Environ. Sci. Technol. 2002, 36, 3020-3024
A New Process for Converting SO2 to Sulfur without Generating Secondary Pollutants through Reactions Involving CaS and CaSO4 H. Y. SOHN* AND BYUNG-SU KIM Department of Metallurgical Engineering, University of Utah, 135 S 1460 E RM 412, Salt Lake City, Utah 84112-0114
Nonferrous smelters and coal gasification processes generate environmentally harmful sulfur dioxide streams, most of which are treated to produce sulfuric acid with the accompanying problems of market shortage and transportation difficulties. Some sulfur dioxide streams are scrubbed with an alkali solution or a solid substance such as limestone or dolomite, which in turn generates wastes that pose other pollution problems. While the conversion of sulfur dioxide to elemental sulfur has many environmental advantages, no processes exist that are environmentally acceptable and economically viable. A new method for converting sulfur dioxide to elemental sulfur by a cyclic process involving calcium sulfide and calcium sulfate without generating solid wastes has been developed. In this process, calcium sulfate pellets as the starting raw material are reduced by a suitable reducing agent such as hydrogen to produce calcium sulfide pellets, which are used to reduce sulfur dioxide producing elemental sulfur vapor and calcium sulfate. The latter is then reduced to regenerate calcium sulfide. Thermodynamic analysis and experimental results indicated that the CaS-SO2 reaction produces mainly sulfur vapor and solid calcium sulfate and that the gaseous product from the CaSO4-H2 reaction is mainly water vapor. The rates of the two reactions are reasonably rapid in the temperature range 1000-1100 K, and, importantly, the physical strengths and reactivities of the pellets are maintained largely unchanged up to the tenth cycle, the last cycle tested in this work. Sulfur dioxidecontaining streams from certain sources, such as the regenerator off-gas from an integrated gasification combined cycle desulfurization unit and new sulfide smelting plants, contain much higher partial pressures of SO2. In these cases, the rate of the first reaction is expected to be proportionally higher than in the test conditions reported in this paper.
I. Introduction Gas streams containing high levels of sulfur dioxide are generated from nonferrous metal smelters and integrated gasification combined cycle desulfurization units (1-4). The sulfur dioxide in relatively high strength streams is converted to sulfuric acid. However, there are often few attractive markets for the sulfuric acid, and thus its production is to fix the pollutant byproduct sulfur dioxide rather than to make * Corresponding author phone: (801)581-5491; fax: (801)581-4937; e-mail:
[email protected]. 3020
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 13, 2002
a marketable commodity (5-7). The current trend in the sulfide smelting industry is to utilize increasingly high oxygen enrichment in the process gas to reduce the heavy energy load to heat nitrogen to the process temperature and reduce the volume of the process gas (3). As a consequence, the modern sulfide smelting technologies produce high-strength sulfur dioxide gas streams. As an example, the new Kennecott copper converting process in Utah generates a steady stream of off-gas containing 35-45% sulfur dioxide (3, 8, 9). For conversion to sulfuric acid, however, this stream must be diluted to 10-14% sulfur dioxide concentration (2, 7, 10). Therefore, it would be highly desirable to develop an alternative process to treat sulfur dioxide streams. Conversion to elemental sulfur would have many advantages because it is inert as a solid and its long distance transportability is excellent. A number of processes for converting sulfur dioxide directly to elemental sulfur have been suggested and developed previously. There are two categories of technologies for this: dry (gas-phase) reduction (4, 11-13) and wet (liquid-phase) reduction (4, 14, 15). Dry reduction processes include the following: 1. The ASARCO Process (11) involves reacting a mixture of SO2, O2, and CH4 at about 713 K, the off-gas from which consists mainly of sulfur dioxide, carbonyl sulfide, and hydrogen sulfide along with sulfur vapor, carbon dioxide, water vapor, nitrogen, and minor quantities of carbon monoxide, hydrogen, and carbon disulfide. This off-gas is further treated before passing through a Claus reactor to recover elemental sulfur. 