetics of calcium-ion exchange of coal and the combustion of the calcium-exchanged coals in a laminar flow furnace under extreme fuel-lean conditions. The purpose of this study was to determine the mechanism of ion exchange of coal and the processes of sulfur dioxide release and recapture in the presence of finely dispersed calcium oxide during combustion. References Borgwardt, R. H., Env. Sci. Tech. 4:59-63 (1970). Case, P. L, Heap, M.-P., McKinnon, C. N., Pershing, D. W. and Payne, R., Prepr. ACS Fuel Chemistry Division 27:158-166 (1982). Freund, H. and Lyon, R. K., Comb. Flame 45:191-203 (1982). Hartman, M. and Coughlin, R. W., AICHE J. 22:490-498 (1976). Pigford, R. L. and Sliger, G., Ind. Eng. Chem. Proc. Des. Dev. 12:B5-91 (1973). CHAPTER 2 CALCTUM-ION EXCHANGE OF COAL 2.1 Introduction Ion exchange of coal to incorporate calcium into the coal matrix for reten- tion of sulfur during combustion has been shown to be promising in recent investigations by Freund and Lyon (1982), and Chang, Flagan and Gavalas (1984). The concept of ion exchange utilizes the presence of acidic functional groups in coal to exchange the hydrogen ions with calcium ions in an alkali medium. One of the factors that determines the success of the method of ion exchange for sulfur emission control is the ability to incorporate a sufficient amount of calcium into the coal sample so that a desired level of Ca/5 molar ratio (Ca/S = 1 for theoretically complete retention of sulfur by exchanged cal- cium) may be obtained. This criterion, in turn, is dependent on the ion- exchange capacity of the coal and the process of ion exchange. The characterization of the ion-exchange capacity of coal has been studied in detail (Fuchs and Sandhoff, 1940; Brooks and Sternhell, 1957; Schafer, 1970; Schobert, 1984). It was found that the occurrence of acidic groups increased as the rank (or carbon content) of the coal sample decreased, thus low-rank coals such as lignites generally possess a sufficient content of acidic groups to achieve Ca/S ratios greater than 1 by calcium-ion exchange. Higher-rank coals such as bituminous coals, on the other hand, do not possess sufficient acidic groups, but they can acquire the required ion-exchange capacity by mild oxidation at the cost of losing some heating value (Brooks and Maher, 1957; Kalema, 1985). A survey on the topic of kinetics of ion exchange revealed that a significant number of studies (Boyd, Adamson and Meyers, 1947; Reichenberg, 1953; Kress- man and Kitchener, 1949; Helfferich, 1962) have been conducted on theoretical and experimental aspects of ion exchange. However, the materials examined in all cases were monofunctional ion-exchange resins whose physical and chemical properties were relatively well characterized. Currently there are no available data on the kinetics of ion exchange of coal, possibly due to the complex nature of the system which involved not only a complicated pore-size distribution but also a range of acidic groups of varying strengths. In the present work the kinetics of calcium-ion exchange of two different types of coal were studied in a batch reactor. The purpose of this study was to elucidate the mechanism and kinetics of ion exchange of coal in order to utilize the ion-exchange properties of coal to incorporate calcium into the coal struc- ture for possible retention of sulfur during combustion. The effects of chemical Kinetics, film diffusion, partial diffusion, pH and partial oxidation were evaluated. 2.2 Experimental 2.2.1 Coal Preparation Two coals of different rank from the Pennsylvania State University Coal Bank were selected for this study. Elermental analyses for the coals, PSOC 623, a Texas Darco lignite, and PSOC 680, an Indiana No. 6 high volatile B bituminous, are presented in Table 2.1. Since the ion-exchange properties of coal are depen- dent on oxygen content, an oxygen content of 14.29% for PSOC 623 and 9.53% for PSOC 680 suggests considerable differences in the acid groups present in the coals. The coal samples were ground to 120x200 U.S. mesh size and treated with dilute acid to remove minerals from the solid. The ground coal was stirred with 1.0 N HCl in a ratio of 10 ml/g for 24 hours at 40°C. The acid treatment was sufficient for the removal of the cations of Ca, Mg, Na, and K associated with the Table 2.1 Elemental Analyses of PSOC 623 (Darco Lignite) and PSOC 680 (Indiana No. 6) on a Dry Basis Provided by the PSU Coal Bank PSOC 623 PSOC 680 % Ash 16.60 11.91 % Carbon 60.67 68.47 % Hydrogen 4.27 4.59 % Nitrogen 1.07 1.52 % Sulfur 1.13 3.58 oe Oxygen (diff.) 14.29 9.53 organic acid sites. After the acid wash, the slurry was filtered and rinsed with demineralized water to neutral pH. The powder was then vacuum dried at 105°C for 24 hours. 2.2.2 Experimental Apparatus and Procedure 2.2.2.1 Coal Oxidation Oxidation of the bituminous coal samples in air was carried out in a fluidized- bed reactor, shown as a block diagram in Fig. 2.1. The reactor was a stainless steel pipe with a length of 5B cm and a diameter of 2.5 cm. The reactor and air preheater were heated externally by semi-cylindrical ceramic elements. The temperature was monitored with an Inconel-sheathed, type K thermocouple probe inserted into the middle of the fluidized-coal bed, as illustrated in Fig. 2.2. Coal samples were introduced into the empty reactor from the feed hopper through the draft tube. The steel wool and glass wool at the top of the reactor serve to contain entrained coal in the exit stream. At the start of a reaction, the gas line to the draft tube was opened and the flow of fluidizing air was directed through the hopper. The coal was carried down the draft table into the bottom of the reactor. When the hopper was empty, the fluidizing air was switched back to the bottom of the reactor. Each oxidation experiment was conducted with approximately 20 g of coal. The tem- perature of the reactor was maintained within 5°C of the desired value. 2.2.2.2 lon Exchange of Coal The experimental apparatus used in this study is shown in Fig. 2.3. The reac- tor vessel was a 400 ml beaker placed in a water bath maintained at a constant temperature. The temperature of the reaction mixture was monitored with an immersed thermometer. A stream of nitrogen was used to purge the reactor of air in order to minimize the oxidation of coal in the presence of alkali and to Air Feed) Water Hopper Condenser Dessicant O fr Reactor Liquid Samples J Cold Trap | Air — > aan Preheater Gas Analysis Fig. 2.1 Fluidized-bed reactor. 10 Ay we & ~>2 tame OVARY va \\ Tee SED e “3 + an} nts ‘ deep ab Yo Se ends . Bate, ee ated eae . Be CO a Was eee Ce ee, ee Td . rs . - 0 74°? we aa mye . « "e t, i} - ry we fee o¥ ate oe of we gn : atoet teu pe, Oe vee cP siouamn eae en) : - -_— ea STEEL WOOL — THERMOCOUPLE GLASS WOOL DRAFT TUBE SAMPLE LINER SCREE Fig. 2.2 Cross-sectional view of reactor (not to scale). ll Ca(OH), I Calomel een ew wabee giles 189% 44% ie Fa Glass a Electrode ok H| Electrode y é ' pH Meter Water Bath — L. N\ i jij. i i . O C) Magnetic Stirrer ty j Fig. 2.3 Ion-exchange apparatus. 12 prevent the absorption of atmospheric COg by the solution. The coal slurry was well mixed with a magnetic stirrer while the water bath was agitated with a vari- able speed stirrer. Approximately 0.5 g of coal sample was added to the reactor along with 10 ml of demineralized water. The coal-water slurry was mixed for ten minutes to ensure that the coal particles were wetted thoroughly, thus reducing film- diffusional resistance on the particle surface. At the start of the reaction, 190 ml of pH-adjusted calcium acetate solution was added to the reactor. The reac- tor was sealed quickly and the reaction slurry was maintained at the initial pH value with the continuous addition of standard 0.03 N Ca(OH)z solution. The amount of Ca(OH)s titrant used was recorded at discrete periods of time. The extent of ion exchange at each time interval was then calculated from the volume of titrant as follows: _ Vea(on)gNca(oH)g ae 1000 w (2-1) where Q = the extent of calcium exchange at time t, g. equiv./g sample. Vca(oH), = volume of Ca(OH)g titrant at time t, ml. Nca(oH), = normality of Ca(OH), solution, g.equiv. /1. w = weight of coal sample, g. Special care was taken to ensure that the pH measurements were accurate and reproducible. The pH was measured with a calomel-glass-electrode pair and a Radiometer pHM-26 pH meter. The calomel electrode was kept filled with saturated-KC] solution in the presence of KCl crystals. The electrode was stored in saturated-KCl solution when not in use. Prior to an experiment, the elec- 13 trodes were placed in a pH 7 buffer solution for half an hour to equilibrate. At the end of the half hour the meter was calibrated to pH 7. The electrode pair was then immersed in a pH 10 buffer solution to adjust the slope of the meter response. In a well-stirred solution, the response time of the pH meter was approximately five seconds. 2.2.3 Analysis of lon-Exchange Capacity of Coal 2.2.3.1 Carboxyl Group Determination The carboxylic-acid-group content of coal was determined by using a pro- cedure developed by Schafer (1970). A 0.5-1.0 g sample of acid-washed 120x200 mesh coal was refluxed for four hours with 75 ml of 1 N barium acetate solution which had been adjusted to pH 8.3 with either dilute Ba(OH)z or HCl. The refluxed mixture was titrated back to pH 8.3 with 0.1 N Ba(OH)z. The procedure of refluxing and titration was repeated once more to ensure that complete exchange had taken place. During refluxing in a basic solution, carboxylic-acid groups in the coal preferentially picked up barium ions and released hydrogen ions to the solution. Thus, the carboxylic-acid-group content of the coal was calculated from the volume of titrant as follows: Vpa(OH)2NBa(OH)2 n> Ng = carboxylic-acid-group content, g. equiv./g sample. Vpa(oH), = Volume of Ba(OH), titrant, ml. Nga(oH), = normality of Ba(OH)g solution, g.equiv. /1. w = weight of sample, g. 2.2.3. Total Acid Group Determination The total acid-group content of coal was determined by adapting a procedure 14 developed by Blom et al. (1957). A 0.5 g sample of acid-washed 120x200 mesh coal was added to a 150 ml plastic bottle containing 50 ml of 0.1 N barium hydroxide solution. -The bottle was purged with nitrogen and then sealed to prevent the oxidation of coal in the presence of alkali and the reaction of Ba(OH)2 with atmospheric COg. The mixture was heated for 16 hours and the coal was then filterd and washed. During the alkali treatment at pH 12.5, all acid groups in coal preferentially picked up barium ions and released hydrogen ions. The excess Ba(OH),2 in the filtrate and washings was titrated by 0.1 N HCl. The total acid-group content of the coal was calculated from the volume of Ba(OH)2 and HCl as follows: _ Vea(or)Npacor), ~ VaewNuci ar = 1000 w (2-3) nr = total acid-group content, g. equiv./g sample. Vpa(oH), = Volume of Ba(OH)g, ml. Nga(OH), = Normality of Ba(OH), solution. g.equiv./l. Vuci = volume of HC] titrant, ml. Nuc} = normality of HCl titrant, gequiv./l. Ww = weight of sample, g. The phenolic-group content of the coal was taken as the difference between the total acid-group content and the carboxylic-acid-group content. 