High temperature hydrogen sulfide and carbonyl sulfide removal with manganese oxide (MnO) and iron oxide (FeO) on .gamma.-alumina acceptors
Post on 18-Feb-2017
Embed Size (px)
Ind. Eng. Chem. Res. 1993,32, 139-149 139
Kojima, T.; Furusaki, S.; Saito, K. A. Fundamental Study on Re- covery of Copper with a Cation Exchange Membrane Part 2- Transfer Rate of Copper and Hydrogen Ion Through a Cation Exchange Membrane. Can. J . Chem. Eng. 1982, 60, 650458.
Lonsdale, H. K. The Growth of Membrane Technology. J . Membr.
Martell, A. E.; Smith, R. M. Critical Stability Constant; Plenum: New York, 1974; Vol. 1.
Maetsuyama, H.; Fujii, K.; Teramoto, M. Selective Separation of Rare Earth Metals by Donnan Dialysis in the Presence of Water-Soluble Complexing Agent. J . Chem. Eng. Jpn. 1991,24,
Morel, F. M. M. Principles of Aquatic Chemistry; Wiley-Intersci- ence: New York, 1983; Chapter 6.
Neihof, R.; Sollner, K. The Physical Chemistry of the Differential &tea of Permeation of Ions Across Porous Membranes. Disscws. Faraday SOC. 1956,21, 94-101.
S C ~ . 1982, 10, 81-181.
Newman, J. S. Electrochemical Systems; Prentice-Halk Englewd Cliffs, NJ, 1960; Chapter 7.
Sudoh, M.; Kameri, H.; Nakamnura, S. Donnan Dialysis Concen- tration of Cupric Ions. J. Chem. Eng. Jpn. 1987,20, 34-40.
Takahashi, K.; Tsuboi, K.; Takeuchi, H. Mass Transfer across Cation Exchange Membrane. J . Chem. Eng. Jpn. 1989, 22, 352-357.
Wallace, R. M. Concentration and Separation of Ions by Donnan Membrane Equilibrium. Ind. Eng. Chem. Process Des. Deu. 1967,
Wen, C. P.; Hamil, H. F. Metal Counterion Transport in Donnan
Wills, G. B.; Lightfoot, E. N. Membrane Selectivity. MChE J. 1961,
Received for review April 28, 1992 Revised manuscript received September 16, 1992
Accepted October 3, 1992
Dialysis. J . Membr. Sci. 1981, 8, 51-68.
High Temperature H2S and COS Removal with MnO and FeO on yA1203 Acceptors
J. Peter Wakker,? Albert W. Gerritsen, and Jacob A. Moulijn* Department of Chemical Engineering, Delft University of Technology, Julianalaan 136, 2628 BL Delft, The Netherlands
A process has been developed to remove H2S and COS at high temperatures (675-1075 K) from fuel gases produced by a coal gasification plant. The process uses MnO or FeO on y-A1203 acceptors which can be regenerated after sulfidation with a gas containing steam. The acceptors should be used in a fixed bed or moving bed reactor. It is important to know what influence the raw fuel gas composition of a coal gasification plant and operating temperature and pressure have on the sul- fidation behavior of the acceptors. Experiments are carried out in a laboratory setup to determine these influences. After an initial deactivation the acceptors can be used over 400 sulfidation-re- generation cycles. Water adversely affects the capacity, whereas CO favors the capacity due to the watewas shift reaction. Trace compounds can influence the capacity negatively as, e.g., HC1, whereas hydrocarbons do not influence acceptor performance. An optimum between capacity and stability is reached a t a operating temperature of 875 K. An increasing pressure does not influence the sulfidation reaction, but the water-gas shift reaction may be positively influenced.
Introduction The composition of a raw fuel gas produced by a coal
gasifier can vary over a wide range, depending upon the type of gasifier (fixed bed, fluidized bed, entrained bed) and the type of coal used. The main compounds produced by a gasifier are CO, H2, COz, HzO, N2, H2S, and COS. There are large differences in gas composition. The hy- drogen concentration varies from 29 to 40 vol %, the carbon monoxide concentration even from 16 to 65 vol % , the carbon dioxide concentration between 1 and 31 vol % , water between 1 and 20 vol %, hydrogen sulfide plus carbonyl sulfide between 0.1 and 1.5 vol %, and methane (and traces of other hydrocarbons) between 0 and 10 vol % (Brainard, 1982a,b; Miles, 1976; Nelson, 1979a-c; Schilling et al., 1979; Schlinger and Richtor, 1980; Shap- pert, 1983; Synthesis Fuel Associates, 1983). The fuel gas can be used to produce electricity, as a synthesis gas or as a reducing gas. In this paper we mainly consider the use of the fuel gas to produce electricity in a coal gasification combined cycle (CGCC) power plant. Before the fuel gas can be used it must be cleaned. A wide variety of existing commercial and well-developed gas processing systems is available for the removal of sulfur, ammonia, and other undesirables from the fuel gas. In general these are low
Present address: Akzo Chemicals bv., P.O. Box 10,7400 AA Deventer, The Netherlands.
temperature processes involving the use of various liquids with either an organic or aqueous base. Consequently, the gas must be cooled before admission to the scrubbing towers. However, when gas cleaning is carried out at high temperatures, the overall efficiency can be improved and lead to lower kilowatt-hour prices. In high temperature gas cleaning the following steps can be distinguished: dust removal, desulfurization, removal of halogens, removal of alkali metals and other trace compounds, and the removal of nitrogeneous compounds.
