Novel Supported Sorbent for Hot Gas Desulfurization Aysel T. Atlmtay,' Lee D. Gasper-Galvln, and James A. Poston
Morgantown Energy Technology Center, Morgantown, West Virglnla 26507 A sorbent which has been prepared by supporting copper oxide and manganese oxide on a silica-rich zeolite was developed for the removal of hydrogen sulfide from coal gas at high temperatures (around 850 "C). The sorbent is regenerable and more durable for fluidized-bed and moving-bed reactors. The sorbent has been shown through sulfidation and regeneration experiments to be capable of reducing the hydrogen sulfide level in coal gas from 2000 to 200 ppmv for about 60-70 min before breakthrough point. Regeneration of the sorbent with a 50 mol % air50 mol % steam mixture is preferred over air regeneration, due to decreased sulfate formation in the presence of steam. The metal coating on the zeolite extrudates, especially the Mn, appears to promote increased resistance to crushing.
Introduction The integrated gasification combined cycle (IGCC) system is one of the most promising advanced energy systems for producing electrical energy from coal because it can provide noticeable improvement in thermal efficiency. The technology for this system is currently being developed a t the Morgantown Energy Technology Center (METC) of the U.S.Department of Energy. When coal is gasified, the sulfur in the coal reacts with steam to form hydrogen sulfide and small amounts of other sulfurcontaining compounds. Hydrogen sulfide is a toxic gas which contributes to the formation of acid rain when it is oxidized to SO2 and/or SOa. It is, therefore, necessary to remove as much of the hydrogen sulfide as possible from the coal gasification stream prior to release to the atmosphere. Additionally, turbines and related equipment need to be protected from the corrosive action of sulfurous compounds in the coal gas. Hydrogen sulfide, which is the most abundant sulfurcontaining compound in coal gas, can react readily with the oxides of alkali earth and transition metals (1). Westmoreland and Harrison (2) reported the results of a thermodynamic screening of the high temperature desulfurization potential of 28 single metal oxides by using the free-energy minimization method. Also, kinetic limits for ZnO, MnO, CaO, and V203 were reported by Westmoreland et al. (3) for desulfurization reactions. ZnO
* Address correspondence to this author at her present address: Environmental Engineering Department, Middle East Technical University, 06531 Ankara, Turkey. 0013-936X/93/0927-1295$04.00/0
0 1993 American Chemical Society
showed a very favorable thermodynamic equilibrium with H2S ( 4 ) , and iron oxide showed easy regenerability with air. Results regarding structural changes in the pure ZnO sorbent at high temperatures were reported by Ranade and Harrison (5). They showed that sintering actually causes the particle to shrink radially; thus, for the same number of particles, the surface area available for reaction is smaller. Additionally, zinc metal produced by reduction of ZnO in coal gas volatilizes above 1290 O F (699 "C). Favorable properties of zinc oxide and iron oxide toward reacting with H2S were combined at the in-house research program of the Department of Energy's Morgantown Energy Technology Center. The combined metal oxide sorbent, zinc ferrite, was shown to reduce HzS levels in gasifier gases to less than 10 ppmv (6). However, zinc ferrite has a temperature limitation. The maximum operating temperature for zinc ferrite is about 1200 OF (649 "C). Focht, et al. (7)showed that the sorbent lost its pore volume during reduction near 1380 O F (749 "C), and the subsequent loss of reactivity was observed during sulfidation. The sorbent also showed a substantial amount of decrepitation above 1200 O F (649 "C). Improvements in the formulation of zinc ferrite have been attempted in an effort to overcome its limitations. An inorganic binder, bentonite, was added to zinc ferrite. Thus, a sorbent with comparable reactivity and capacity to the original sorbent, but with a higher temperature limit than the original zinc ferrite, was obtained. This sorbent can withstand temperatures up to 1275 O F (691 "C), which is not very satisfactory for high temperature fluidizedand entrained-bed gasifiers. The need for developing a sorbent having better properties (durability and reactivity) than the original or modified zinc ferrite led researchers to try other metal oxides. Various combinations of