2.3 Results Analysis of the acid-washed PSOC 623 sample shows that the lignite coal selected for this study has a total acid-group content of 5.16 meq/g. The carboxylic-acid-group content is determined to be 2.46 meq/g, thus the 15 difference of the two values establishes the phenolic-hydroxyl-group content at 2.70 meq/g. The effect of particle size on the kinetics of calcium-ion exchange of the lig- nite coal at pH 8.3 is exhibited in Fig. 2.4. The results reveal a significant increase in the extent of exchange achieved as a function of time as the particle size decreased from 50xB0 mesh (177-297 ym) to 120x200 mesh (74-125 wm). Figure 2.5 shows a slight dependence of the extents of exchange on the concen- tration of calcium ions in the exchange solution. Analysis of the data indicates that as the concentration of calcium ions increases from 0.05 N to 1.0 N, the ini- tial rate of exchange increases, but the final rates of exchange are similar. The same type of behavior is observed in the effect of temperature as shown in Fig. 2.6. Figure 2.7 shows that as the pH of the ion exchange solution is increased, a corresponding increase in the extent of calcium exchange is obtained since some of the phenolic groups are also being exchanged as well as carboxylic-acid groups. The total acid-group content of PSOC 680, the bituminous coal, is determined to be 2.41 meq/g with 0.41 meq/g of carboxylic-acid groups. Table 2.2 summar- izes the results of partial oxidation of the bituminous coal sample at 195°C and 260°C for various periods of time. Both carboxylic-acid-groups and phenolic hydroxyl group contents are observed to reach maximum values as the time of oxidation was increased from 2 to 24 hrs at a reaction temperature of 195°C. Both acid-group contents are also observed to decrease as the reaction time was increased from 2 to 8 hrs at 260°C. Figure 2.8 shows the experimental data for coal samples oxidized at 195°C. The results indicate that the extent of exchange achieved in a given time increases with the time of oxidation. The rates of calcium-ion exchange are “Cy429] N T°O UL Jo22 pue €°g Hd Ye €29 I0Sd $O abueyoxa wnoLed yo Jua}Xa ay} UO aZLS aLILqued Jo }d9Js39 p'z ‘BLY (NIW) SWIL 16 S¢2'0 Of Se Oe Gl Ol S 0 — . i S°0 Fon) er ~~ x2 8 SL°0 Cc OTr 4S°¢ x HSSW O8Xx0S HSSW OO0¢exXO0¢c! * a ! (Le02°B/*Ainba*bu ) AJNNVHIXY 30 LNALXA 17 "€°8 Hd ye (Ysa QOZXOZT) €29 I0Sd 40 abueyoxa WNLILRD JO JUAPXS BYP UO UOLZEUZUBDUOD WNLI[PD JO POIZIA G’z “Bly 0 c (NIW) JWIL 09 oS OP os 0 l T l S2°0 71 . a ™ aD 8 GL°O ¢ OTL 44V) dO NOTLVYLNAINO? J 1 tL NO°* 1 NOL°O NSO°O I Vv x a (,e09°6/*Atnba*bu ) JINVHOXG dO LNALXY 18 "[4+429] N T'O pue €°8 Hd Fe (Ysa QOZXOZT) €29 30Sd $0 abueydxa WNLo_ed Jo YUazZXd JY UO auNzpeuadWiay JO 499447 QZ “HL (NIW) 3WIL o¢ G2 02 Sl Ot G 0 or I T l T T #0 m x HS°O mt G2°O ar oO “TI hom nm SOL QO i = 3 mM a _, GLO a 42 2 f cans . ~ OTF 4S°% 5 do IG * x 5 Jocc * a ~ i | l | | 19 “£4429] N T°O pue Jo.zz2 32 (YSeW OOZXOZT) E29 I0Sd $0 abueyoxa winto,ed yo yuayXa ayy uo Hd yo zYDaJJQ =oZ*Z BLY (NIW) SWIL 09 OS Ov o¢ 02 Ol 0 T T T T T —#O SZ°0 See 4S2°e x S°Olt Hd * £°8 Hd ° | l | ia! (,e09°6/*Atnba*6u) JINVHOXY 40 LNALXJ 20 Table 2.2 Effect of Partial Oxidation on the Content of Acidic Groups in PSOC 680 Temperature Time of Carboxyl Hydroxyl Total of Oxidation Oxidation Groups Phenolic Groups Acid Groups (°C) (hr) (mg.equiv./g.) (mg.equiv./g.) (mg.equiv./g) - - 0.41 2.00 2.41 195 2 0.59 2.79 3.38 195 8 2.21 3.54 5.75 195 24 2.13 4.06 6.19 260 2 2.31 3.73 6.04 260 8 1.89 3.00 4.89 21 [4480] N T'O pue e°8 Hd 32 089 90Sd 40 abueyoxa WNL d [ed JO JUaIXI 94} UO J,GGT 32 UOLZEPLXO yo ZDaJyR 8°Z “BLA (NIW) WIL 09 OS Ov o¢ q | rt pS pa ooo Fg ee po 8 Eee Ree 4 ae OP I = ee enw ,Q28" aid wn p20 | 15961 LV NOILvaIXO 2 yy g + v 4S wut Myo og | | ¢{e09°6/*Atnba* Bu) QT X 39NVHOX9 40 LN3LX] 22 “[4480] N T'O pue c°g Hd 32 089 30Sd 40 abueyoxa WNLILED JO ZX JY} UO JQGHT 7 UOLZEPLXO JO 39149 6° “BLA (NIW) SWIL os Ov of 02 Ol 0 nd bee l l I I 00 l ;O1 201 (NIW-1¥00°9/D3) jot x ALvd 23 [4420] N T’O pue e°g Hd 22 089 30Sd go abueyoxa wNLd,ed JO Ua}Xa ay} UO 9,09Z 72 UOLZEPLXO 4O YDasSSq OT'2 “Bld (NIW) 3WIL 09 Os Ov o¢ 02 Ot 0 8° Vv74S 9092 Lv Notivaixo { a a MY ° oO I ! { | 1 9 x (1v09°9/D3) pol X FONVHOXA JO LN3LX3 24 “C4420] N T°O pue €°8 Hd 2e 089 J0Sd yo abueyoxa WNLILeD JO a}eA ay} UO 3,99Z 7B UOLZEPLXO JO 4Da44q TI'Z “BL4 (NIN) JWIL 09 OS Ov o¢ 02 01 0 T T T T T 001 (NIN-1¥09°9/03) /OT X Alva i ly Noluvaixo (YH 8° Vv g Z 092 {iy 7 + MY * a X | | i I | y Ol 25 shown to increase as the time of oxidation increases, Fig. 2.9. Similar results are observed for coal samples oxidized at 260°C as exhibited in Figs. 2.10 and 2.11. 2.4 Discussion Calcium-ion exchange of acid groups in coal with calcium acetate may be written as a reversible chemical reaction: 2RH + Ca(CHgCO0), {* CaRg + 2CHsCOOH (2-4) where RH is either a carboxylic-acid group or a phenolic hydroxyl group. The equilibrium can be shifted toward the right by neutralizing the acetic acid formed during the reaction. If the pH value of the exchange solution is main- tained at pH 6.2 to 8.3, then complete exchange of carboxyl groups may be achieved (Schafer, 1970). At pH values higher than 8.3, phenolic hydroxyl groups, weaker acids than carboxylic acids, will also undergo ion exchange. In terms of the ionic reaction, the equilibrium can be expressed as: 2R7H*(s) + Ca**(aq.) . Ca**(R7)g + 2H*(aq.) (2-5) The overall ion-exchange mechanism may be divided into five steps: i. Diffusion of calcium ions through the solution up to the coal particles. ii. Diffusion of calcium ions inside the coal particles. iii. Chemical exchange between calcium ions and acid groups inside the coal particles. iv. Diffusion of the displaced hydrogen ions out of the interior of the coal parti- cles [reverse of step (ii)]. v. Diffusion of the displaced hydrogen ions through the solution away from the coal particles [reverse of step (i)]. Since the hydrogen ions released during ion exchange are neutralized readily 26 with the addition of Ca(OH)z, the cation exchange resembles an acid-base reac- tion which is generally very fast. Thus, step (iii) is expected to be fast in com- parison to diffusion processes, and the rate-controlling mechanism should be either film diffusion, steps (i) and (v), or particle diffusion, steps (ii) and (iv). Each of the diffusion mechanisms consists of two coupled processes due to con- ditions of electroneutrality surrounding the coal particles which cannot be violated. 2.4.1 Reaction Rate of Ion Exchange The reaction rate for calcium-ion exchange of coal is not very well under- stood. If the reaction rate is assumed to follow the mass action relationship in Eq. (2-5), then the rate expression can be approximated by ~ GRED = ig RH}*[Cat*] - ky[CaReI [HF 6) where k; and k, are the forward and reverse reaction rate constants, respec- tively. Mass balance of the ion-exchange sites gives [RH]o = [RH] + 2[CaRg], (2-7) where [RH]p = initial condition concentration of exchange sites, mol/l. [RH] = concentration of exchange sites at time t, mol/l. [CaRg] = concentration of exchanged sites, moles/I. The ion-exchange experiments were conducted at constant pH values, thus [H*] is constant. The exchange solution also had an excess of calcium ions in order to keep [Ca**] relatively constant throughout each experiment. Examination of the reaction rate terms in Eq. (2-6) reveals that the forward reaction rate is much faster than the reverse reaction rate because [H*] is 27 several orders of magnitude smaller than other terms in the expression. Thus, the reaction rate can be approximated by an irreversible second order reaction in [RH]? with — d@[RH] = ky [RH]? where ky = k,[Ca**]. The rate equation can be integrated to yield 1 1 * - = kyt, [RH] [RH]o , (2-9) which is rearranged to 1 — 1° T-F 1 = ky[RH]ot (2-10) [RH]o - [RH] 4, where F = is ine fractional conversion of the ion-exchange sites. [RH]o The ion-exchange experiments for PSOC 623 (120x200 mesh) conducted at pH 8.3 are analyzed according to the second-order rate expression given in Eq. (2- 10). Figure 2.12 shows that the reaction does conform to second-order kinetics for a time interval up to about seven minutes. Analysis of the slopes of the straight line fit indicates that the second order reaction rate constant is 17 l?mole*sec™ at 22°C and 49.7 l?mole~*’sec"! at 51°C. The activation energy is estimated at approximately 7.0 kcal/mole. The deviation of the reaction rate from second order kinetics suggests that the actual reaction rate may be more complicated than a simple second order reaction with respect to the concentration of ion-exchange sites, however chemi- cal kinetics does not appear to be the rate-limiting process at higher time inter- vals. Figure 2.13 shows that the rate of ion exchange does not vary with tem- perature, which also supports the observation that the rate is not controlled by chemical kinetics. Experiments conducted with a larger particle size, 50x80 mesh, revealed a strong dependence of the rate on particle size, as shown in Fig. 1/ (1-F) 28 PSOC 623 120X200 20 7 | ‘ | ‘ ml , | ‘ Oo - 22° A ~- 51° 16e = / ler 4 y “ f / a / 4 7 / Br / - As Vid 7 Va “7 gee 4-e Lc oO _ - ao 2g fg 7 0 n | 4 l L | " I n 0 10 20 30 40 50 T (MIN) Fig. 2.12 Test of second order dependence of calcium exchange of PSOC 623 with respect to the concentration of carboxylic-acid groups. 29 "€29 JO0Sd 50 abueyoxa wld, ed 40 a3e4 ay. UO duN}euaduia}? Jo 799aJ43 (NIW) 3WIL o¢ GZ 02 St 01 S €I°2 “Bly 001 — | J J ] ai TIT 7 oO ° N N 1 we L. 101 Losi ty 201 cOl (NIW-1¥09°9/04) gol X ave 30 Oe "€29 J0Sd $0 abueyoxa WNLILeD JO 9zZeA ayy UO AZLS aLILYUuRd jo 299d4q PI'Z “BLY Ge (NIW) SWIL O¢ cl Ol S 0 ¥ Trrry? q { ( | L re lori hy i HSSW O8X0S °* x HS3W OOZXOZ1 * a”: I l L ! 2 901 1:01 201 cOl (NIW-1¥O9"9/D3) OT x aLvY 31 2.14. Since a chemical-reaction-controlled rate process would be independent of particle size, the observed dependence of the rate on particle size further indi- cates that the process of calcium-ion exchange of coal is not controlled by chemical kinetics. 2.4.2 Film Diffusion In a film-diffusion-controlled process, there is a film of thickness 6 surround- ing the particles in which a concentration gradient exists. The flux of diffusing Species is (2-11) where D = diffusion coefficient of the diffusing species. C = concentration of the diffusing species. The total flowrate of diffusing species across the film is then = enny| S| ; where Q, = total amount of diffusing species in time t. R = radius of the spherical particles. If the concentration gradient across the film is linear, [2c = = (Cp ~Cs), (2-13) and the concentration of the diffusing species on the surface of the particle, Cs, is equal to zero. Eq. (2-12) may be written as 32 d R oe = 4nR'Do 3 D R= constant , (2-14) where Cg = concentration of the diffusing species in the bulk solution. Equation (2-14) suggests that the extent of ion exchange is a linear function of time. Figure 2.4 clearly exhibits a strong nonlinear dependence of the extent of exchange as a function of time, thus film diffusion is not expected to be the rate-limiting step. The effect of film-diffusion can also be examined by evaluating the rate of mass transfer to the coal particles, r=k,,AC, , (2-15) where km, = mass transfer coefficient, A = surface area of the coal particles. The mass transfer coefficient can be obtained from the expression Nu = Hcp (2-16) where Nu = Nusselt number, assumed to be equal to 2. d, = particle diameter. The experiments were conducted with particles of 120x120 U.S. mesh size, so the average particle diameter is assumed to be 100 wm. The diffusion coefficient of calcium ions diffusing through a solution of 0.1 N calcium acetate can be estimated by (Perry and Chilton, 1973) 33 ém | CyVw (2-17) where D, = diffusivity of the diffusing species in a dilute solution. ¥+ = mean ionic activity coefficient based on molarity. C, = g-mole water /cc of solution. Vq = partial molar volume of water, cc/g-mole. [tw = viscosity of water. j4 = viscosity of solution, The diffusion coefficient of