Behavior of Chloride during Coal Combustion in an AFBC System

Transcription

1 Energy & Fuels 1999, 13, Behavior of Chloride during Coal Combustion in an AFBC System Wei Xie, Wei-Ping Pan,* and John T. Riley Materials Characterization Center, Department of Chemistry, Western Kentucky University, Bowling Green, Kentucky Received July 15, 1998 Two 1000-h burns were conducted with the 12-in. (0.3 m) laboratory AFBC system at Western Kentucky University. Operating conditions similar to those used at the 160-MW AFBC system at the TVA Shawnee Steam Plant located near Paducah, KY, were used. A low-chlorine (0.012% Cl and 3.0% S) western Kentucky No. 9 coal and a high-chlorine (0.28% Cl and 2.4% S) Illinois No. 6 coal were used in this study. Four different metal alloys [carbon steel C1020 (0.18% C and 0.05% Cr), 304 SS (18.39% Cr and 8.11% Ni), 309 SS (23.28% Cr and 13.41% Ni), and 347 SS (18.03% Cr and 9.79% Ni)] were exposed uncooled in the freeboard area at the entrance to the convective pass, where the metal temperature was approximately 900 K. A two-phase investigation was carried out in order to study the fate of chlorine during coal combustion in an AFBC system and to study the susceptibility of boiler components to corrode in combustion gases containing hydrogen chloride. As determined from emission and ash studies, the temperature in an AFBC system plays a key role in the retention of chloride, which is more favorable at low operating temperatures. A small amount of scale failure was observed on the other three samples in both test runs. On the basis of the SEM-EDS mapping results, there was no localized chloride distribution observed on the surface of the coupons, either in the scale failure area nor on the rest of the metal part. Some trace amount of chloride was found but was evenly distributed on the surface of the coupons. There was no concentration of chloride on the spot of scale failure. The scale failure might be due to sulfur attack and/or the effect of erosion. Further study with higher chlorine-content coals for more conclusive information is needed. Introduction The occurrence of furnace wall and superheater corrosion in fluidized-bed combustor systems has caused some operational and economic concerns. 1 It is generally accepted that chlorine may play a role in this corrosion. To predict the performance of high-chlorine coals in these combustion systems, it is necessary to have a better understanding of the different corrosion mechanisms in which chlorine and sulfur may be involved. 2 It is also important to evaluate the critical point of coal chlorine content which may cause initial corrosion. To better understand the combustion behavior of chloride during coal combustion in an AFBC system, a comprehensive research project was performed at Western Kentucky University. The project concentrated on the effect of coal chloride content on HCl emission reduction, on the effect of chloride on the absorption of SO x emissions, on the effect of operating parameters on the distribution of chloride in the fly ash and bed ash, (1) Minchener, A. J.; Lloyd, D. M.; Stringer, J. The Effect of Process Variables on High-Temperature Corrosion in Coal-Fired Fluidized Bed Combustors. In Corrosion Resistant Materials for Coal Conversion Systems; Meadowcroft, D. B., Manning, M. I., Eds.; Applied Science Publishers: New York, 1983; Chapter 15, p 299. (2) Mayer, P.; Manolescu, A. V.; Thorpe, S. J. Influence of Hydrogen Chloride on Corrosion and Corrosion-Enhanced Cracking Susceptibility of Boiler Construction Steels in Synthetic Flue Gas at Elevated Temperature. In Corrosion Resistant Materials for Coal Conversion Systems; Meadowcroft, D. B., Manning, M. I., Eds.; Applied Science Publishers: New York, 1983; Chapter 5, p 87. and on the effect of high-temperature corrosion for different metals in fluidized beds during coal combustion. The laboratory-sized atmospheric fluidized-bed combustor (AFBC) at Western Kentucky University (WKU) was designed to serve as a flexible research and development facility to evaluate combustion performance and to estimate the effects of flue gas emissions. The WKU AFBC system was configured to simulate the 160 MW AFBC system at TVA s Shawnee Steam Plant near Paducah, KY, and the operating conditions for the WKU system are similar to those used in the Shawnee Plant s system. The chlorine content of coal varies from just a few parts per million to thousands of parts per million. Emissions of chloride from coal-fired plants can range from 50 to several thousand parts per million, depending on the original concentration in the coal, the type of combustor, and any pollution control equipment installed. It has been estimated that 94% of the chloride in coal is volatilized during combustion, generally being emitted as gaseous HCl. 3 In an AFBC system, limestone may be able to capture the chlorine. Limestone degenerates to CaO, and the CaO reacts with HCl to produce CaCl 2. 