2. The Allied Chemical Process (11), like the ASARCO Process, uses natural gas as the reducing agent. About half of the sulfur dioxide is reduced directly to sulfur and is removed by condensation. The remainder is reduced principally to hydrogen sulfide and removed in the Claus plant. 3. The Boliden Process (12) and the Lurgi Process (11) use hot coke to reduce sulfur dioxide in roaster gases. All the processes described above have been discontinued because the plants are complex and expensive. 4. Under development at Lawrence Berkeley Laboratory (4) is a process that catalytically reduces sulfur dioxide at near atmospheric pressure, using reducing gases such as H2, CO, and methane. The product gas consists of elemental sulfur, water, and CO2. The disadvantage is that the product gas contains, in addition to sulfur, large amounts of water and carbon dioxide, which cause the separation of the sulfur to be difficult. 5. The Direct Sulfur Recovery Process (DSRP), under development at Research Triangle Institute (4, 13), is designed to convert sulfur dioxide into elemental sulfur at a temperature of 823 K and a pressure of about 20 atm. Sulfur-dioxidecontaining gas is reduced by a synthesis gas such as raw coal gas in the presence of a selective catalyst. However, this process is effective only at high pressures. 6. Another process for dry reduction is the Eisenlohr process (13), in which hydrogen is added to the feed stream and the gas mixture is passed through a bed of cobalt oxide, nickel oxide, molybdenum oxide, or tungsten oxide. Disadvantages of this process include the high costs of the catalysts, the possibility of heavy metal pollution, and the added requirement of the Claus plant to process the product gas mixture after elemental sulfur is recovered. Wet reduction processes, all of which require relatively complicated and expensive equipment, include the following: 7. The IFP (14) ammonia absorption process starts with the absorption of sulfur dioxide in an ammonia solution. The spent absorbent containing ammonium sulfite and 10.1021/es011459b CCC: $22.00
2002 American Chemical Society Published on Web 05/11/2002
bisulfite is decomposed by heating in an evaporator, and the resulting ammonia, sulfur dioxide and water vapor are sent to an IFP liquid-phase Claus reactor into which hydrogen sulfide is injected. The hydrogen sulfide and sulfur dioxide react to form elemental sulfur, and the ammonia, which is not affected, passes through the reactor and is recycled to the absorber. 8. In the Citrate Process (15), recovery of elemental sulfur from sulfur dioxide emissions in waste gases is done by the use of an absorbing solution containing thiosulfate, which is formed in the regeneration step. The pregnant absorbent is reacted with hydrogen sulfide to yield elemental sulfur. 9. The Bio-FGD process developed by Paques and Hoogovens (4) consists of four basic steps: absorption, anaerobic reaction, aerobic reaction, and sulfur separation. Sulfur dioxide contained in a flue gas is removed in an absorber fitted with water sprays. The mostly sulfitescontaining liquid is then fed to two biological reactors, where the sulfur-containing compounds are converted to sulfur. In the first anaerobic reactor, reduction with the aid of sulfatereducing bacteria forms hydrogen sulfide. In the second reactor, H2S is selectively oxidized to sulfur under a controlled biological environment. However, sulfur cake produced in this process contains many impurities. Because of the various problems associated with each of these previous processes, none are currently used commercially. The present research is concerned with developing a new process for converting sulfur dioxide gas to elemental sulfur that is relatively simple and without generating any solid wastes. The proposed scheme involves a cyclic process of the reduction of sulfur dioxide by calcium sulfide and the regeneration of calcium sulfide by the reduction of the product calcium sulfate with a suitable reducing agent such as hydrogen. The overall process would begin with calcium sulfate as the starting material.