calcium ions in a solution of 0.1 N calcium acetate is then estimated to be 7.1 x 10°§ cem®/sec. The mass transfer coefficient is thus 1.42 x 10° cm/sec. The experiments were conducted with 0.25 g of coal with a density of approximately 1.3 g/cc. So the total external surface area of the coal particles is 115 em®. The estimated rate of mass transfer to the external sur- face of the coal particles is 4.9 x 107 moles/min, or 9.B x 10 eq./min. This rate is far greater than the observed rates of ion exchange. Figure 2.15 also shows that the observed rate of exchange does not vary as the concentration of calcium ions is increased from 0.05 eq/l to 1.0 eq/l. Therefore, film diffusion is not the rate-limiting step for calcium-ion exchange of coal. 2.4.3 Particle Diffusion The diffusion of calcium ions through coal particles in the presence of very fast film diffusion and reaction can be approximated by the shrinking-core model. This model assumes that reaction occurs first at the outer surface of the particle. The zone of reaction then moves into the solid, and leaves behind completely converted material and inert solid, which is referred to as ash. Thus, 34 "€Z9 90Sd $0 abueydxa wntojeo $O 9PRU BYR UO UOLPCUZUBIUOD WNLIELD yo Ydassq GIl°zZ ‘Bt4 (NIW) SWIL 09 OS Ov Os O¢ of 0 I I T T I T | a ¥ T rery tT 90! ,Ol 201 cOl (NIW-1¥00°9/03) gol * aLve 35 at any time there exists an unreacted core of material which shrinks in size dur- ing reaction. If the shrinkage of the unreacted core is much slower than the flow rate of reactant, which is calcium ions for the case of ion exchange, toward the unreacted core, then the unreacted core can be assumed to be stationary in considering the concentration gradient of calcium ions in the ash layer at any time. The rate of reaction of calcium exchange is then given by its rate of diffusion to the reaction surface, dNco, _ amr? _ at mr°Qc, = constant , (2-18) _ and the flux of calcium ions within the ash layer can be expressed by Fick's law, Qc, = De =—, ° dr (2-19) where De = effective diffusion coefficient. C = concentration of calcium ions at a distance r from the center of the parti- cle. Integrating Eq. (2-18) across the ash layer gives — aNca ——- = 4mDeC , where R = radius of the particle. r, = radius of the unreacted core. The final solution is given by Levenspiel (1972) as £21-3(11 -F*4 4 2(1 -F), T (2-21) where 36 F = the fractional conversion of ion exchange sites in coal. The time required for complete conversion of ion exchange sites in a coal parti- cle is r= pak” where Pp = molar density of ion exchange sites in the particle. b = stoichiometric coefficient for ion exchange sites, which is equal to 2 when , the reactant is calcium ion. The experimental data for the calcium-ion exchange of carboxylic acid groups in PSOC 623 are tested against Eq. (2-21), as shown in Fig. 2.16. The results show that the data for the 50xBO mesh coal particles follow the behavior for particle diffusion according to the shrinking-core model. The time for complete conver- sion of carboxylic acid groups is obtained from the slope of the linear relation- ship to be 6.6 hrs. The effective diffusion coefficient is then estimated to be 1.6 x 10°§ cm/sec. The data for the 120x200 mesh particles deviate from the parti- cle diffusion expression of Eq. (2-21). However, a reasonable agreement can be obtained after the reaction has proceeded for longer than 15 minutes. The time for complete conversion based on the data in the linear region is estimated as 3.8 hrs. The corresponding effective diffusion coefficient is approximately 4.9 x 1078 em?/sec., The deviation of the data for ion exchange of 120x120 mesh coal particles from the prediction of the shrinking-core model may be attributed to the effect of electric fields from ionic interactions (Helfferich and Plesset, 1957). The elec- tric potential in the particles increases as the diffusion of ions proceeds, thus resulting in a nonlinear diffusion process with a diffusion coefficient which varies 37 PSOC 623 [Ca**]=0.1N 4 l ! l 0 - 120X200 MESH & - 50X80 MESH 35k _ LOK ~ 20 = a LOK oa 4 - 0 “~ 7 a isk / : iL of 7 A ae 0 ot | | | 0 5 10 15 20 25 TIME (MIN) Fig. 2.16 Test of particle-diffusion limitation on calcium-ion exchange of PSOC 623. 3C 38 with the concentration of ions in the particles. 2.4.4 Effect of Partial Oxidation Partial oxidation of the bituminious coal sample, PSOC 6B0, at moderate tem- peratures in air is observed to increase the content of carboxylic-acid and phenolic hydroxyl groups in the sample as indicated in Table 2.2. The increase in the number of ion-exchange sites makes it possible to achieve higher Ca/S molar ratios in the coal for the purpose of sulfur emission control by the exchanged calcium during combustion of the coal. Table 2.3 shows that the level of Ca/S that can be obtained by exchanging carboxylic-acid groups increases from 0.18 to 1.03 simply by first oxidizing the sample at 260°C for two hours. Exchange of all the acid groups of the same sample increases the Ca/S level from 1.08 to 2.70. The increase in the carboxylic-acid-group content results in the rates of calcium-ion exchange at pH 8.3 as shown previously in Figs. 2.9 and 2.11. How- ever, the corresponding fractional conversions of the carboxylic groups is observed to decrease in Fig. 2.17 as the coal samples are oxidized. Analysis of the data indicates that the rates of ion exchange for the oxidized coals satisfy the criterion of particle diffusion as shown in Fig. 2.18. Table 2.4 shows that the experimentally determined diffusion coefficients for calcium and hydrogen ions in oxidized PSOC 680 are in the range of 5.0 x 107?! to 3.5 x 107!°cm?/sec, which is approximately two orders of magnitude smaller than the diffusion coefficient obtained for diffusion in PSOC 623. Since the electrical field generated by the interaction of ions in the pore structure is most probably the cause of the observed low diffusivities compared to diffusivities of noncharged species (diffusion coefficients on the order of 107? to 10°%em?/sec), the experimental results suggest that the electrical potential in the pore structure of PSOC 6B0 is 39 Table 2.3 Effect of Oxidation on Ca/S Content Obtainable by Calcium-Ion Exchange of PSOC 680 Coal T Time Carboxyl Groups All Acid Groups Sample (°C) (hr) Ca/S Ca/S PSOC 623 - - 3.49 7,32 PSOC 680 - - 0.18 1.08 PSOC 680 195 2 0.26 1.51 PSOC 680 195 8 0.99 2.57 PSOC 680 195 24 0.95 2.77 PSOC 680 260 2 1.03 2.70 PSOC 680 260 8 0.85 2.19 0. 40 PSOC 680 120X200 MESH OS 0.4 I I I T T Oo - -Acid Washed x - Oxidized at 260°C for 2 hrs. (35- 4S - Oxidized at 260°C for 8 hrs. | © - Oxidized at 195°C for 2 hrs. O - Qxidized at 195°C for 24 hrs. 9) 10 20 30 40 50 t (min) Fig. 2.17 Effect of partial oxidation on the fractional conversion of carboxylic-acid groups in PSOC 680. 60 PSOC 680 (120X200 MESH) 4] pH=8.3 4 T — 7 7 7 - ACID-WASHED - OXIDIZED AT OXIDIZED AT - OXIDIZED AT - OXIDIZED AT q0o0epa0 { I ’ oT ' r 260°C FOR 2 HRS. 260°C FOR 8 HRS. 195°C FOR 2 HRS. 195°C FOR 24 HRS Y TIME (MIN) Fig. 2.18 Test of particle-diffusion limitation on calcium-ion exchange of PSOC 680. 42 Table 2.4 Calcium-Ion Exchange of Carboxylic-Acid Groups in Partially Oxidized PSOC 680 (120 x 200 mesh) Oxidative Treatment . D, x 191° (°C) (hr) (hrs.) (cm"/sec) - - 40, 0.78 195 2 89, 0.50 24 88. 3.5 260 2 145. 1.2 8 73. 2.0 43 higher than in PSOC 623, which could be due to a pore-size distribution contain- ing smaller pores. 2.5 Conclusions The kinetics of calcium-ion exchange of coal was studied in an effort to eluci- date the mechanism of ion exchange of coal. Jt was observed that the rate of ion exchange was dependent on particle size and pH value of the solution. Con- centration of calcium ions and temperature of the exchange solution were found to affect only the initial rate of exchange. Analysis of the results indicates that particle diffusion is the rate-limiting step of the overall process of ion exchange of coal and that chemical reaction may be important in the first few minutes of reaction. Application of a particle diffusion criterion shows that the diffusion coefficient of calcium ions in 50x80 mesh PSOC 623, a lignite coal, is approximately 1.6 x 10° cm?/sec. Partial oxidation of PSOC 680, a bituminous coal, was observed to increase both carboxylic-acid and phenolic hydroxyl group contents in the sample so that higher Ca/S5 levels may by achieved by ion exchange. Analysis of the particle diffusion behavior shows that the diffusion coefficient of calcium and hydrogen ions in partially oxidized PSOC 680 (120x200) is in the range of 5.0 x 107"! to 3.5 x 10°'cm*/sec. The observed low diffusivities suggest the possible buildup of an electric field within the pore structure of coal as ion exchange progresses. 44 References Blom, L., Edelhausen, L. and van Krevelen, D. W., Fuel 836:135-153 (1957). Boyd, G. E., Adamson, A. W. and Myers, Jr., L.5., J. Am. Chem. Soc. 69:2836-2848 (1947) Brooks, J. D. and Maher, T. P., Fuel 36:51-62 (1957). Brooks, J. D. and Sternhell, S., Aust. J. Appl. Sei. 8:206-221 (1957). Chang, K. K., Flagan, R. C. and Gavalas, G. R.,, presented at the Western States Section/The Combustion Institute, Palo Alto, October 22-23, 1984. Freund, H. and Lyon, R. K., Comb. Flame 45:191-203 (1982). Fuchs, W. an Sandhoff, A. G., Fuel 12:45-69 (1940). Helfferich, F., ‘lon Exchange", New York, N.Y. (1962). Helfferich, F. and Plesset, M. S., J. Chem. Phys. 28:418-428 (1958). Kalema, W. 5., "A Study of Coal Oxidation", PhD Thesis, California Institute of Technology (1985). Kressman, T. R. E. and Kitchener, J. A., Disc. Faraday Soc. 7:90-105 (1949). Levenspiel, 0., "Chemical Reaction Engineering", 2nd ed., N.Y., 1972. Perry, R. H. and Chilton, C. H., ed. ‘Chemical Engineers’ Handbook”, 5th ed., 1973. Reichenberg, D., J. Am. Chem. Soc. 75:589-597 (1953). Schafer, H. N.S., Fuel 49:197-213 (1970). Schafer, H. N. S., Fuel 49:271-280 (1970). Schobert, H. N., ed., 'The Chemistry of Low-Rank Coals", ACS Symposium Series, No. 264 (1984). 