4 In an AFBC system, capture of chloride by limestone in the combustion zone depends on the (3) Shao, T. M.S. Thesis, Western Kentucky University, (4) Heidbrink, J. M.S. Thesis, Western Kentucky University, /ef CCC: $ American Chemical Society Published on Web 02/26/1999

2 586 Energy & Fuels, Vol. 13, No. 3, 1999 Xie et al. temperature of the zone and the ratio of calcium-tosulfur. A study by Liang and others 5 showed that chloride capture has a large variation with temperature moving from a low of 18% gaseous HCl at 700 C to 99% HCl at 950 C. The resulting product is almost entirely in the form of liquid CaCl 2. Munzner and Schilling 6 studied the effect of limestone in a bench-scale AFBC system and showed that a greater recapture of chloride occurred with larger excesses of limestone or when the Ca/S ratio was greater than 2. The combustion of coal in a fluidized-bed combustor with a limestone bed is one method of controlling sulfur emissions. With increasing use of high-chloride coal for combustion in an AFBC system, the addition of limestone may help reduce chloride emission. Therefore, in situ sulfur and halogen capture by limestone is the major advantage of fluidized-bed combustion. It is possible that chlorine might also affect the detailed chemistry of the processes of sulfur capture in fluidizedbed combustors 7,8 and also might play a role in sulfur deposition in ashes. One of the main goals for this project is to evaluate the critical point where the coal chloride content may cause initial corrosion. The chloride content of most coals burned in U.S. coal-fired boilers has recently been limited to less than about 0.3% by weight to avoid aggravation of the potential for fireside corrosion. However, no formal report has been published to support this limit, although there have been several studies conducted on British coals which indicate that aggressive fireside corrosion can be correlated to the use of high-chlorine coal. (5) Liang, D. T.; Anthony, E. J.; Leowen, B. K.; Yates, D. J. Proceedings, 11th International Conference on FBC, Montreal, Canada, April 21-24, 1991; Vol. 2, pp (6) Munzner, H.; Schilling, D. H. Proceedings, 8th International Conference on FBC, Houston, TX, March 18-21, 1985; Vol. III, pp (7) Johnson, I.; Lence, J. F.; Shearer, J. A.; Smith, G. W.; Swift, W. M.; Treats, F. G.; Turner, C. B.; Jonke, A. A. Support Studies in Fluidized-Bed Combustion; Argonne National Laboratory Quarterly Report ANL/CEN/FE-79-8, (8) Griffin, R. D. A New Theory of Dioxin Formation in Municipal Solid Waste Combustion. Chemosphere 1986, 15. (9) Orndorff, W. W.; Su, S.; Napier, J.; Bowles, J.; Li, H.; Li, D.; Smith, J.; Pan, W.-P.; Riley, J. T. Studies with a Laboratory Atmospheric Fluidized Bed Combustor. Proceedings, 12th International Coal Testing Conference, Cincinnati, OH, 1996; pp (10) Pan, W.-P.; Riley, J. T. Behavior of Chlorine during Coal Combustion in an AFBC System; Final Report to Electric Power Research Institute (W ), Figure 1. Schematic diagram of Western Kentucky University s AFBC system. Table 1. Proximate and Ultimate Analysis Values a for the Coals and Limestones Used in the Study KY limestone Proximate Analysis (%) moisture ash Volatile Matter Fixed Carbon Ultimate Analysis (%) Ash Carbon Hydrogen Nitrogen Sulfur Oxygen Chlorine (ppm) BTU/lb n/a a Moisture is as-received; all other values are reported on a dry basis. Experimental Section Two 1000-h burns were conducted with the 0.3-m (12-in.) laboratory AFBC system at Western Kentucky University. A schematic diagram of the AFBC system as used for the tests and the full description of the AFBC system (Figure 1) have been previously presented by Pan and co-workers. 