metal oxides, such as Zn, Cu, Al, Ti, Fe, Co, Mo, and V were tried at Massachusetts Institute of Technology (MIT) and Jet Propulsion Laboratory (JPL) in an effort to develop a durable and high capacity sorbent for removal of H2S from hot coal-derived gases (8).Basically, the sorbents tested were divided into two groups: (1) zinc-based sorbents and (2) copper-based sorbents. Zinc-based sorbents included zinc ferrite (ZF), zinc-copper ferrite (ZCF),and zinc titanate (ZT). Copperbased sorbents included copper aluminate (CA) and coppepiron aluminate (CFA). Research results(8)showed that zinc-based sorbents were more promising than copperbased sorbents in the temperature range of 1000-1350 O F (538-732 "C). One of the most promising sorbents among Envlron. Sci. Technol., Voi. 27, No. 7, 1993
1285
the zinc-based sorbents was zinc titanate [maximum operating temperature of 1340 O F (727 "C)]. It was found that the Ti02 phase helps to stabilize the ZnO phase, and volatilization of Zn at temperatures higher than 1290 O F (699 "C) has been reduced (about 1wt % Zn is lost in 1000 h of operation). However, Ti02 does not react with H2S. Therefore, the capacity of zinc titanate is lower than that of zinc ferrite due to a dilution effect. Extrudates of regenerable mixed metal oxide sorbents (copper-modified zinc ferrite, zinc titanate, copper aluminate, and copper-iron aluminate) were tested in a high temperature-high pressure (HTHP) fixed-bed reactor by ResearchTriangle Institute (RTI) (9). Reductions in H2S concentration from >10 000 ppmv to 0 I-
3
600
3
200
0 20
0
60
40
80
100
TIME ( m i n )
Figure 4. Effect of temperature on hydrogen sulflde breakthrough curves for muttlple sulfldatlon of SPlC5M5 sorbent. 1200 r
--c 2 n d Sulfidotlon
+3 rd
-L -v
1000
,
Sulfldotion
-0- 4 t h Sulfidotlon -G5th Sulfldotion -0-
SPZCSMS
Izx)oF ( 6 7 7 o c )
6 th Sulfldation
0
0 0
800
lI
W
/
0 -I
v) 3
600 W c ( I
>
I I-
9
400
3
0
200
0
0
20
40
60
80
100
I20
140
160
TIME (min)
Figure 5. Hydrogen sulflde breakthrough curves for muttlple sulfldations of SP2C5M5 sorbent at nominal 1250 O F (677 'C) with 50 mole steaml50 mole air used for regeneration: regenerabllky.
was 2 in. The gas lines between the reactor outlet and the condenser inlet were heated via heating tapes and were heavily insulated to prevent the condensation of steam in the gas lines. The exit gas from the reactor was filtered, cooled down in the condenser,and sampled for gas analysis. The tail gas was sent through the absorbers to remove the sulfurous gases before it was discharged to the atmosphere. The experiments consisted of a sequence of sulfidation and regeneration runs. The reactor temperature was held constant during the sulfidation and regeneration experiments. Reactor pressure was maintained at 5 f 3 psig. Three temperatures were tested; 1000,1250,and 1600 OF (538,677,and 871 OC,respectively). The space velocities for the sulfidation and the regeneration runs were 2000 and 600 hr-l, respectively. Regeneration of the sulfided sorbent was conducted with air only and with a 50 % air/ 50% steam (by volume) mixture. Breakthrough during 1298
Envlron. Sd. Technol., Vol. 27, No. 7, 1993
sulfidation runs was defined as 200 ppmv H2S in the effluent gas (however,for environmental purposes it would be preferable to have pre-breakthrough averages of 50 ppmv or less). After the sulfidation run was stopped, the reactor was purged with nitrogen for at least 15 min. Regeneration was carried out a t the reaction temperature. The regeneration cycle was stopped when the SO2 concentration of the effluent gas was below 500 ppmv. The sorbents were subjected to five sulfidation/regeneration cycles to establish sulfur sorption capacity and regenerability. Selected sorbent samples were analyzed for total weight, metal and sulfur content, crush strength, surface area, average pore diameter, pore volume, and pore size distribution both before and after reaction. Gasgrab samples obtained during experimentation were analyzed by gas chromatography. Also Gas-tech precision gas detection
COPPER OXIDE SORBENT AT 16OO0F 1871 'C
.............. I ST -_ --- 2 N D 3RD --- - --- 4 T H -.. - 5TH --- ---- 6 T H *
-*(
:-..--0
20
40
RUN RUN RUN RUN RUN RUN
I
I
I
I
6D
eo
100
120
TIME
I40
(mid
Figure 8. Hydrogen sulfide breakthroughcurves for muitiple sulfldatlons of zeolite-supported copper oxlde sorbent at nomlnal 1600 O F (871 OC).
tubes were used to determine the HzS and SO2 concentrations in the inlet and outlet gas streams.