45 CHAPTER 3 COMBUSTION OF CALCTUM-EXCHANGED COALS 3.1 Introduction Sulfur emission contro] in coal combustion is currently practiced by various processes of post-combustion scrubbing. Because of the high cost and complex operation of these scrubbing processes, alternative methods of sulfur removal have been under active investigation. The most successful of these, fluidized combustion in a bed of calcined limestone or dolomite, is already commercial in industrial boilers and might soon become commercialized in utility boilers. At the relatively low temperatures of fluidized combustion, sulfur oxides combine with the calcium or magnesium oxide particles to form stable sulfates. Fluid- ized combustion has several other advantages as well, but its application is lim- ited to new power plants. Other methods of sulfur control suitable for existing pulverized coal-fired boilers rely on injecting the calcium sorbent with the coal and removing the spent sorbent with the ash. At the high temperatures of the pulverized coal flame, calcium sulfate is not thermodynamically stable, hence sulfur oxide removal must exploit either rate limitations in the release of sulfur oxides or the capture of sulfur oxides released in the cooler post-flame part of the furnace, where the equilibrium of sulfur oxide capture becomes favorable. Two modes of calcium sorbent injection have been investigated. The first involves the injection of ground limestone along with the coal either in a conven- tional furnace or in a furnace employing staged combustion (Case et al., 1982). This method has the merit of extreme simplicity and low cost but has yielded sulfur removal in the range of 40-604, insufficient to meet existing environmen- tal standards. Evidently, the kinetics of sulfur oxide capture in the post-flame region of the furnace are not sufficiently rapid as observed by Borgwardt (1970), 46 perhaps due to the slow diffusion within the partially sintered calcium oxide par- ticles (Pigford and Sliger, 1973; Hartman and Coughlin, 1976). Another method that has been investigated involves the calcium sorbent in closer association with the coal particles. One version is to impregnate the coal with a calcium solution and precipitate calcium carbonate right in the pores of the particles. The other method is to add calcium by exchanging acidic groups in the coal with calcium cations, using a suitable solution. This exchange may be represented by 2RH + CA** + RCa*tR” + 2Ht (3.1) where RH is a carboxylic or phenolic group in coal. Lignites have sufficient acidic groups for accepting as much as 6-8% calcium by weight. Bituminious coals do not possess sufficient acidic groups but they can acquire the required ion exchange capacity by mild oxidation (~ 200°C) at the cost of losing some heating value. Adding calcium by ion exchange produces very fine, atomic scale, dispersion in close association with the sulfur source resulting in more rapid kinetics of sulfur capture and reduced diffusional limitations due to deactiva- tion of CaO by solid product layers. Freund and Lyon (1982) studied the combustion of calcium-exchanged coal as a function of fuel equivalence ratio and found sulfur removal as high as 90% at sufficiently high equivalence ratios. At equivalence ratios used in conven- tional combustion the sulfur removal fell to 60% or lower. In view of this, they proposed this mode of calcium addition in connection with staged combustion. In the first fuel-rich stage sulfur would be captured in the form of sulfide. If this fuel-rich stage is followed by an oxygen-rich stage, the sulfide would be con- verted to sulfate provided the temperature remains sufficiently low. 47 In the present work the combustion of calcium-ion-exchanged coals was stu- died in a laminar flow furnace under extreme fuel-lean conditions. The purpose of this study was to determine the processes of sulfur dioxide release and recap- ture in the presence of finely dispersed calcium oxide. The effects of residence time, oxygen concentration, amount of calcium-additive and furnace tempera- ture were evaluated. The results are discussed qualitatively in terms of a specu- lative mechanism for the release and recapture of sulfur oxide. 3.2 Experimental 3.2.1 Coal Preparation The two coals used in this study were PSOC 623, a Texas lignite, and PSOC 680, an Indiana high volatile B bituminous. The coal samples were ground to 230 x 325 U.S. mesh size and treated with dilute acid to remove minerals from the solid. The ground coal was stirred with 1.0 N HCLin a ratio of 10 ml/g for 24 hours at 40°C. The acid treatment was sufficient for the removal of the cations of Ca, Mg, Na and K associated with carbonates, sulfates, clays and organic acid sites. Aluminosilicate material and pyrite are not dissolved but may be removed physically. After the acid wash, the slurry was filtered and rinsed with dem- ineralized water to neutral pH. The powder was then vacuum dried at 105°C for 24 hours. A procedure developed by Schafer (1970) for determining the carboxylic-acid- group content and the total acidity of low-rank coals was adapted to incorporate calcium ions into the coal samples by ion exchange. Bach acid-washed coal sam- ple was mixed with a 0.1 N calcium acetate solution at 25°C for 4 to 24 hours. The pH of the mixture was kept at 7.0 to 11.0 with Ca(OH)zg to control the amount of calcium ion exchanged. After calcium exchange the slurry was filtered and washed with demineralized water to remove excess calcium. The 48 treated coal was then vacuum dried at 105°C for 24 hours. The ion-exchange capacity of the bituminous coal was insufficient to incorporate enough calcium into the coal for a meaningful Ca/5S ratio. Thus the sample was prepared by first oxidizing the coal at 160°C for 65 hours in an air-fiuidized-bed reactor. The oxidized coal was then acid-washed and ion exchanged. 3.2.2 Experimental Apparatus and Procedure The experiments were conducted in a laminar flow furnace as shown in Fig. 3.1. The furnace tube was made of high purity alumina with a length of 60 cm and a diameter of 5cm. The furnace was heated by radiation from six U-shaped molybdenum disilicide heating elements surrounding the alumina tube in a hot zone of 20 cm length. The vertical ternperature profile of the furnace was obtained with a series of six thermocouples attached to the furnace wall as shown in Fig. 3.2 and with a high temperature probe inserted into the furnace from the bottom. A coal feeder, shown in Fig. 3.3, was used to feed coal into the furnace at a rate of 0.3 g/hr via a water-cooled injector located at the top of the hot zone. A small fraction of the total gas flow entered through the injector; the balance of the gas entered in the annular region between the injector and the furnace wall. A low-density alumina disc with 40 equally spaced 0.3 mm holes was placed 5 cm above the injector tip to serve as flow straightener. The residence time of coal particles was controlled by varying the flow of gas through the straightener. The amount of coal combusted during each 4-hour experiment was deter- mined by weighing the test tube containing coal in the feeder before and after the experiment. The cumulative 502 emitted during a combustion run was determined by passing the combustion gas through two gas-washing bottles in series containing a solution of 1% HzOzg as shown in Fig. 3.4. In this medium 502 > Oo 1) Main combustion air 2) Coal and gas 3) Water-cooled injector 4) Flow straightener 5) Heating element 6) Alumina furnace tube 7) Water- cooled sample probe and dilution air Fig. 3.1 Combustion furnace (mot to scale). SA \ “ 7 51 +” FURNACE Hf === To VIBRATOR ~t————— FINE GAUGE TUBING GAS —— IN O-RING SLIDING SEAL TO ee VIBRATOR Tae “he —— TEST TUBE Tre] COAL ET MOVING SYRINGE [ }<— PUMP STAGE Fig. 3.3 Coal feeder (not to scale). 52 *squoutiedxa uotysnquod 10jy waysks mOoTy y'¢ *3Tgq Saaggnyds “os <——°N pesiog pessexucy 4 oe | WH @ O peitiog (d) xX Ox rd To 4IV pel110g @& 53 was converted to HgSO, which was then determined potentiometrically. Sulfur content in the recovered ash was analyzed using a Leco furnace in accordance with the high temperature combustion method described in ASTM Methoa 3177- 75. From the bottom of the furnace an insulated and water-cooled sample collection-probe extended up into the furnace. The inner surface of the sample probe was made of porous stainless steel tube. The combustion gas was cooled and diluted with dried, filtered air through the walls of the probe. A filter assembly with quartz filter paper was placed at the end of the probe to collect ash samples. 3.2.3 Chemical Analysis of Coal 3.2.3.1 Sulfur Determination The total sulfur content of coal and coal ash was analyzed during a Leco fur- nace in accordance with the high temperature combustion method described in ASTM. Method 3177-75. A0.5 g sample was measured into a boat made of refrac- tory material and burned in the Leco furnace at 1350°C with a stream of oxygen. After 15 minutes of combustion the flow of oxygen was substituted by a stream of helium for 10 minutes. The flow of helium was added to the ASTM procedure in order to accelerate the decomposition of sulfates, in particular calcium sul- fate, into 50g and CaO. The sulfur oxides emitted during combustion and decomposition were scrubbed with a solution of 1% hydrogen peroxide, produc- ing sulfuric acid. The hydrogen peroxide solution was then titrated with dilute sodium hydroxide to the initial pH value of 4.5. The sulfur content of the coal or coal ash was calculated from the volume of titrant as follows: _ VweouNnaoy Ss x 1.603 (3.2) 54 where 5 = wt % sulfur in sample. Vweon = Volume of NaOH titrant, ml. NyaoH = normality of NaOH solution, g.eq/liter. w = weight of sample, grams. For coals that contain high levels of chlorine, HCl is evolved during combus- tion and contributes to the acidity of the efflUent gases. Corrections for chlorine is described in the ASTM procedure. The coals used in the present study contained less than 0.02% chlorine, therefore corrections were not neces- sary. The method was calibrated with reagent grade CaSOQ,-2H20, and an accuracy within 1% was obtained. Analyses of coal samples gave a standard deviation of 5%, due mainly to the inhomogeneity of samples. 3.2.3.2 Calcium Determination The calcium content of coal was analyzed by atomic absorption spectropho- tometry. Coal ash was prepared by following ASTM Method D-2795-69. The ash sample was dissolved in hydrofluoric acid, treated with concentrated sulfuric and nitric acids and then diluted to make the solution, which was employed to determine the calcium content of the ash. A series of samples was prepared with various added amounts of a standard calcium solution. The samples were then analyzed by a Varian Techtron Model AA-6 Atomic Absorption Spectropho- tometer. The absorbance for each sample at a wave length of 422.7 nm and spectral band pass of 0.2 nm was obtained in an acetylene/nitrous oxide flame. In the absorbance range of 0.2-0.8, absorbance is directly proportional to con- 55 centration, Xo C, = v_ = mA, ° (3.3) where Co = concentration of calcium in original sample, mg/ml Xp = wt of calcium in sample, mg V, = volume of solution in sample, ml m = proportionality constant, mg/ml A, = absorbance of sample. In samples with added standard calcium solution, Xp + AX; S = yoray, = ™ o FAN; (3.4) where AX; = wt of calcium in added standard solution, mg AV; = volume of added standard solution, ml. The calcium content of the original sample Xp can be obtained by solving the algebraic equation relating a pair of samples with different added amounts of standard calcium solution, Xp + AX; | ion _ mA _ Vo + AV; C; mA Xp + AV; | | (3.5) Vot AV; which yields 56 A; Xi(Vo + AV}) — AX; A, (V, + AV;) a. (Vo + AV)) ~ (Vo + AV) (3.6) i The calcium content of the coal sample is then VolXoNV, o) *Ca = 100w (3.7) where Vr = total volume of the diluted ash solution, ml w = wt of coal, mg. 3.3 Results Table 3.1 lists the results of sulfur and caicium analyses of the calcium-ion- exchanged coals used in the combustion experiments. A comparison of the sul- fur contents of the coals with the analyses provided by the Penn State Univer- sity Coal Bank given in Table 3.2 shows a difference in the total sulfur contents, particularly with the bituminus coal, PSOC 680. This variation in the sulfur con- tent is due mainly to the coal preparation procedures which involve oxidation of the bituminous coal at 160°C for 65 hours prior to acid washing and ion exchange. Since it is unlikely that appreciable amounts of organic sulfur can be removed under the mild oxidizing condition during coal preparation, therefore, the oxidative pretreatment probably oxidized pyritic sulfur to sulfates which is subsequently removed during acid washing. The agreement in the organic sulfur content of the original bituminous coal with the total sulfur content of the treated coal suggests that most of the sulfur in the calcium-ion-exchanged coal is in the form of organic sulfur species. The results of sulfur emissions observed during the combustion experiments are normalized to the total sulfur of the calcium-treated coals as given in Table 3.1. 