9,10 The combustor s operating parameters (air/water flow, coal/lime feed, bunker weight, temperatures, and pressure) are controlled and logged to file with a Zenith 150 MHz computer utilizing the LABTECH software version 3.0. During combustion runs, any needed changes in the parameters can easily be entered into the computer by accessing the correct control screen and making the necessary corrections on line. A 1000-h burn was done with a low-chlorine (0.012% Cl and 3.0% S) Western Kentucky No. 9 coal (WKU 95011), which is the same type of coal as that supplied to the TVA plant in A second 1000-h burn was conducted with a highchlorine (0.28% Cl and 2.4% S) Illinois No. 6 coal (WKU 95031). Analytical values for the coals, as well as the limestone from Kentucky Stone in Princeton, KY, are given in Table 1. The coal and limestone both were air-dried before being crushed to -4 mesh (4.75 mm). The limestone was also used as the bed material in the AFBC system. During combustion runs, limestone was fed into the system at a constant rate. The summary of steady-state run conditions for both test runs are shown in Table 2. Four different metal alloys (carbon steel C1020, 304 SS, 309 SS, and 347 SS) were studied in this project. Elemental compositions of these alloys are given in Table 3. Each metal coupon had a 5.08 cm outside diameter and was cm thick with a cm diameter hole in the center. A set of metal coupons was placed at 3.35 m above the fuel injection port, which is 10 cm below the convective pass heat exchange tubes. The coupons were held in place by a machinable tungsten rod (powder metallurgically prepared) and separated by ceramic mounts. Two sets of each coupon (eight total) were used in each run. Each specimen within a group was rotated as to position in an array every 250 h during the test burn. The temperature near the metal coupons was approximately

3 Behavior of Chloride During Coal Combustion in an AFBC System Energy & Fuels, Vol. 13, No. 3, Table 2. Summary of Steady-State Run Conditions primary zone gas velocity (ms -1 ) 0.33 (298 K) 1.24 (1120 K) fuel feed rate (kg h -1 ) (for coal 95031) (for coal 95011) limestone feed rate (kg h -1 ) (for coal 95031) (for coal 95011) Ca/S molar ratio Temperature (K) bed temp in 0.56 m in 0.97 m in 1.90 m (probe 1) in 2.60 m (probe 2) in 3.30 m (probe 3) Table 3. Analytical Composition (Percent by Weight) of Four Alloy Metals element C1020 C304 C309 C347 C Cr Cu Mn Mo Ni P S Si Fe balance balance balance balance others N-0.06 N N CO-0.03 CO K. The coupons were weighed before and after the run and examined using SEM-EDS. The procedure for cleaning the specimens was to blow deposits off using compressed nitrogen until a constant weight was reached. The SEM analysis was performed using a JEOL JSM-5400 SEM. Attached to the SEM for energy-dispersive X-ray analysis (EDS) was a KEVEX Sigma 1 system with a Quantum detector for elemental analysis down to carbon. The following instrument operating parameters were used for the SEM/EDS analysis: electron beam energy, 20 kev; working distance, 24 mm; sample tilt angle, 0. A total of eight spots (35X) on each specimen were analyzed. The concentration of S, O, Cr, Cl, Ca, Mn, Ni, Fe, Na, Mg, Al, Si, and K were determined. The highest and lowest sulfur content spots are presented and discussed in the next section. Small pieces (2 cm 0.3 cm) were cut from specimens using a LECO VC-50 precision diamond saw. The small pieces were mounted using a LECO PR-32 mounting press (pressure: 4 psi, 20 min heating, and 40 min cooling, Lucite powder was used). The mounted specimens were ground and polished using a LECO AP-60 8-in. Grinder/Polisher, using three different SiC grit sizess400 PSA (600 s), 600 PSA (300 s), and 800 PSA (300 s)sat 300 rpm and 70 lbs of pressure. The specimens were polished with 1-µm diamond compound/red wool felt cloth/oil microid diamond compound extender (120 s, 300 rpm and 65 lbs). The mounted specimens (cross section) were examined using SEM/EDS. The results obtained by EDS analysis were corrected by computer for corrections such as adsorption, fluorescence, and atomic weight. Results and Discussion A two-phase investigation was carried out in order to study the fate of chlorine during coal combustion in an AFBC system and to study the susceptibility of boiler components to corrode in combustion gases containing hydrogen chloride. Effects of Temperature. The effect of temperature on the emission of chloride is illustrated in Figure 2. Figure 2. Effect of temperature on the emission of hydrogen chloride at different heights above the fuel injection port. Figure 3. Effect of bed temperature on the concentration of sulfur in the fly ash of the AFBC system. Figure 4. Effect of bed temperature on the chloride content in the bed ash of the AFBC system. More HCl was observed when the temperature was raised. The capture of HCl by limestone is more difficult than the capture of SO Also, the reaction between HCl and CaO is more favorable at lower temperatures. 5 The increased concentration of HCl with increasing distance above the fuel injection ports might result from secondary combustion (the secondary air was injected 0.8 m above the setter plate) occurring in the freeboard region. The effects of bed temperature on the chlorine contents in fly ash and bed ash are shown in Figures 3 and 4. It is obvious that the chlorine content in the source coal is a decisive factor in the distribution of chloride