2. Choice of Reaction System The following two types of processes would satisfy the requirements stated above for a new process for converting sulfur dioxide to elemental sulfur without any secondary environmental problems: (1) A process in which the reaction of sulfur dioxide produces a gas phase product containing essentially pure sulfur: The solid product in this case must be reusable in the process, because it is undesirable to generate large amounts of waste solids that require disposal. (2) A process in which the reaction of sulfur dioxide produces a solid product containing excess sulfur which can easily be decomposed in a separate step to yield pure sulfur gas and a reusable solid residue: In this case, the off-gas from the reaction step should only contain species that can be vented directly or with a minimum amount of treatment such as afterburning. With this in mind, a number of possible reaction systems, including the NiS-SO2, Fe-SO2, FeS-SO2, FeS-C-SO2, and CaS-SO2 systems, were examined based on thermodynamic analysis (16). Equilibrium compositions at various temperatures were calculated to find the most suitable reaction system for converting sulfur dioxide to elemental sulfur without generating secondary environmental problems. The thermodynamic analysis was performed by the use of HSC chemical software developed by Outokumpu Research Oy, which is based on the principle of the Gibbs free energy minimization. Based on the results of these thermodynamic analyses, the CaS-SO2 system showed the best promise in terms of costs and meeting the first of the two possible requirements discussed above. This reaction is represented by
CaS(s) + 2SO2(g) ) CaSO4(s) + S2(g)
(1)
FIGURE 1. Equilibrium composition for the reaction of 1 mol of CaS with 1 mol of SO2 together with 0.1 mol of H2O at a total pressure of 1 atm. The calculated equilibrium compositions for this system, in which several other possible compounds in the Ca-SO(-H) system were considered, are shown in Figure 1. This figure shows the equilibrium amounts of the major species present when 1 mol of SO2 is reacted with 1 mol of CaS together with 0.1 mol of H2O, considering that some SO2 streams may contain water vapor depending on the source. It is seen that at temperatures up to about 1050 K the solid product will be calcium sulfate and the gaseous product will be mainly sulfur as S2, the amounts of other sulfur species S3 and higher being negligible, mixed with small amounts of hydrogen sulfide and unreacted sulfur dioxide. After sulfur is separated by condensation, these gases would be recycled to the feed stream, scrubbed, or fed to a sulfuric acid plant, if one is available nearby. The amounts of CaO, SO3, and other compounds considered were negligible. If the feed sulfur dioxide stream contains oxygen, it reacts with CaS to form calcium sulfate. If the sulfur dioxide stream is dry, as expected for those generated from nonferrous smelters, the gaseous product will not contain hydrogen sulfide. The product calcium sulfate is reduced to calcium sulfide, as will be explained later. This system presented the greatest potential as the candidate for a new process for converting sulfur dioxide to elemental sulfur and thus was investigated further by carrying out experimental work.
3. Thermodynamics of the Reduction of Calcium Sulfate to Calcium Sulfide The hydrogen reduction of calcium sulfate, presented below, is of interest as a means of regenerating calcium sulfide from calcium sulfate produced in the CaS-SO2 system as well as producing sulfur from gypsum and anhydrite:
CaSO4(s) + 4H2(g) ) CaS(s) + 4H2O(g)
(2)
The calculated thermodynamic data presented in Figure 2 indicate that the products indeed are CaS and water vapor together with a small amount of hydrogen sulfide. The small amount of CaO that might be generated would be converted to CaSO4 when the solids are reused for reaction 1.
4. Experimental Work Experiments were carried out using a thermogravimetric analysis (TGA) apparatus, which is often and conveniently used to study gas-solid reactions (16, 17). It consisted of three major partssan electrobalance, a reactor, and a gasdelivery system. The weight changes taking place during the reaction were measured continuously using a Cahn balance (Model 1000). In the course of the experiments, the balance chamber was purged continuously with nitrogen to prevent the intrusion of reactant gas and heat into it. VOL. 36, NO. 13, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3021
FIGURE 2. Equilibrium composition for the reaction of CaSO4 (1 mol) with H2 (3 mol) at a total pressure of 1 atm.