A number of proceases for high temperature regenerative H2S removal is currently under investigation. The most important ones use zinc ferrite (e.g. Ayala and Marsh, 1991), zinc titanate (e.g. Gangwal et al., 1989) and copper based sorbents (Desai and Brown, 1990). The main dis- advantages of these processes are the limited applicability and the regeneration route. When these sorbent materials are used in highly reducing gases, as, e.g., the case with the Shell gasification process, the metal oxide is reduced into the metallic state, or metal carbides are formed. This results in loss of strength and capacity of the sorbent.
Regeneration of these sorbents is carried out with an oxygen/nitrogen mixture. The regeneration is a highly exothermic reaction, and therefore only low oxygen con- centrations can be used. SOz is formed in low concen- trations. Further, the regeneration is complicated by sulfate formation. Due to the oxidative character of the regeneration, inevitable hydrogen consumption takes place
0888-5885/93/2632-0139$04.00/0 0 1993 American Chemical Society
140 Ind. Eng. Chem. Res., Vol. 32, No. 1,1993
during the next sulfidation. This hydrogen can be better used to produce electricity.
Two processes which possibly can produce elemental sulfur during regeneration are the mixed metal oxide process (Anderson and Hill, 1988) and iron oxide on silica (Wal, 1987), but both processes have not been proven yet.
The process discuased in this paper can overcome several of the disadvantages of the above processes. The acceptors tested are able to remove H2S from the fuel gas to levels less than 20 ppm according to the following reaction: MeO/7-A1203 + H2S e
Me = Mn or Fe (1) It has been proven by Wakker (1992) that the active compound in this reaction is a metal aluminate, MeA1204 (Me = Mn or Fe), making steam regeneration of the sul- fided acceptor material possible. Also other reactions occur, e.g., the water gas shift reaction and COS formation and capture. COS may be removed by direct reaction with the acceptor, or may be converted to H a which reacts with the acceptor according to reaction 1. The performance of the acceptors is strongly dependent upon gas composition and process variables such as temperature. Especially the water concentration is expected to have a strong influence on the acceptor performance as water shifts the equilib- rium in reaction 1 to the left. Therefore, the acceptor can be regenerated with a gas containing steam and H2S is formed back in a high concentration.
Moreover, the gas produced by a coal gasifier can contain a number of compounds which may deactivate the acceptor or influence the acceptor performance negatively by blocking active sites. Among these compounds are heavy metals, HC1, HF, NH3, and HCN, and hydrocarbons. Because it is impractical to add heavy metals in low con- centrations to a synthetic gas mixture, one best can observe the influence of heavy metals by carrying out experiments under realistic conditions. These experiments are de- scribed elsewhere by Wakker and Gerritsen (1990b). The expected HCl concentrations for the Shell and Texaco gasification processes, fed with Australian Wambo coal, are 600 and 500 ppm and the HF concentrations 224 and 188 ppm, respectively. Wambo coal only contains 0.05 wt % Cl and 0.01 wt % F on a dry base (Alderliesten et al., 1990). Tests done with HC1 and hydrocarbons will be discussed in this paper as well.
Experimental Section Acceptor Preparation. The acceptor material is pre-
pared by wet impregnation. The carrier (7-A1203, AKZO/Ketjen 001-1.5E, ground and sieved, particle di- ameter 0.25-0.42 mm, surface area 260 m2/g) is added to the impregnation solution (3.3 mL of solution/g of carrier) and shaken a few times during impregnation. The im- pregnation time is about 16 h. For preparation of a manganese containing acceptor, a 2 M manganese acetate (Merck, pa.) solution is used. An iron containing acceptor is prepared from an 1 M iron(II1) ammonium oxalate (Riedel de Hiihn, p.a.) solution. After impregnation the material ie filtered, dried at room temperature during 5-6 h, and calcined in static air a t 575 K during 60 h. The manganese is now present as Mn02 or Mn203, and iron, as Fez03 or Fe30,. The last step in preparation is reduc- tion, carried out in situ in the reactor a t reaction tem- perature to obtain the active metal aluminate, MnAl2O4 or FeA1201. The manganese containing acceptors consisted of 7.10-8.23 wt % Mn on 7-A120, and the iron containing acceptor of 4.64 wt % Fe on 7-Al2OP Fresh acceptors have a surface area of 180-200 m2/g and an average pore di- ameter of 4-5 nm.