4 588 Energy & Fuels, Vol. 13, No. 3, 1999 Xie et al. Figure 5. Effect of coal type on the emission of sulfur dioxide at different temperatures in the AFBC system. in the ash. The low-chlorine coal (95011) released less hydrogen chloride during combustion. As a result, there were almost no temperature effects on the absorption of HCl in both the fly ash and bed ash from the combustion of this coal. For the high-chlorine coal (95031), both Figures 3 and 4 show that chloride retention is more favorable at low temperatures. 11,12 Effects of Coal Chlorine Content on the Retention of SO 2. Figure 5 shows the results of sulfur dioxide emission from tests with two different coals (95011 and 95031). It is obvious that efficient sulfur capture is obtained between 1120 and 1170 K for both coals. The lower sulfur oxide emission for coal observed at the higher temperature may be due to the effect of high chlorine content. 13 It may also be due to the interaction between SO 2 and HCl in the oxygen-rich conditions. 14 Several reactions may be involved. In the oxygen-rich conditions and high temperature, HCl can react with oxygen to form Cl 2, a reaction known as the Deacon Reaction. Figure 6. Effect of coal type on the emission of hydrogen chloride at different temperatures in the AFBC system. Figure 7. Effect of the Ca/S ratio on the sulfur content in the fly ash. 4HCl + O 2 f 2Cl 2 + 2H 2 O (1) When SO 2 is presented via coal sulfur combustion, an interesting and important reaction is one in which SO 2 may be attacked by Cl 2 to form SO 3 and HCl: Cl 2 + SO 2 + H 2 O f 2HCl + SO 3 (2) Therefore, more SO 2 can be converted to SO 3 and then be absorbed by CaO. SO 2 absorption may also be due to the formation of voids, allowing the diffusion of HCl and SO 2 toward the inside of limestone particles. 15 Effect of Coal Type and Ca/S Ratio. There is good agreement between HCl emission and the chlorine (11) Julien, S.; Brereton, C. M. H.; Lim, C. J.; Grace, J. R.; Anthony, E. J. Fuel 1996, 75 (4), (12) Bramer, E. A. Flue Gas Emission from Fluidized Bed Combustion. In Atmospheric Fluidized Bed Coal Combustion; Elsevier: The Netherlands, (13) Johnson, I.; Lence, J. F.; Shearer, J. A.; Smith, G. W.; Swift, W. M.; Treats, F. G.; Turner, C. B.; Jonke, A. A. Support Studies in Fluidized-Bed Combustion; Argonne National Laboratory Quarterly Report ANL/CEN/FE-79-8, (14) Xie, W.; Han, W.; Pan, W.-P.; Riley, J. T. The Effect of Halides on Emissions from Atmospheric Fluidized Bed Combustion. Presented at the 83th Kentucky Academy of Science Meeting, Morehead, Kentucky, (15) Matsukata, M.; Takeda, K.; Miyatani, T.; Ueyama, K. Simultaneous Chlorination and Sulphation of Calcined Limestone. AIChE J. 1996, 51 (11), Figure 8. Effect of the Ca/S ratio on the chloride content in the fly ash. content in the coals, as illustrated in Figure 6. In other words, the coal with higher chlorine content showed greater emissions of HCl. Figures 7 and 8 show the effect of the Ca/S ratio on the sulfur and chloride retention in ash. One can see from these figures that the Ca/S ratio has more influence on the sulfur retention for the high-sulfur-content coal (95011). With an increase in the Ca/S ratio, the sulfur content in fly ash increased. Also, it can be seen that there is little effect of the Ca/S ratio on the chlorine content in fly ash. On the other hand, Figure 8 shows that the Ca/S ratio is more important for the capture of chloride compared to sulfur for the high-chlorine-content coal (95031). It is assumed that the HCl is probably