FIGURE 3. X-ray pattern of intermediate product solid from the CaS-SO2 reaction.
Sulfur dioxide gas (99.9%) was supplied by Union Carbide Co., and hydrogen (99.9%) and nitrogen (99.9%) gases were purchased from Air Products Co. The calcium sulfide powder, manufactured by Alfa AESAR Co., was 99.9 wt % pure, and the calcium sulfate powder (99.5%) was obtained from Aldrich Chemical Co. Both materials were of -44 µm size as determined by sieving. For testing the effect of a nickel catalyst, calcium sulfate powder was impregnated with nickel by mixing 5 g of calcium sulfate particles in 100 mL of nickel nitrate solution (5 wt %) for 24 h. It was then dried at 443 K for 3 h in an oven and, when needed, reduced to calcium sulfide by hydrogen at 1153 K for 1 h in a horizontal furnace. Reagent grade nickel nitrate [Ni(NO3)2‚6H2O] manufactured by Alfa AESAR Co. was used to prepare the solution. The calcium sulfide powder is very reactive to moisture and was thus stored in an evacuated desiccator. Pellets were prepared from a powder mixture containing 85% nickel-impregnated calcium sulfate, 10% ammonium carbonate, and 5% bentonite as a binder. Ammonium carbonate was added to produce pellets of high porosity. The powders were thoroughly mixed. After adding to it distilled water of 5% weight, the mixture was again thoroughly mixed and hand-rolled to spherical pellets. The green pellets were dried in air at 443 K for 2 h, followed by being heated at 923 K for 6 h to burn off any organic residues and to indurate them. By this procedure, strong spherical pellets of about 0.5 cm and 44-52% porosity were obtained. The reaction cycle was repeated to the tenth cycle, during which the physical integrity of the pellets remained intact.
FIGURE 4. X-ray pattern of solid product from the hydrogen reduction of CaSO4.
5. Results and Discussion 5.1. Reaction Products. For the CaS-SO2 reaction, a gas mixture containing nitrogen and sulfur dioxide at a partial pressure of 8.7 kPa was passed over the calcium sulfide held in an alumina boat at 1073 K for various reaction times. Figure 3 shows the X-ray pattern of the final solid phase. It is seen that the solid contained only calcium sulfate and unreacted calcium sulfide. The XRD pattern of the solid collected by cooling the product gas revealed only sulfur (16). These results indicate that sulfur gas can be produced by the reaction between calcium sulfide and sulfur dioxide without any appreciable side reactions. For the CaSO4-H2 reaction, hydrogen at a partial pressure of 8.7 kPa was passed over the calcium sulfate powder at 1073 K for various lengths of time. Figure 4 represents the result of X-ray analysis for the system, which indicates that only calcium sulfide was detected in the solid product. 5.2. Reaction of Sulfur Dioxide with Calcium Sulfide Pellets. The conversion at a particular time was determined 3022
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 13, 2002
FIGURE 5. Effect of reaction temperature on the reaction of CaS pellets with SO2 in the first cycle (pSO2 ) 25.8 kPa). by dividing the mass change of the solid sample at the time by the theoretical maximum total mass change. External mass transfer and interpellet diffusional effects were eliminated by using a sufficiently large gas flow rate and sufficiently porous pellets (16). Shown in Figures 5 and 6 are the conversion-vs-time relationships for the reaction at various temperatures under a sulfur dioxide partial pressure of 25.8 kPa of calcium sulfide pellets prepared by the hydrogen reduction of fresh nickelcatalyzed calcium sulfate pellets and those regenerated in the second cycle [reaction in the third cycle]. At 1088 K, the conversion of fresh nickel-catalyzed calcium sulfide pellets was around 40% in 10 min under sulfur dioxide partial pressure of only 25.8 kPa. These curves were reproducible within ( 3.0%. It is seen in Figure 6 that the reaction rate
ature (16), the results of which yielded the following rate expression for the conversion-vs-time relationship
(
XCaS ) λ‚Ln 1 +
)
kapp1t λ
(3)
where
kapp1 ) (1/2)‚k1‚pSO2
(4)
As an example, the rate parameters in the third reaction cycle were FIGURE 6. Effect of reaction temperature on the reaction of CaS pellets with SO2 in the third cycle. (pSO2 ) 25.8 kPa).