57 Table 3.1 Sulfur and Calcium Analyses of Coal Samples Used in the Combustion Experiments Wt% Dry Basis Molar Ratio Ash | S ‘Ca Ca/S. PSOC 680 21.1 1.64 7.07 3.45 PSOC 623 A 22.4 0.81 2.50 2.46 PSOC 623 B 20.2 0.81 0.82 0.81 PSOC 623 Acid-Washed 13.4 1.00 0.06 0.05 Table 3.2 Sulfur Analysis of PSOC 680 and PSOC 623 on a Dry Basis Provided by PSU Coal Bank % Pyritic % Sulfatic % Organic % Total PSOC 680 1.60 0.32 1.66 3.58 PSOC 623 0.19 0.12 1.01 1.35 58 Figure 3.5 shows the retention of sulfur during combustion of the calcium- exchanged bituminous coal, PSOC 680 (Ca/S = 3.4), at a furnace temperature of 1600°K. The extent of sulfur capture was reduced greatly as the gas flowrate increased from 0.5 l/min to 4.1 L/min corresponding to residence times of 9 seconds to 1.1 seconds in the hot zone of the furnace. The results also show that as oxygen concentration was increased from 13% to 40%, sulfur capture was increased correspondingly. In all experiments, combustion of the coal was com- plete since analyses of the ash samples showed no residual carbon. Sulfur ana- lyses of ash samples from the experiments with PSOC 680 in 40% Os are presented in Table 3.3. Sulfur balance varied from 115.7% at a flowrate of 4.1 t/min to 86.7% at a flowrate of 0.51/min. The variations in sulfur may be par- tially due to sample inhomogeneity. Low sulfur balances may also be due to par- tial loss of ash deposited on the insulation for the sample probe. Figures 3.6 and 3.7 show the effect of increasing calcium content on sulfur eapture during the combustion of the lignite coal, PSOC 623. As Ca/S ratio increased from 0.81% to 2.46%, sulfur capture increased proportionally. Figure 3.8 shows the retention of sulfur during combustion of PSOC 680 (Ca/S = 3.4) at a furnace temperature of 1400°K. The results indicate that the effect of gas flowrate on sulfur capture is not as great as at a furnace tempera- ture of 1400°K compared to the results obtained at a furnace temperature of 1600°K. 3.4 Discussion The evolution of sulfur during combustion of coal may be described by the following steps: 59 PSOC 680 - Ca/S=3.4 - 1606 120 T T T T 0 - 40% Q2 x = 31% Q2 OS - 21% O2 190rF Oo - 13% Q2 ~ _ o S Bor - bud oO - & 60F \ ’ - oO oa => Le S 40F a) 20 0 0 FLOWRATE (L/MIN) Fig. 3.5 Effect of gas flowrate and oxygen concentration on sulfur retention during combustion of calcium-ion-exchanged PSOC 680 (Ca/S = 3.4) in a laminar flow furnace. 60 Table 3. 3 Sulfur Balance for the Combustion of PSOC 680 (Ca/S = 3.4) at 1600°K in 40% 0 2 Gas . Flowrate % S in Gas %S in %S (L/min) as SO, Ash Balance 4.1 103.3 12.4 115.7 2.1 79.9 17.4 97.3 1.0 40.5 52.7 93.2 0.5 16.3 70.4 86.7 61 -PSOC 623 - Ca/S=0.8 - 16006°K 120 T T I T Oo - 40% Oz x oo- 31% Q2 190Fr ce Ss 80br 4 Lu oc > joe & 60F cm ra) oc a) Lo = 40 - wn 20+ \ ~ Aa 0 1 2 3 4 5 FLOWRATE (L/MIN) Fig. 3.6 Effect of gas flowrate and oxygen concentration on sulfur retention during combustion of calcium- ion-exchanged PSOC 623 (Ca/S = 0.8) in a laminar flow furnace. 62 PSOC 623 - Ca/S=2.5 - 1606°K 120 T T T T A - 21% Qo 100k Oo - 13% 0; 4 SN Boe 4 Lu oO > — & 60- ao = (=) cc x => ion (ea) w 20r 4 i) 0 5 FLOWRATE (L/MIN) Fig. 3.7 Effect of gas flowrate and oxygen concentration on sulfur retention during combustion of calcium- ion-exchanged PSOC 623 (Ca/S = 2.5) in a laminar flow furnace. 63 “ PSOC 686 - Ca/S=3.4 - 1400°K 120 T T ] Oo - 40% 02 10GR = ee S ogoh 7 - 9 Lo oc > =< 60 oO = od Qo fad Z 4 = 40r 4 4 op) 4 4 20r 4 0 I I i I 0 1 2 3 4 5 FLOWRATE (L/MIN) Fig. 3.8 Effect of gas flowrate on sulfur retention during combustion of PSOC 680 (Ca/S = 3.4) at a furnace temperature of 1400°K. 64 i. Reducing Conditions: R-S +> HeSt + R’ R-Ca > Cad +R’ CaO + H2S ~ CaS + H20 CaO + S02 + 3C0 » CaS + 38C0.2 CaO + COS + CaS + CO, ti, Oxidizing Conditions: HS + 2.0, + S02 + H0 R—-S* + Oz > SOs + CO CaS + 202 + CaSO, 3 COS + 5-Op + SO; + COz i CaO + SO, + 5 Oz ” CaSO, formed by the reaction of CaO and SO, in an oxygen-rich environment. (3.8) (3.9) (3.10) (3.11) (3.12) (3.13) (3.14) (3.15) (3.16) (3.17) During devolatilization and early stages of combustion, the interior of the char particles is under reducing conditions, thus sulfur may be evolved as volatile sulfur species or may form calcium sulfide through steps (3-10) to (3-12). As oxygen penetrates the char particles, gaseous sulfur species are oxidized to 50, and calcium sulfide is oxidized to calcium sulfate. Calcium sulfate may also be Reaction (3-17) can proceed in either direction depending on the ternpera- carbonyl sulfide. May also be inorganic sulfur species. ture and the partial pressures of 502 and Og. The calculated equilibrium SO, partial pressures based on reaction (3-17) is plotted against 1/T in Fig. 3-9. * May also be volatile organic sulfur species such as mercaptans and (PPM) Ps O02 65 T(°K) 1500 1400 1300 104 . : —T T | : Fo oo - 10% Op : - A - 20% Oo ~ i © - 40% 02 7 fr O - 60% Q2 1 109 4 102? “q 10h L s _ - Oo 4 190 { i a i | 9-95 6 6.5 7 7.5 8 8. 1/T X10% (°K"!) Fig. 3.9 Calculated equilibrium SO, partial pressure for the : 1 reaction CaO + sO, +5 0, = CaSO, . 66 During combustion experiments with a fixed coal feedrate of 1.2 g/hr, the SOg2 partial pressures ranged from 20 ppm at a gas flowrate of 4.1 2/min to 160 ppm at a gas flowrate of 0.5 2/min. Thus the formation of CaSO, will be unfavorable thermodynamically at temperatures above 1350°K to 1450°K as flowrates vary from 4.12/min to 0.51/min. Figure 3.9 also shows that as oxygen partial pres- sure increases, equilibrium SOz partial pressure decreases correspondingly. At the heated section of the furnace the wall temperature is maintained at 1600°K and the particle temperature is considerably higher. At these temperatures all sulfur species would be released as SOg under conditions of equilibrium, thus sulfur capture probably occurs when the particles are cooled downstream of the hot zone such that the formation of CaSQ, is thermodynamically favorable. 3.4.1 Effect of Residence Time The temperature profiles of the furnace at the experimental flowrates are presented in Fig. 3-10. The thermodynamically favorable zone for the formation of CaSO, begins as the furnace temperature falls below 1400°K, which is approxi- mately 27 cm from the injector. The zone ends as the particles reach the sam- ple collector which is 34 cm from the injector. The reaction rate for the forma- tion of CaSO, from CaO and SO, decreases rapidly as temperature decreases below 750°C (Borgwardt, 1970). Thus the introduction of cool dilution air through the sample probe serves to quench further reactions as the furnace temperature is lowered to 1000°K and SOz2 concentration is reduced. The relevant temperature for the formation or decomposition of CaS0, is the particle temperature. Calculations of the particle temperature were made with a coal combustion model developed by C. L. Senior (1984) based on the work of Libby and Blake (1979). The calculated peak particle temperatures are plotted against oxygen concentration in Fig. 3.11. The times required for 95% burnout (°K) 2000 67 FURNACE TEMPERATURE PROFILE 1750r 1S0C Re 1250 Op xo I Lf 1 L/MIN 1 L/MIN O L/MIN 5 4, 2. i. QO. L/MIN i i ! I T | aN 10 1S 20 DISTANCE FROM INJECTOR (CM) Fig. 3.10 Furnace temperature versus distance from coal injector. (°K) PEAK TEMPERATURE 68 CALCULATED PARTICLE TEMPERATURES 2600 T T 2406 4 o 2206+ / - / 5 2060 S _ / 1859 ao we] — —_j a” a 1600F - 1400 l l l 0 19 20 30 40 50 % OXYGEN Fig. 3.11 Theoretical calculation of particle temperature for the combustion of a bituminous coal at a furnace temperature of 1600°K. 69 of carbon were calculated to be 25 msec for combustion with 40% Og to 6B msec with 13% Og at a furnace temperature of 1600°K. The particles are combusted completely and cooled to the furnace temperature in approximately 100 msec to 200 msec, only a fraction of the particle residence time in the hot zone which ranges from 1.1 sec to 9 sec. Therefore, the particle temperature through the lower portions of the furnace was assumed to be similar to the furnace tempera- ture as illustrated in Fig. 3.12. The particle residence time through the zone where the temperature favors capture of SO2 by CaO was calculated for each of the flowrates. The results for the combustion experiments conducted at a furnace temperature of 1600°K showing the effects of the redefined particle residence time are presented in Figs. 3.138 - 3.15. In all cases, the extent of sulfur capture increases with a corresponding increase in particle residence time. Figure 3.12 shows that the furnace and particle temperatures range from 1400°K to 1000°K in the reactive zone of the furnace which is approximately 7 cm long, as designated by D' where the temperatures are lowered to 1400°K, and D" where cool dilution air is introduced through the sample probe. Thus the observed sulfur capture is the cumulative result of the reaction of SOz and CaO with Og over a temperature range of 400°K. In the case of strongly fuel-lean conditions, the oxygen concentration is in excess and remains relatively constant throughout the furnace. Thus if the reaction between the gas and CaO is of first order with respect to SO2 (McClellan et al., 1970), then the reaction rate can be expressed as dCso, ~ Tap = K Gsoe: (3-18) which gives 70 ~~ ea, a \ OS™ ol lentes ee FURNACE TEMPERATURE pony CASO), STABLE PSS cor- - mt = v D’ DISTANCE FROM INJECTOR Fig. 3.12 Illustration of particle temperature profile (4%) SULFUR CAPTURE 71 PSOC 686 - Ca/S=3-.4 - 1600°K 1290 T T T T T x o- 3i%Z Op 4 - 21% O2 1OGF oO - 13% 0: = D Sor a — Jo x 60 Oo + / 40- / - 20r o 4 0 i i 1 0 1 2 x) 4 fs) 6 RESIDENCE TIME (S) Fig. 3.13 Effect of particle residence time on sulfur retention for the combustion of PSOC 680 (Ca/S = 3.4) at various oxygen concentration. (%) SULFUR CAPTURE 72 PSCC 623 - Ca/S=0.8 - 1600°K 120 7 T o - 40% O> 4 - 21% 02 [O0F 4 BOF _ 50r 4 40r 4 | “| ——— 4 ae ts 0 P L l l 0 j 2 3 4 5 6 RESIDENCE TIME (5S) Fig. 3.14 Effect of particle residence time on sulfur retention for the combustion of PSOC 623 (Ca/S = 0.8) at various oxygen concentration, (%) SULFUR CAPTURE 73 PSOC 623 - Ca/S=2.5 - 1600°K 120-4 Do - 40% 0, x - 31% Oo S - 21% Qo 100k o - 13% Q3 _ 80F _ 60 _ 40h 7 20Fr _l 0 0 6 4 RESIDENCE TIME (S) Fig. 3.15 Effect of particle residence time on sulfur retention for the combustion of PSOC 623 (Ca/S = 2.5) at various oxygen concentration. Co, (3-19) where Cso, = the concentration of 503 in the gas. k* = the first order reaction rate constant which includes the effect of oxygen concentration, the properties of the ash particles (e.g. calcium content in the ash), and is averaged over the temperature range of the reactive zone —from 1400°K to 1000°K. The observations at high flowrates, or low residence times, show little or no sul- fur capture. This result suggests that all the sulfur in the original coal sample is in the form of 50, therefore the concentration of SOg in the gas before recap- ture occurs, CS, can be calculated from the rate of coal combusted and the flowrate of gas through the furnace. The SQz2 analyzed in the exit gas gives the concentration at the quench point, assuming reaction does not continue beyond the sample probe. Figure 3.16 shows a linear dependence of In(Cso,/-$o,) with respect to time for sulfur capture during combustion of PSOC 680 (Ca/S = 3.4). Figures 3.17 and 3.18 indicate similar linear relationships during combustion of PSOC 623 (Ca/S = 2.5) and PSOC 623B (Ca/S = 0.8), respectively. These results show that sulfur capture is a first order reaction with respect to 502 in the gas. According to Eq. (3-19), the slope of the straight line of In(Cgo,/go,) versus time is the negative value of k’, the first order reaction rate constant. The experimentally determined first order reaction rate constants are tabulated in Table 3.4. There is no variation in the reaction rate constant at 18% and 21% Oz concentration, suggesting a zeroth order dependence on oxygen at low oxygen 75 PSOC 686 - Ca/S=3.4 - 1600°K 1.5 T oO - 40% Q2 xoo- 3i% 0: 4 - 21% 02 0.75 o - 134 0: Ok ~O0.75+ -1.5- -2.25b ~3 ] l ! 