5 Behavior of Chloride During Coal Combustion in an AFBC System Energy & Fuels, Vol. 13, No. 3, Figure 9. Effect of time of exposure to AFBC combustion gases from low-chlorine coal on the weight of metal coupons. Figure 10. XRD data for alloy 304 before and after two test runs for coals and captured in the low bed temperature region when the flue gas is passing through the heat exchange tube region because the reaction between HCl and CaO is more favorable at the lower temperature. 11,12 The composition of the fly ash (collected from the wet cyclone) in both cases is Ca(OH) 2, which is due to CaO reacting with water, CaSO 4, CaCO 3, and CaCl 2. This is the only place CaCl 2 was identified in this study. Effect of Chlorine Content on Corrosion. The C1020 specimen was cracked after 250-h of operation in both test runs. Figure 9 illustrates the weight change data for the alloys in the first 1000-h burn with the lowchlorine coal. The type 347 specimen showed the highest weight gain among the other three samples. However, alloy 304 showed the most oxide scale ( 20 µm) on the surface, followed by alloy 347 ( 12 µm), while alloy 309 ( 10 µm) showed the least oxide scale on the surface. These observations are based on the color (reddish) and SEM studies. 10 The outer scale was chiefly hematite (Fe 2 O 3, red-brown oxide, d A ) 2.70, 2.52, 1.69, 1.84), which was identified using X-ray diffraction spectroscopy, as illustrated in Figure 10. The inner layer was enriched in chromium oxide. Figure 11 shows line scans of elements (Fe, Cr, and Ni) present in the cross-section of alloy 347. Internal precipitates in the metal just below the metal-oxide interface, sometimes in alloy grain boundaries, had a globular appearance, which is typical of chromium sulfide, but EDS results showed neither Cr nor S (or anything else) to be especially concentrated in them. Also, a few spot analyses were made to determine whether any of the internal precipitates Figure 11. Scale formed on alloy 347 after 1000 h of exposure at 1144 K to combustion gases from low-chlorine coal. (a) Fe X-ray line scans, (b) Cr X-ray line scans, (c) Ni X-ray line scans. present in alloys were chromium sulfide, and the mapping results showed no sulfur enrichment with chromium. However, the results showed sulfur was somewhat enriched with calcium. Ni contained in alloys seems to provide no help with regard to corrosion prevention, since it was consumed in the outer layer.

6 590 Energy & Fuels, Vol. 13, No. 3, 1999 Xie et al. Very similar results were observed in the case of alloy specimens 304 and 347. Slight oxide spallation was observed in all three alloys. The degree of spallation for the three alloys followed the order 304 > 347 > 309 (the least). This is the reason for the fluctuation of weight change shown for alloy 304. The weight gain is due to oxidation, and the weight loss is due to oxide spallation. The two reactions compete with each other. The spallation might be due to the cooling effect (cooling of the samples from the reaction temperature) 16 or sulfur attack and the effect of erosion. 17,18 The EDS results for alloy 347 indicated that the amount of sulfur and calcium on the surface increased with burning time. These results may be due to the presence of deposits on the surface of the coupons. The concentration of oxygen on the coupons increased with time of combustion but leveled off after 500 h. The results for iron are opposite from the oxygen results. These results may indicate the formation of oxide scale followed by the spallation of oxide. There was no chloride identified on the surface of any metals that were exposed during the combustion of the low-chlorine coal. Fly ash passed through the specimens during the entire run. Thus, the effect of erosion should also be taken into consideration. 10 Alloy 309 is the best corrosion-resistant material (less oxide scale and scale spallation) among the three alloys tested under our experimental conditions. The performance of alloy 309 may be due to the high amount of chromium in the material. Figure 12 shows the weight changes for the three alloys in the second 1000-h burn with the high-chlorine coal (0.28% Cl). Weight gains were observed for alloys 309 and 347 before 500 h. In the case of the 304 alloy, the weight remained almost constant in the first stages of the test burn. There was no chloride (EDS results) observed on the surface of coupons before 500 h. A small amount of oxide spallation was observed on all three samples, which is similar to the results obtained with the low-chlorine coal in the first 1000-h test. On the basis of the EDS results, the elemental distribution on the surface presents a similar trend with the first test runs for all three alloys. However, weight loss was observed in all three coupons after 500 h. Chloride was identified on the surface of the coupons. The higher chloride (around 1 wt %) contents were observed after the first few cleaning (by blowing with N 2 ) cycles. The chloride content on the surface of the coupon