Results and Discussion The effect of temperature on sorbent capacity is shown in Figure 4. The H2S breakthrough curves indicate that the breakthrough times were 38,49,and 17 min at 1000 O F (538 OC), 1250 O F (677 "C), and 1600 O F (871 "C), respectively. The sorbent has greater capacity a t 1250 O F (677 "C) and 1600 O F (871 "C) than at lo00 O F (538OC). However, the pre-breakthrough HzS levels are lower at lo00 O F (538"C) and 1250 O F (677 OC) than a t 1600 O F (871 "C). Although the pre-breakthrough HzSlevel is lower at lo00 O F (538"C), the sorbent has more total capacity a t the higher temperatures. Two runs were conducted at 1250 O F (677 "C) with different samples from the same batch of SPlC5M5 sorbent to establish the extent of reproducibility of the experimental method. Twelve detector tube readings of H2S were taken, 5 min apart, for each run. The average difference between readings for the two runs was 30ppmv. Thus, the experimental method appears to be reasonably reproducible. Sorbent regenerability was tested at 1250 O F (677 OC). Two and one-half cycles were conducted with only air as the regenerant gas, and five cycles were studied using a 50 mol % steam/50 mol % air mixture for regeneration. Two criterion were used to assess the regenerability of the sorbent: (1)a comparison of the HzSbreakthrough curves and (2) a comparison of the SO2 outlet curves during the sulfidations. When the sorbent was regenerated with air, the second and third sulfidation breakthrough curves were somewhat similar. However, the pre-breakthrough average HzS outlet concentration was approximately 50 ppmv less for the third sulfidation than for the second sulfidation. Figure 5 shows the HzS breakthrough curves after the sorbent was regenerated with steam/air. The curve for the first sulfidation was not shown, as the reactor bypass was accidentally left open during the run. The third through sixth sulfidations, conducted with steam/air
regeneration, gave similar breakthrough curves. Thus, Figure 5 shows that the 50% steam/50% air (by volume) regeneration appears to give good regenerability. Outlet SO2 concentrations were compared for the second and third sulfidations (set I) using air regeneration, with the second, third, and fourth sulfidations (set 11) using steam/air regeneration. After 15 min on-stream, set I showed outlet SO2 concentrations between 500 and 600 ppmv. After the same period, set I1 yielded SO2 concentrations ranging between 30 and 450 ppmv. These results indicate that more sulfate was formed during air regeneration than during steam/air regeneration. This is not unexpected,sincethe presence of steam reduces the overall oxidative properties of the regenerant stream. In general, if SO2 is released in large quantities when reducing gases are introduced after oxidative regeneration, this indicates that sulfates have been formed during the regeneration. The release of such sulfurous gases must be avoided during sulfidation. It has been suggested in studies on zinc ferrite that a clean reducing gas could be used for reductive regeneration of the sorbent after oxidative regeneration and prior to sulfidation. However, the economics for carrying out a reductive regeneration step are poor, due to the capital cost of additionalvesselsand the introduction of complications into the overall operation of the system (14).