0 ] 2 3 4 5 TIME (S) Fig. 3.16 First order dependence of sulfur capture with respect to SO concentration during combustion of PSOC 680 (Ca/S = 3.4) at 1600°K. (C/Cg) Ln 76 PSOC 623 - Ca/S=2.5 - 1600°K Op x 0D 1 i { I I j { i i | | Fig. 3.17 1 2 3 4 fs) 6 TIME (S) First order dependence of sulfur capture with respect to SO concentration during combustion of PSOC 623A (Ca/S = 2.5) at 1600°K. (C/Cg) 77 PSOC 623 - Ca/S=0.8 - 1600°K lee 1 O - 40% OQ, A - 214 OQ: 0.8b a 0.4- _ an SS ; —_== =. -O.4F - TIME (S) Fig. 3.18 First order dependence of sulfur capture with respect to SO2 concentration during combustion of PSOC 623B (Ca/S = 0.8). 78 Table 3.4 First order reaction rate constants for the retention of SO, during combustion of coals at 1600°K. PSOC 680 Py (atm) (Ca/S = ae 2 K*, (sec) 0.40 0.40 0.31 0.21 0.21 0.034 0.13 0.035 PSOC 623A (Ca/S = 2.5) 1) * : = K , (sec PSOC 623B (Ca/S = 0.8) 1) * - K , (sec 0.17 0.12 0.061 0.065 0.067 0.052 0.030 79 levels, As oxygen concentration increases above 21%, the reaction rate constant also increases. The observed trends of reaction rate constant and oxygen con- centration suggest that the order of reaction is not constant with respect to oxygen concentration at higher oxygen concentrations. The effective reaction time for the formation of calcium sulfate in the present study ranges from 0.6 seconds to 5.5 seconds. The short reaction times for all of the experiments combined with the the observed results of little or no sulfur capture at 0.6 second raise the possibility of diffusional resistances for the reac- tion of SOg with CaO dispersed in the coal ash. The diffusional processes are: the diffusion of SO2 through the bulk gas, the diffusion of SO, through the pores of the ash particles, and possibly the diffusion of SOg through a shell of solid ‘reaction product which forms on the active solid surface as the reaction proceeds. The characteristic time for the diffusion of SOz through the bulk gas to the ash particles can be approximated by dD’ (3-20) where d\ = the distance that 50, has to diffuse through the bulk gas to reach the ash particles. D = the mass diffusivity of SOz2 through the bulk gas. An upper limit on tp can be estimated by examining the experimental conditions at a gas flowrate of 4.1 l/min. In an average experiment, 1.3 g of coal is fed into the furnace over a period of 4 hours. The average particle size is 50 um in diam- eter, thus there are approximately 1.5 x 10” particles fed through the furnace along with 9 x 10° cm of gas. If we assume the particles do not break up and 80 the particles are distributed uniformly in the furnace, then an upper limit for A is approximately 0.2 cm. The mass diffusivity for binary mixtures of gases can be calculated by the Chapman-Enskog formula for gaseous state at low density (Hirschfelder et al., 1954), 3 [i 1] EF * ip | pois? pap (3-21) Dap = 0.0018583 in which Dap = the diffusivity of A in B. p = the pressure of the system. Gap = the Lennard-Jones collision diameter. Q pap = a dimensionless function of the intermolecular potential field for one molecule of A and B. For SOz in air, D varies from 1.05 em®/sec at 1000°K to 1.84 cm?/sec at 1400°K. The upper limit of the bulk gas diffusional characteristic time is then 38 msec, which is considerably faster than the reaction time scale. This result suggests that the diffusion of 502 through the bulk gas does not pose a resistance to the formation of CaSO, if the assumption of uniform distribution of particles is valid. Diffusion of SOg within a spherical particle of ash in the absence of chemical reaction may be expressed as: ac _ De 8 [2 de ~ 2? @r ér |’ (3-22) where 81 c = the concentration of 5Oz2 at time t and radial position r in the spherical particle. D, = the effective diffusivity. If the concentration of SOg is assumed to be uniform at the surface of the parti- cle, then the boundary and initial conditions can be approximated by: c=c, at r=R (t= 0) ce=0 for t< 0 (O< r< R) where c, = the concentration of SO2 at the surface of the particle. R = the radius of the particle. The equations may be non-dimensionalized to give: ge - 1 8 je oe at r° oF ar | (3-23) t=1 at r=1 (t= 0) @¢=0 at t< 0 (O< F< 1), with =.~.¢ YP z t eé=—,r=s andt= , Cs R tp, where tp, = Er the characteristic time for intraparticle diffusion. e The solution of Eq. (3-23) is (Crank, 1956): — hed —1 jn art) = et) - 44 2_y¥ (1)? sin(nnte Pn Cs TH ns. 1 (3-24) At the center of the particle (F = 0), the solution reduces to: (3-25) The concentration of SOz at the center of the particle reaches 99% of the value at the surface when t ™ 0.55 tp, The effective diffusivity of S02 through the ash particle can be estimated by applying the parallel-pore model for a mono- disperse system: a (3-26) where & = porosity of the ash particle. o = tortuosity factor. D = diffusivity of 502 in the ash particle. If the diffusion of SOg in ash particles can be approximated by the diffusion of 502 in calcined limestone, then the value of D/ at 900°C was calculated by Hart- man and Trnka (1980) to be 0.081 cm®/sec, based on the data of Campbell et al. (1970) for the Knudsen diffusion of carbon dioxide in porous limestone with a tortuosity factor of 1.5. The ash from the combustion of 50 wm coal particles is present either in discrete particles of up to 5 wm in diameter or in an agglomerate of many individual particles of approximately 40 wm in diameter as observed under a Scanning Electron Microscope. The porosity of individual par- ticles is very low due to melting of ash under high temperatures, On the other hand, the porosity of agglomerates is much higher due to void spaces between particles. If the general physical properties of ash are assumed to be similar to those of calcined limestone, then the porosity can be estimated as 0.5. The effective diffusion coefficient of SOg through ash is then D, = 0.041 cm®/sec. The characteristic time for diffusion of 502 in a 40 wm ash agglomerate is then 83 estimated to be to, % 10 sec. A similar characteristic time for diffusion in a 5 pm ash particle would put the porosity at 0.03, which is possibly within order of magnitude of the actual porosity of ash particles. The diffusion of 50, through a layer of solid reaction products in the sulfation of calcined limestone was determined by Pigford and Sliger (1973) to be negligi- ble for particles smaller than 100 wm. Since the ash particles in the combustion of coal are generally smaller than 50 ym, it is assumed that diffusion of 50, through solid product layer is negligible. The small characteristic times for diffusion suggest that chemical reaction is the rate-limiting step for the capture of SO in the case of uniformly distributed ash particles. Examination of the ash samples collected in the absence of dilution air revealed a nonuniform distribution of ash deposited on the quartz filters. The majority of the ash particles appeared to be located near the center of the filter in a circular pattern. Thus it is possible that the distribution of ash particles is not radially uniform throughout the furnace, but is actually concentrated in a cylinder extending from the coal injector to the sample collector. This pattern of ash distribution does not affect the characteristic times for diffusion of 50, through ash particles, however it does change the characteristic time for diffusion of 50, in the bulk gas. The diffusion of SO, through the bulk gas of the furnace in the absence of chemical reaction can be expressed in cylindrical coordinates as: dc 1 9a | ac | — = Dp —ir— where 84 ce = concentration of 502 at a distance r from the centerline of the furnace in time t. D = mass diffusivity of SOg through the bulk gas. If chemical reaction is fast relative to gas diffusion (a simplifying assumption in order to examine characteristic time of gas diffusion), then the initial and boun- dary conditions are: 1.C.: ¢ = C§o, at t = 0. B.C.i:c=Oatr=R, (t> 0). B.C.2: dcfr = Oatr=R, (t> 0). where Co, = concentration of SO, in gas phase before entering reaction zone. Ry = radius of the furnace tube. R, = radius of cylinder in which ash particles are concentrated. Using the method of separation of variables gives: e(r.t) = R(r)T(t). (3-2B) Equation (3-27) is transformed into RT = D(R'T+ ERT). r (3-29) or TLR IR.» DT R ‘rR x (3-30) where A is an eigenvalue. The time dependence of the solution can be obtained from T +DNVT = 0 (3-31) 85 which yields T(t) = aye Dat ; (3-32) where a, is a constant of integration. The other half of Eq. (3-30) gives r*R" + rR'+)?r?R = 0, (3-33) the solution of which is: R(r) = agdo{Ar) + agYo{Ar) , (3-34) where ag and ag are integration constants. The boundary conditions give (3-35) and B.C.2: —agAJ,(AR;y) + agAY,(AR;) = 0, (3-36) which yields the spatial dependence of the solution: Yo(AR Jo{AR;) (3-37) The general solution is obtained by combining Eqs. (3-32) and (3-37), = Yo(AnR:) | pret e(r,t) = A Jo(Anr) — Yo{Anr) fe ” The furnace radius R; is 2.5 em and if R, is assumed to be 1 cm, then the eigen- values can be solved from Eqs. (3-35) and (3-36). The first two eigenvalues are d, 0.9 em! and Ax ¥ 3.1em7!. Since Af is sufficiently larger than A?, the gen- eral solution can be approximated by the first term in the series: Zo\Ar) _ Daft e(r,t) = A; (3-38) The initial conditions yield 86 a = a f¥oNRD e(r.o) = Cho, = A, [Fo sue ~ Yon). (3-40) Combining Eqs. (3-39) and (3-40) gives the relative concentration of S02 as a function of time, eft) _ ,-0.8in Co, (3-41) The decrease in SOg concentration is plotted versus time in Fig. 3.19, which covers the range of temperatures in the reaction zone. At an average tempera- ture of 1200°K, the concentration of S02 in the annular region between the cylinder of ash particles and the furnace wall decreases 50% in 0.6 second and 90% in 2.0 seconds. Since the reaction time for CaSO, formation in the combus- tion experiments ranges from 0.6 second to 5.5 seconds, the diffusion of SOg in the gas phase can present a resistance to the reaction rate if the distribution of ash particles is centered in the inner region of the furnace. 3.4.2 Effect of Oxygen Concentration The effects of varying oxygen concentration on sulfur capture are also exhi- bited in Figs. 3.13-3.15. The results show that as oxygen concentration increases, there is a corresponding increase in the observed sulfur capture. The experimentally determined first order reaction rate constants for reaction (3- 18) are plotted in Fig. 3.20. For the three coal samples examined in this study, the first order rate constants remain constant as oxygen concentration is increased from 13% to 21%. This trend suggests that sulfur capture is zeroth order with respect to oxygen in the range of 13% to 21% oxygen. However, at oxygen levels above 21%, the first order rate constant increases as oxygen con- centration in the gas increases. Analysis of the effect of oxygen reveals that the order of reaction is not constant with respect to oxygen concentration. 1. 87 = 1000°K T = 1200°K 1400°K 0.5 1.0 1.5 2.0 2.5 Time (sec) Fig. 3.19 Relative concentration (C/Co ) of SO, as a function of time for the diffusion of S0,*through bulk gas as described in equation (3-41). (SEC™!) K* CA T | T | Oo - PSOC 680 (Ca/S=3.4) x - PSOC 623A (Ca/S=2.5) 4 ~ PSOC 623B (Ca/S=0.8) —f— ray 0 | | 0 0.1 QO. 3 Q. 53 0.4 Q. PO. (ATM) 88 FIRST ORDER RATE CONSTANTS Fig. 3.20 The effect of oxygen on first order reaction rate constants. 