1 2447 ) 0.529 exp λ T
( )
(5)
and
(
k1 ) 180 exp -
FIGURE 7. Effect of sulfur dioxide partial pressure on the reaction of CaS pellets with SO2 in the third cycle (T ) 1073 K).
FIGURE 8. Effect of reaction temperature on the hydrogen reduction of CaSO4 pellets in the fifth cycle (pH2 ) 86.1 kPa). increased with temperature up to 1088 K but became lower as the temperature was further raised. This was due to the sintering of the pellets at these higher temperatures, as verified by an SEM analysis of the pellets (16). The effect of sulfur dioxide partial pressure is shown in Figure 7, from which the rate dependence was determined to be first order (16). It is noted that the reaction rate was measured and is presented in terms of the conversion of solid CaS under constant sulfur dioxide partial pressures. The rate of the conversion of a given amount of sulfur dioxide, which is of the ultimate interest, depends on the relative amount of CaS it is contacted with, the larger the latter the faster the conversion. The rate and extent of conversion can be readily calculated by the use of the complete rate expression discussed below that includes the rate dependence on sulfur dioxide partial pressure. Considering that 2 mol of SO2 react with 1 mol of CaS, the SO2 conversion rate should be quite rapid with a reasonable excess molar ratio of CaS to SO2. A complete rate analysis was then carried out for the reaction in the temperature range 1023-1088 K by combining these effects of sulfur dioxide partial pressure and temper-
12 200 kPa-1‚min-1 T
)
(6)
The calcium sulfide pellets made from fresh calcium sulfate pellets had somewhat higher reactivity, but the rate parameters for the third through the tenth cycles remained close to those given above (16). This is important because the solids must be reusable for repeated cycles to avoid generating solid waste. 5.3. Regeneration of Calcium Sulfide Pellets by the Hydrogen Reduction of the Produced Calcium Sulfate Pellets. The conversion-vs-time relationships for the hydrogen reduction of nickel-catalyzed calcium sulfate pellets in the fifth cycle are shown in Figure 8 for various temperatures under a hydrogen partial pressure of 86.1 kPa (atmospheric pressure at Salt Lake City). At 1088 K the reduction was essentially completed in 20 min. These curves were reproducible within ( 3.0%. The hydrogen reduction of nickel-catalyzed calcium sulfate pellets was always allowed to proceed to completion. However, the reaction of the product calcium sulfide that follows a pore-blocking mechanism, as indicated by eq 3 would have taken too much time to go to completion. Thus, this reaction was interrupted at certain conversion levels at different cycles, which means that regenerated calcium sulfate pellets always contained some calcium sulfide (as indicated by the “fractional conversion” of 0.498 at zero time). It is noted, however, that the regenerated nickel-catalyzed calcium sulfate portion can be reduced relatively fast. The rate dependence of this reaction on hydrogen partial pressure was very low at 0.20 due to the strong adsorption of hydrogen on nickel surface. A complete rate analysis was carried out (16), the results of which yielded the following rate expression for the conversion-vs-time relationship:
[ ( -ln
)]
o 1 - XCaSO 4
1 - XCaSO4
1/1.52
) kapp2t; o XCaSO4 ) XCaSO at t ) 0 (7) 4
where 0.20 kapp2 ) (1/4)‚k2‚pH 2
(8)
and
(
k2 ) 2.8 × 107exp -
20 300 kPa-0.20‚min-1 T
)
(9)
The fresh nickel-catalyzed calcium sulfate pellets had somewhat lower reactivity, but the rate parameters for the fifth through the tenth cycles remained close to those given VOL. 36, NO. 13, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3023
above (16). Again, this is important because the solids must be reusable for repeated cycles to avoid generating solid waste, as mentioned earlier.