decreased to about 0.1% after cleaning was complete. This may indicate that most chloride exists in the ash deposits. Ash deposits on the gas heat exchange tubes (average temperature around 770 K), which were just a few inches above the uncooled coupons, were collected at different locations (bottom, middle, top, right, and left tubes). The sulfur and chloride distributions in the deposits on the heat (16) Verma, S. K. Oxidation/Sulfidation Behavior of Ferritic 446, Austenitic 310, and Cobalt-Base Alloy 6B in High-Temperature Exposures in A Coal Gasification Atmosphere. Proceedings, Symposium on Corrosion in Fossil Fuel Systems; Wright, I. G., Ed.; The Electrochemical Society, Inc., 1983; Vol. 83-5, pp (17) Wright, I. G.; Stringer, J. Fly Ash Erosion of Boiler Convection Banks. Presented at EPRI/Florida Power and Light Seminar on Use of Coal in Oil-Design Utility Boilers, Dec 2-4, 1980, Section 5, (18) Tabakoff, W.; Kotwal, R.; Hamed, A. Erosion Study of Different Materials Affected by Coal Ash Particles. Wear 1979, 52, Figure 12. Effect of time of exposure to AFBC combustion gases from a high-chlorine coal on the weight of metal coupons. Figure 13. Chloride and sulfur concentrations at different locations on the heat exchange tubes in the AFBC combustor. exchange tubes are shown in Figure 13. Most of the chloride was deposited on the outside tubes (bottom, left and right). As for sulfur in the deposits, the same results were not observed, since sulfur was evenly distributed in the deposits on the tubes. This may indicate that the outside tubes may be more easily attacked (if any) by chloride than the inner and top tubes. On the basis of the mapping results, chloride was evenly distributed on the surface of the metal coupons. There was no significant precipitate of chloride in the outer and inner scale or the cross section of the coupons. There was no significant difference in the morphologies of cross sections of coupons exposed in the two different runs. There was no evidence to indicate chloride was the cause of any corrosion problem. However, there was more scale spallation observed in the high-chlorine coal test run than was observed in the low-chlorine coal test run. This may be due to the higher erosion rate occurring in the case of high-chlorine coal. The higher erosion rate was due to the higher ash contents (10.78% vs 9.37%) and the higher coal feeding rate (10.44 vs 8.41 Kg/h, to keep the Ca/S ratio ) 3) in the high-chlorine coal. This is evidence to suggest that sulfur and chlorine may both enhance attack on the metal coupons, but the data are not conclusive. Conclusions Considering the results of this study, some observations and conclusions that can be made are as follows. The capture of HCl by limestone is more difficult than

7 Behavior of Chloride During Coal Combustion in an AFBC System Energy & Fuels, Vol. 13, No. 3, the capture of SO 2. There is no significant change in the emission of HCl when the Ca/S ratio is varied. The bed temperature in an AFBC system plays a key role in the retention of sulfur and chloride in ash. When the bed temperature is too high, less SO 2 is absorbed in the bed ash and more is absorbed in the fly ash. The chloride content of a coal is an important factor in the retention of chloride in the ash. Chloride retention in both the fly ash and bed ash is more favorable at low operating temperatures. When the sulfur or chlorine content in coal reaches a certain point, the Ca/S ratio in the combustion mixture will be an important factor in the absorption of SO 2 and HCl. In the combustion of a highchlorine coal, more sulfur is captured in the bed ash than in the combustion of a low-chlorine coal. Of the four alloys tested, the high-chromium ( 23%) 309 alloy steel forming Cr 2 O 3 on the surface is the most resistant to corrosion of the four materials tested. Scale spallation of alloy coupons was observed in both 1000-h test burns with low- and high-chlorine coals. The second test burn with the high-chlorine coal showed more scale spallation than that obtained with the first run with the lowchlorine coal. This increase in spallation may be due to erosion by the higher ash content of the high chlorine coal. There is no conclusive evidence to indicate chlorine may cause corrosion problems under our experimental conditions. Acknowledgment. The financial support for this work received from the Electric Power Research Institute (Contract No. W ) and the technical assistance from the Illinois Clean Coal Institute is gratefully acknowledged. EF