5 89 Although S02 absorption on CaO in the ash according to reaction (3-17) appears to be the dominant mechanism of sulfur capture, other mechanisms may also occur. One possibility is the vaporization of calcium in a high- temperature environment such as that surrounding a burning char particle (Quann and Sarofim, 1982). The calcium vapors diffusing outside the particle would reoxidize to form CaO aerosol particles of very high surface area that would be effective in capturing SOz. Calcium vaporization can occur by the reduction of calcium oxide in a reduc- ing environment during the combustion of char particles, CaQ(s) + CO(g) 2 Calg) + COo(g) . (3-42) The equilibrium partial pressure of calcium is determined from the equilibrium constant _ Poo.P ca Poo (3-43) At high temperatures, the primary product of carbon combustion is carbon monoxide (Mulcahy and Smith, 1969). Thus if the only source of COs is from reaction (3-42), the partial pressure of CO, can be assumed to be equal to the partial pressure of calcium. The equilibrium partial pressure of calcium is then Poa = (KP¢o)'# . (3-44) To estimate the amount of calcium vaporized, it was assumed that the rate of vaporization is limited by diffusion from the particle surface to the free stream. In this case the rate of calcium vaporization can be calculated by (Senior and Flagan, 1982) (3-45) 90 where R, = rate of carbon combustion, Yca = mass fraction of calcium vapor at the particle surface, and the Peclet number, P,, at the particle surface is defined as R Pe = 4nr,cD ' (3-46) where rp = radius of the coal particle. c = gas concentration at the particle surface. D = diffusivity of gas at the particle surface. A particle combustion model developed by C. L. Senior (1984) was used to calcu- late the transient particle temperature, partial pressure of CO at the particle surface and the combustion rate. The equilibrium constant K was calculated at the average particle temperature T, with thermodynamic data from the JANAF Tables (Dow Chemical Company, 1971) and Stern and Weise (U.S. National Bureau of Standards, 1966). The calcium mass fraction Yp, was then calculated from (3-45). The rate of calcium vaporization was integrated with respect to time from ignition to 95% burnoff to obtain the total amount vaporized during combustion. The calculated calcium vaporization is summarized in Fig. 3.21 by plotting the cumulative percent of calcium vaporized as a function of oxygen during combustion of a coal containing 2.5% calcium. The estimated calcium vaporiza- tion is in good agreement with observed data reported by Quann (1982), Fig. 3.22, during combustion of a lignite. The extent of vaporization increases with oxygen content since higher particle temperatures at high oxygen concentra- 91 CALCIUM VAPORIZATION 6 ] CALCIUM VAPORIZED vA a 0 l l l 0 10 20 30 40 Z OXYGEN Fig. 3.21 Calculated calcium vaporization during combustion of a coal with 2.5% Ca at a furnace temperature of 1600°K. 90 92 DATA FROM QUANN (1982) 10¢ fc T T fF 9 - MONTANA LIGNITE (1.44% Ca) : F 6AM = 6TLLLINOIS BITUMINOUS (0.8% Ca) 4 - o “ oad sz =| SO 10h be am NOE J a L J =) _ ° 4 < L > 4 | = L 4 > S A <x oO ne 10S oO 7 0 2Q 40 60 80 106 % OXYGEN Fig. 3.22 The effect of oxygen on vaporization of calcium during combustion of pulverized coals at a furnace temperature of 1750°K (data of Quann). 93 tions accelerate the vaporization. However, the total amount of calcium vapor- ized under experimental conditions examined in this study is in all cases less than 5%. This analysis shows that the effect of calcium vaporization on sulfur capture is relatively ‘small. 3.4.3 Effect of Calcium Content The three coal samples used for this study were prepared with varying amounts of calcium to examine the effect of Ca/S ratio on sulfur retention and the efficiency of the ion-exchanged calcium in capturing sulfur. The results for combustion experiments conducted with 40% Og are compared in Fig. 3.23. The extent of sulfur capture is observed to increase approximately proportional to the increase in the Ca/S ratio of the coal sample, although there does not appear to be any noticeable difference due to coal types other than the effect of Ca/S content of the coal. This same trend is also apparent in the experiments conducted with 31% Og as shown in Fig. 3.24. Earlier discussions have indicated that the capture of SOz2 by CaO in coal ash is chemical reaction-rate limited during combustion of 50 wm coal particles. Thus theoretically, the extent of sulfur capture should be independent of cal- cium content in the coal if there is a sufficient excess of calcium. The observed dependence of sulfur capture on calcium content suggests that the efficiency of calcium is rather low, possibly due to the inaccessibility or deactivation of part of the calcium content. Huffman et al. (1981) reported that calcium in coal acts as a fluxing agent in accelerating the melting of ash under reducing conditions at 1200°K to 1400°K and under oxidizing conditions at 1500°K to 1700°K. The particle temperatures achieved in this study far exceeded the melting temperature of calcium- enriched ash, and the glassy, spherical appearance of the ash particles when 94 Combustion at 160CSK in 40% O02 120 — T 1 T 0 - PSOC 665 - Ca/S=3.4 x - PSOC 623A - Ca/S=2.5 4S - PSOC 6238 - Ca/S=0.8 196F - i) XS scr 4 lu oc 2 — on ~| x 60 x (oe; jaa a) Li t = ~ 5 49 oO x 3 RESIOENCE TIME (S) Fig. 3.23 Effect of calcium content on sulfur retention for combustion of coal in 40% 05. 95 Combustion at 160C°K in 3i% O2 T T l T 3 - PSOC 685 - Ca/S=3.4 x - PSCC 623A - Ca/S=2.5 & - PSOC 6238 - Ca/S=0.8 x x a G z | | l i 120 196 = BoP Li od > oe & 607 oO cc - ue = 40 wn 20 0 0 Fig. 3.24 RESIDENCE TIME (S) Effect of calcium content on sulfur retention for combustion of coal in 31% 05. 96 examined under a microscope indicate that a considerable extent of ash melting occurs, thereby consuming part of the calcium. Calcium oxide has also been found to react with alumina and silica to form aluminate and silicate at approx- imately 1500°K (Cobb, 1910). In the presence of both alumina and silica, cal- cium aluminosilicates are formed. Since the major components of coal ash are alumina and silica, thus a fraction of the calcium is lost due to the formation of aluminosilicates and becomes unavailable for sulfur capture (Raask, 1982). Baker and Attar (1981) and Newman (1941) further found that CaSO, decom- poses faster when mixed with silica and aluminosilicates, thus suggesting an even lower efficiency of sulfur capture by the calcium in coal. The possible loss of calcium due to the formation of aluminosilicates is consistent with the observed dependence of sulfur capture on the calcium content of coal samples. However, calcium deactivation cannot be quantified without knowledge of the extent of formation of calcium aluminosilicates. 3.4.4 Effect of Furnace Temperature The effect of furnace temperature on sulfur capture during combustion of PSOC 680 (Ca/S = 3.4) is shown in Fig. 3.25. At the most favorable conditions for sulfur capture, 0.5 l/min gas fiow rate and 40% Oo, there is virtually no difference in the results for experiments conducted at furnace temperatures 1400°K and 1600°K. This is because at sulfur capture levels of above 90%, any actual difference would be relatively small. However, as gas flowrate increases, the extent of sulfur capture for experiments at 1400°K increases rapidly com- pared to experiments conducted at 1600°K. This observed trend is most prob- ably due to the combined effects of longer reaction times and lower particle temperatures at the lower furnace temperature. The reaction time was defined earlier as the residence time of particles SULFUR CAPTURE (%) 97 120 T ’ 1 r 1400°K fe) X—-——-=- 1600°K 100 60 + 20} FLOWRATE (2/MIN) Fig. 3.25 Effect of furnace temperature on sulfur retention during combustion of PSOC 680 (Ca/S = 3.4). 98 through the furnace zone where the temperatures favor capture of SO2 by CaO. Since the furnace temperature of 1400°K is approximately the temperature at which the formation of CaSQ, becomes thermodynamically favorable, the reac- tion time for sulfur capture with a furnace temperature of 1400°K is then the residence time of the particle in the furnace from the location where the parti- cle temperature approaches that of the furnace to the point where dilution air is introduced. The results of the experiments conducted at a furnace tempera- ture of 1400°K showing the effects of the redefined particle residence time are presented in Fig. 3.26. Analysis of the data according to first order kinetics with respect to S02 gives first order reaction rate constants of 0.067 sec™ at 40% Og and 0.016 sec™! at 21% Og. The observed rate constants are smaller than those obtained at a furnace temperature of 1460°K due to the difference in the temperature-time history of the particles, particularly since the reaction rate is time averaged over the entire temperature range. The maximum reaction rate of the sulfation of calcined limestone was reported to be approximately at 1220°K (Borgwardt, 1970; Borgwardt and Harvey, 1982). The reaction rate decreases as temperatures diverge from i1220°K. Thus, increasing the overall reaction time by increasing the particle residence time at 1400°K would effectively decrease the average reaction rate. The particle temperatures attained during pulverized combustion at 1400°K are much lower than at 1600°K, as illustrated by the calculated temperatures in Fig. 3.27. The lower particle temperatures and the higher levels of sulfur cap- ture at low reaction times for experiments conducted at 1400°K suggest possible effects from other mechanisms of sulfur capture than just the recapture of S02 by CaO in the ash. Under reducing conditions during the early stages of combustion, CaS is formed by the reaction of CaO with gaseous sulfur species (Freund and Lyon, 1982) as indicated by reactions (3-10) to (3-12). The sulfide is 99 PSOC 680 - Ca/S=3.4 - 1406°K 120 cr T T T o - 40% 0Q2 & - 21% Oo 100Fr = ee erro © S sok ~ Lid ow > — oS 60F - ro) a > Le _ = — > 40 7) 20Fr 4 0 | i I ! 0 4 8 12 16 20 RESIDENCE TIME (S) Fig. 3.26 Effect of residence time on sulfur retention during combustion of PSOC 680 (Ca/S = 3.4) at a furnace temperature of 1400°K. (°K) PEAK TEMPERATURE 2600 2400 2200 2000 1800 1600 1400 0 Fig. 3.27 Theoretical calculation of particle 100 CALCULATED PARTICLE TEMPERATURES o - 1600°K 4 “- 1400°K QO A | l 10 20 30 40 vA OXYGEN temperature for the combustion of a bituminous coa? with a diameter of 50 um. 90 101 subsequently oxidized to CaSOQ,. In the case of higher furnace temperature of 1600°K, the CaSO, formed readily decomposes to CaO and SOg. In the case of the lower furnace temperature of 1400°K, the rate of CaSO, decomposition would be considerably slower at lower particle temperatures. Thus the overall sulfur cap- ture would be due to the capture of sulfur during combustion of coal and the post-combustion recapture of 502 by CaO in the ash particles. 