MeS/y-A1203 + H20
Instrumental Details. The apparatus used to obtain the sulfidation-regeneration cycles is described in detail by Wakker and Gerritsen (1990a). The laboratory setup consists of a gas dosage section, a fixed bed reactor, and an analysis section. In the gas dosage section oxygen and water are removed from all gases, except for H2S, HC1, and hydrocarbons which are used without further purification. After passing mass flow controllers, the gas flows are mixed and flow to the reactor (a quartz tube). The reactor feed contains 0-70 vol % CO, 0-40 vol '3% H2, 0-15 vol ?& H20, 0-10 vol % hydrocarbons, 0-1 vol % HC1,0-1 vol % COS, 0-1 vol % H2S, the balance being Np The tests are carried out a t temperatures between 675 and 1075 K, using 3 g of acceptor material. The total flow rate was always kept at 67.5 pmol/min. When tests are carried out with water in the feed gas, small fluctuations in the H2S concentration leaving the reactor are observed due to a not completely stable H20 dosage. After breakthrough, the H2S dosage is stopped and the acceptor material is regenerated with a gas containing steam. This is done by injecting liquid water into the reactor where it evaporates and is heated to reaction temperature before reaching the acceptor material. The amount of H2S leaving the reactor is mea- sured by a continuous titration system. An additional gas chromatographic analysis of the reactor effluent compo- sition is carried out when COS, C02, and H20 concentra- tions also have to be measured. After regeneration a new sulfidation-regeneration cycle is started.
When experiments were carried out a t elevated pressure, the quartz reactor tube was replaced by a stainless steel (RVS 316) reactor with the same internal diameter. This reactor could only be used up to 725 K because at higher temperatures the steel reacts with H2S, thus causing in- correct experimental data. Therefore another reactor, made of a ferritic stainless steel (an Al-CI-Fe alloy), which does not react with H2S up to temperatures of 1075 K, has been used as well. The internal diameter of this reactor was 12 mm. For the same bed length and residence time to be obtained, 5.1 g of acceptor was used and the flow rate was increased to 178 pmol/min. The pressure was varied in the range of 0.1-0.5 MPa.
Results and Discussion Acceptor Deactivation. Whenever a fresh acceptor
is tested, a stable performance is obtained only after an initial period of deactivation. During this period the suLfur removal capacity decreases with time. As shown in Figure lA, the deactivation process is quite strong during the first five to ten sulfidation-regeneration cycles a t 875 K; the breakthrough capacity (the amount of sulfur captured at breakthrough of the acceptor bed, or, in other words, the amount of sulfur captured at the moment the H2S con- centration in the reactor exit rises above 100 ppm) of a fresh manganese containing acceptor decreases from about 1.7 wt % S to about 0.8 wt % S. The decrease in break- through capacity between the first and the second cycle is caused by the sulfidation of MnO present as small crystallites which can be sulfded but not regenerated with steam. Temperature programmed sulfidation (TPS) (Wakker, 1992) measurements showed that the formation of MnA120, is relatively slow; not all MnA1204 is formed after reduction of the acceptor in the reactor. Moreover, part of the MnA120 cannot be sulfided as ita reactivity is too low at 875 K. Only when higher temperatures are applied (e.g. 1075 K) can all MnA1204 be sulfided.
In the next 40 sulfidation-regeneration cycles the deactivation slows down: f i y the breakthrough capacity becomes constant a t about 0.65 wt % S, but it may be higher for other acceptor batches. The decrease in
Ind. Eng. Chem. Res., Vol. 32, No. 1, 1993 141
c M 8
1 .oo R #
0.00 .... 0 50 100 150 200 250 300 350
Cycle number [-I Figure 1. Deactivation of MnO/y-A1203 and Fe0/y-Al2O3 at 875 K: (A, top) = 7.1 wt % Mn on y-A120,; (B, bottom) A = 4.6 wt % Fe on yA1203.
breakthrough capacity during these 40 cycles is caused by a decrease in the surface area of the acceptor; the surface area decreases from about 200 to about 140 m2/g. The sulfur capture capacity remains almost constant at this level for at least 400 sflidation-regeneration cycles. The same change in the breakthrough capacity can be observed for the iron containing acceptor (Figure 1B). Due to its lower metal content a fresh acceptor has a breakthrough capacity of about 1 wt % S. This value decreases to a stable capacity of about 0.45 wt % S.
The deactivation is negligibly slow at 657 K, whereas it is much faster and stronger at 1075 K. Deactivation is not caused by an accumulation of sulfur on the acceptor: the amount of sulfur regenerated equals the amount of sulfur captured, except for the first cycle.
Influence of Temperature on Acceptor Capacity. The influence of temperature on acceptor capacity can be obtained in two different ways. The first way is to stabilize the acceptor at a certain high temperature, here 875 or 1075 K, and then to measure the breakthrough capacity at lower temperatures, here from 675 to 875 K and from 875 to 1075 K. The results obtained in this way are shown in Figure 2 f...