Acknowledgments This work was supported in part by a Faculty Research Grant from the University of Utah Research Committee.
Literature Cited (1) Dalton, S. M. in Symposium Series on Desulfurization 3; Kyte, W. S., Ed.; Chameleon Press: London, 1993; pp 67-77. (2) Friedman, L. J. In Sulfur Dioxide Control in Pyrometallurgy; Chatwin, T. D., Kikumoto, N., Eds.; TMS: Warrendale, PA, 1981; pp 205-220. (3) Asteljoki, J. A.; Bailey, L. K.; George, D. B.; Rodolff, D. W. J. Metals 1985, 37(5), 20-23. (4) Kwong, V.; Meissner, R. E. Chem. Eng. 1995, 102, 74-83. (5) Weisenberg, I. J.; Winkler, F. M.; Burckle, J. O. In Sulfur Dioxide Control in Pyrometallurgy; Chatwin, T. D., Kikumoto, N., Eds.; TMS: Warrendale, PA, 1981; pp 33-53. (6) Burckle, J. O.; Worrell, A. C. In Sulfur Dioxide Control in Pyrometallurgy; Chatwin, T. D., Kikumoto, N., Eds.; TMS: Warrendale, PA, 1981; pp 55-65. (7) Agarwal, J. C.; Loreth, M. J. In Sulfur Dioxide Control in Pyrometallurgy; Chatwin, T. D., Kikumoto, N., Eds.; TMS: Warrendale, PA, 1981; pp 67-89. (8) Bouillon, D. F In Restoration and Recovery of an Industrial Region; Gunn, J. M., Ed.; Springer-Verlag: New York, 1995; pp 275-285.
3024
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 13, 2002
(9) George, D. B. In Sulfide Smelting 2002; Stephens, R. L., Sohn, H. Y., Eds.; TMS: Warrendale, PA, 2002; pp 3-13. (10) Puricelli, S. M.; Grendel, R. W.; Fries, R. M. In Sulfide Smelting ’98: Current and Future Practices; Asteljoki, J. A., Stephens, R. L., Eds.; TMS: Warrendale, PA, 1998; pp 451-462. (11) Sander, U. H. F.; Fischer, H.; Rothe, U.; Kola, R. Sulphur, Sulphur Dioxide and Sulphuric Acid; The British Sulphur Corporation Ltd.: London, 1984; pp 90-95. (12) Katz, M.; Cole, R. J. Ind. Eng. Chem. 1950, 42, 2258-2269. (13) Nelson, S. G. In Processing and Utilization of High-Sulfur Coals; Parekh, V, B. K., Groppo, J. G., Eds.; Elsevier: New York, 1993; pp 543-553. (14) Semrau, K. In Sulfur Removal and Recovery from Industrial Processes; Pfeiffer, J. B., Ed.; Mills-Frizell-Evans Co.: Maryland, 1975; pp 1-22. (15) Korosy, L.; Gewanter, H. L.; Chalmers, F. S.; Vasan, S. In Sulfur Removal and Recovery from Industrial Processes; Pfeiffer, J. B., Ed.; Mills-Frizell-Evans Co.: Maryland, 1975; pp 192-211. (16) Kim, B.-S. Ph.D. Dissertation, University of Utah, Salt Lake City, UT, 1999. (17) Szekely, J.; Evans, J. W.; Sohn, H. Y. Gas-Solid Reactions; Academic Press: New York, 1976; pp 209-213.
Received for review December 5, 2001. Revised manuscript received April 4, 2002. Accepted April 12, 2002. ES011459B