3.5 Conclusions The retention of sulfur during the combustion of calcium-ion-exchanged coals was studied in a laminar flow furnace under strongly fuel-lean conditions. It was observed that the extent of sulfur retained in coal ash increased with increasing particle residence time and increasing oxygen concentration in the combustion gas. Sulfur capture was also observed to increase in proportion to an increase in the Ca/S content of the coal. Analysis of the results indicates that chemical reaction is the rate-limiting step in the capture of SOQg by CaO in coal ash to form CaS0,. Diffusion of SOz through the combustion gas was also found to have a possible effect on the reaction rate due to the inhomogeneity of the distribution of coal ash. The findings of this study suggest that in the reducing environment generated during early stages of combustion, a fraction of the sulfur in coal may react with CaO to form CaS, As the particle environment becomes oxidizing, CaS is oxidized to CaSO, which then readily decomposes to release SOg. However, if the furnace temperature is sufficiently low, e.g. 1400°K or lower, then the rate of decompostion is slower, thus part of the sulfur is never released into the combustion gas. At the high temperatures attained during combustion, CaO or CaSO, can react with alumina and silica present in the ash, therefore removing a portion of the calcium from further reactions and decreasing the efficiency of 102 the original calcium conte..t from capturing sulfur. The major fraction of sulfur retention is derived from the recombination of SOzg and Oz with CaO in the lower region of the furnace where temperatures are favorable for the stability of CaSQ,. 103 References Baker, D. C. and Attar, A, Env. Sci. Tech. 15:2BB-293 (1981). Borgwardt, R. H., Env. Sci. Tech. 4:59-63 (1970). Borgwardt, R. H. and Harvey, R. D., Env. Sci. Tech. 6:351-360 (1972). Campbell, F. R., Hills, A. W. D. and Paulin, A., Chem. Eng. Sci. 25:929 (1970). Case, P. L, Heap, M. P., McKinnon, C. N., Pershing, D. W. and Payne, R., Prepr. ACS Fuel Chemistry Division 27:158-166 (1982). Cobb, J. W., J. Soc. Chem. Ind. 29:250-259 (1910). Crank, J., "Mathematics of Diffusion’, Oxford University Press (1956). Dow Chemical Company, Thermal Research Laboratory, JANAF Tables, 2nd ed., National Bureau of Standards, Washington, D.C. Freund, H. and Lyon, R. K., Comb. Flame 45:191-203 (1982). Hartman, M. and Coughlin, R. W., AIChE J. 22:490-498 (1976). Hartman, M. and Trnka, D., Chem. Eng. Sci. 35:11B89-1194 (1980). Hirschfelder, J. O., Curtiss, C. F. and Bird, R. B., "Molecular Theory of Gases and Liquids”, Wiley, New York (1954). Huffman, G. P., Huggins, F. E. and Dunmyre, G. R., Fuel 60:585-597 (1981). Libby, P. A. and Blake, T. R., Comb. Flame 36:139-169 (1979). McClellan, G. H., Hunter, S. R. and Scheib, R. M., Spec. Tech. Publ., No. 472, p. 32, ASTM (1970). Mulcahy, M. F. R. and Smith, I. W., Rev. Pure Appl. Chem. 19:81-107 (1969). Newman, E. 5., Journal of Research of the NBS 27:9-196 (1941). 104 Pigford, R. L. and Sliger, G., Ind. Eng. Chem. Proc. Des. Dev. 12:85-91 (1973). Quann, R. J. and Sarofim, A. F., 19th Symp. (Int.) Combustion Institute, Pittsburg, pp. 1429-1440 (1982). Quann, R. J., "Ash Vaporization under Simulated Pulverized Coal Combustion ° Conditions", PhD Thesis, Massachusetts Institute of Technology (1982). Raask, E., Prog. Energy Comb. Sci. 8:26-276 (1982). Schafer, H.N.S., Fuel 49:197-213 (1970). Schafer, H. N.S., Fuel 49:271-280 (1970). Senior, C, L. and Flagan, R. C., Aerosol Sci. Tech. :371-383 (1982). Senior, C. L, ‘Submicron Aerosol Formation during Combustion of Pulverized Coal’, PhD Thesis, California Institute of Technology (1984). Stern, K. H. and Weise, E. L., "High Temperature Properties and Decomposition of Inorganic Salts", NSRDS-NBS 7, U. S. National Bureau of Standards, Washing- ton, D.C. (1968). 105 CHAPTER 4 CONCLUSIONS AND RECOMMENDATIONS 4.1 Conclusions 1. Lignite coal possesses significant ion-exchange capacity which can be used to introduce calcium ions into the coal matrix in a molecular scale, uniform distribution. The ion-exchange capability is due to the presence of carboxyl and phenolic hydroxyl groups in the coal. For exchange of the lignite, coal can achieve Ca/S ratios greater than 7. Bituminous coal contains a consid- erably lower ion-exchange capacity than lignite coal. The concentration of available carboxyl and phenolic groups was increased by oxidation in air at 195°C and 260°C. Ion exchange of the oxidized bituminous coal could result in Ca/S contents greater than 2.5. 2. The initial rate of ion exchange increased as the calcium-ion concentration or temperature of the reaction was raised. This higher initial rate could be due to faster reaction rate and higher driving force for diffusion of ions across a liquid film surrounding coal particles. 3. The rate of ion exchange increased appreciably as the pH value of the exchange solution was increased. The increase in rate at higher pH values is due mainly to the higher ion-exchange capacity as more acidic groups readily undergo cation exchange. 4. Particle diffusion was found to be the rate-determining step for the process of ion exchange. Analysis of the results with a simple model for particle diffusion indicated that the diffusion coefficient of calcium ions was approxi- mately 1.6 x 10-8 cm?/sec for diffusion in PSOC 623. The diffusion coefficient in PSOC 680 was estimated at 5.0 x 10°" to 3.5 x 107!em?/sec. This difference could probably be due to the different electric potentials in the 106 pore structures generated by ionic interactions. 5. Emission of SQg during fuel-lean combustion of calcium-ion-exchanged coals was observed to decrease with increasing particle residence time. Analysis of the results suggested that SO. was released during combustion of the coal samples and was captured by CaO in the cooler region of the furnace where thermodynamics together with kinetics favor the formation of CaS0,. 6. Chemical reaction was found to be the rate-determining step in the capture of SQz by CaO in coal ash. Diffusion of SO2 from the gas phase to the ash particles could also affect the reaction rate due to the inhomogeneity of the radial dispersion of coal ash in the combustion furnace. 7, Sulfur retention was observed to increase as oxygen concentration in the combustion gas was increased. This trend may be partially due to vaporiza- tion of calcium at the high particle temperatures attained during combus- tion in higher oxygen concentrations, which would then result in CaO aerosol particles of high surface area that would be effective in capturing 502. 8B. The efficiency of sulfur capture by the exchanged calcium was found to be lower than theoretically possible. This result could be due to a loss of cal- cium from the formation of aluminosilicates and thus becoming unavailable for sulfur capture. 4.2 Recommendations 1. The present study examined the kinetics of ion exchange of carboxylic-acid groups in coal. A further study examining the kinetics of exchange of all acid groups in coal would be informative. 2. The diffusivity of ions in coal were found to be several orders of magnitude slower than in an electrolyte solution.-A study of the surface area and pore 107 size distribution of the coal as a function of extent of ion exchange can be useful in the understanding of the diffusion process. . The efficiency of sulfur capture by exchanged calcium was affected by the presence of alumina and silica in the coal. The interference can be elim- inated by preparing synthetic coals in which ion-exchanged calcium is the only mineral present. The particle temperature and vaporization of calcium should also be measured to fully understand the role and efficiency of cal- cium in the retention of sulfur. . A detailed surface area study of the coal and ash from combustion should be conducted in order to examine the effect of particle surface area on the efficiency of sulfur capture. The temperature profile of the reaction zone should also be incorporated into the analysis of the results so that kinetic parameters may be obtained. 108 APPENDIX FURNACE CALIBRATION The overall SOz ‘balance in the combustion furnace was determined by a series of calibration runs with a gas of known 502 concentration under experi- mental conditions. At each flowrate, the SOg concentration in the gas was set by a mixture of air and 1005 PPM SOQ, in air to simulate the SOzg concentration in the furnace during combustion of coal. At the start of a run, the flow of air was introduced into the furnace through the inlet for main combustion gas, while the flow of 1005 PPM SQz2 was fed through the coal injector. The exit gas mixture was scrubbed through two gas-washing bottles containing a solution of 1% H2gOxz. The total amount of 502 collected in a four-hour run was analyzed by titration of the HgOe solution. The first calibration run was conducted with the furnace at room tempera- ture. The overall SQzg balance obtained was 99.7%, which suggested that there were no leaks in the system. The results for the calibration runs conducted at experimental temperatures are as follows: Flowrate A502 (l/min) Balance 4.1 92 2.1 B9 1.0 8B 0.5 B4 The results of the calibration runs show that there is a loss of 50, through the furnace at high temperatures. Since the system is closed and has no leaks, the 109 losses can only be due to the lower part of the furnace where the alumina- insulated sample collector is located. If the recombination of 502 with CaO to form CaSO, occurs in the cooler region of the furnace and the reaction is negli- gible as the combustion gases are cooled and diluted with air entering the sam- ple probe, then the actual sulfur emission can be calculated as (A-1) where * = corrected sulfur emission. S = measured sulfur emission. y = correction factor as determined by calibration experiments. Tables A.1 to A.4 summarize the results of the combustion experiments taking into consideration the losses of 502 in the furnace. In several cases, the correc- tion results in sulfur balances exceeding unity, which could be due to several factors: experimental error, inhomogeneity of coal samples and a simplification of the correction. Duplicate runs conducted at most cases showed the results to be reproducible to approximately 52%. 110 Table A.1 Corrected data for retention of sulfur during combustion of PSOC 623A (Ca/S = 2.5) at a furnace temperature of 1600°K 20 Flowrate % S emission % S emission ° *2 (1/min) (measured) (corrected) % S capture 40 4.1 76.6 ‘83.3 16.7 2.1 77.0 86.5 13.5 1.0 57.6 65.5 34.5 0.5 34.0 40.5 59.5 30.6 4.1 82.7 89.9 10.1 2.1 77.0 86.5 13.5 1.0 58.4 66.4 33.6 0.5 42.4 50.5 49.5 21 4.1 94.7 102.9 0 2.1 86.0 96.6 3.4 1.0 74.7 84.9 15.1 0.5 62.3 74.2 25.8 13.1 4.1 93.8 102.0 0 2.1 83.4 93.7 6.3 1.0 71.8 81.6 18.4 0.5 60.6 72.1 27.9 Table A.2 Corrected data for retention of sulfur during combustion of PSOC 623B (Ca/S = 0.8) at a furnace temperature of 1600°K % 0 Flowrate % S emission % S emission ° <2 (1/min) (measured) (corrected) % S capture 40 4.1 96.0 104.3 0 2.1 93.5 105.1 0 1.0 77.8 88.4 11.6 0.5 61.5 73.2 26.8 30.6 4.1 99.3 107.9 0 2.1 88.5 99.4 0.6 1.0 79.2 90.0 10.0 0.5 65.6 78.1 21.9 21 4.1 117.5 127.7 0 2.1 92.3 103.7 0 1.0 82.2 93.4 6.6 0.5 73.0 86.9 13.1 lll Table A.3 Corrected data for retention of sulfur during combustion of PSOC 680 (Ca/S = 3.4) at a furnace temperature of 1600°K 20 Flowrate % S emission % S emission o €2 (1/min) (measured) (corrected) % S capture 40 4.1 95.0 103.3 0 2.1 71.1 79.9 20.1 1.0 35.6 ~ 40.5 59.5 0.5 11.8 14.0 86.0 30.6 4.1 91.9 99.9 0.1 2.1 82.7 92.9 7.1 1.0 51.3 58.3 41.7 0.5 31.3 37.3 62.7 21 4.1 97.0 105.4 0 2.1 98.6 110.8 0 1.0 80.1 91.0 9.0 0.5 71.7 85.4 14.6 13.1 4.1 94.1 102.3 0 2.1 94.1 105.7 0 1.0 79.1 89.9 10.1 0.5 71.6 85.2 14.8 Table A.4 Corrected data for retention of sulfur during combustion of PSOC 680 (Ca/S = 3.4) at a furnace temperature of 1400°K % 0 Flowrate % S emission % S emission 2 2 (1/min) (measured) (corrected) % S capture 60 4.1 15.9 17.3 82.7 0.5 3.7 4.4 95.6 40 4.1 42.8 46.5 53.5 2.1 36.5 44.0 59.0 1.0 20.0 22.7 77.3 0.5 14.6 17.4 82.6 21 4.1 66.0 71.7 28.3 2.1 64.3 72.2 27.8 1.0 52.2 59.3 40.7 0.5 48.1 57.3 42.7 30.7 0.5 43.6 51.9 48.1