In order to determine the effect of combining the two metal oxides, namely, Cu oxide and Mn oxide, in the sorbent, separate experiments were run with Cu oxide and Mn oxide each separately supported on the zeolite SP-115 with the same sorbent preparation technique. The same gas composition was used as the reacting gas. The results of these experiments were given in Figures 6 and 7, respectively. As shown in the figures, the sulfur absorption capacity for Cu oxide was much greater than that for Mn oxide at 1600 O F (871 OC). By combining Cu oxide with Mn oxide (SP2C5M51, the efficiency and capacity of the sorbent was increased over that obtainable by using the zeolitesupported copper oxide alone. The result for the sorbent Environ. Scl. Technol.. Vol. 27, No. 7, 1993
1299
200c MANGANESE OXIDE SORBENT AT 1600'F
. . ~.......... .. --- - ---_ _ _ -- - - -- -
I500
-d
I ST
RUN
2NO 3RD
RUN
4TH
RUN
5TH
RUN
6TH
RUN
(871 ' C J
RUN
8 000 W
0 J
ln 3 W z
Pe 500 I
0
20
60
40
80
I20
IO0
140
T M E (min)
Flgure 7. Hydrogen sulfide breakthrough curves for multlple sulfldations of reolltasupported manganese oxMe sorbent at nomlnal 1800 O F (871 "C). 350
-' -5
2 n d Sulfidation ---b 3 r d 4th A 5th --t 6 t h
300
a
Sulfidation Sulfidation Sulfidation Sulfidotion
SP2C5M5 '2500F ( 6 7 7 0 c 1
i,
8
250
W
0 J
,7
200
> _1
z
20
I50
c
Y
2
100
50
0
0
20
40
60
80
I20
100
140
160
T I M E (rninl
Figure 8. Outlet concentrations of carbonyl sulfide for multiple sulfktlons of SP2C5M5 sorbent at nomlnal 1250 O F (677 "C) wlth 50 mol % steam/5O mol % air used for regeneration: the effect of cycle number (after steam/air regeneratlon) on carbonyl sulflde formatlon.
SP2C5M5 was available at 1250 O F (677 "C)as opposed to the result for zeolite-supported Cu oxide or Mn oxide alone at 1600 O F (871 "C). A better performance is expected from the sorbent SP2C5M5 at 1600 OF (871 "C) simply because of the temperature effect explained in Figure 4. Moreover, the addition of MnO to CuO increases the structural stability of the sorbent. As it will be explained in the later part of this section, the crush strength of zeolite SP-115has increased with the addition of metal coating, and Mn appears to promote increased resistance to crushing. Although no carbonyl sulfide was fed to the reactor in the simulated gas mixture, appreciable concentrations were 1900
Environ. Sci. Technol., Vol. 27, No. 7, 1993
detected in the outlet gas during sulfidation. The COS concentration was usually greater than 50 ppmv and often as high as 100-250 ppmv in the outlet gas. It is wellknown that certain zeolites can catalyzethe COS formation reactions. CO and C02 can react with H2S according to the following reactions (11):
H2S 4- co, HZS
+ CO
-P
--*
cos + H,O COS
+ H,
Kl,
K
= 2.2 x
K i m K = 3.2
X
(1) (2)
At temperatures between 440 O F (227 "C)and 1340 O F (727 "C), the thermodynamics of reaction 2 are more favorable than those of reaction 1. Figure 8 shows that COS production became a minimum after about 3&40
400
350
-L j L
-Nominal --.bNominol +Nominal -Nominal
lOOO+, 111 1250oF. 111 I S O e , ' Is1 1 2 S O t , 111
spI
CsM,
250
8 w
0
j 200 v)
i
& m
5
IS0
0 IJ
3 c
O
100
I
20
40
1
I
so
eo
I
TIME (min)
Figure 9. Outlet concentrationsof carbonylsulfldefor sorbent SPlC5M5 for (a-c) single sulfldations and (d) a second suifidatlon after air regeneration: the effect of temperature and cycle number (after air regeneration) on the carbonyl sulfide formation.
min on-stream and then increased. The initial decrease in COS may have been due to reduced availability of CO, because of the competing reactions in which CO reduces copper oxides. The later increase in COS formation may have been caused by an increase in the availability of CO after the copper oxides have been reduced. The effect of increased temperature on COS formation can be seen in Figure 9. It appears that 1250 OF (677 "C) is a more favorable temperature for COS formation than either 1000 O F (538 "C) or 1500 O F (816 "C). The figure also suggests that COS production may increase with the number of cycles. More data need to be collected to explain this result. The effect of a single loading of the metal acetate solution versus a double loading of the metals is shown in Figure 10. As expected, the double-coated sorbent, SP2C5M5, has a breakthrough time that is nearly double that of the single-coated SPlC5M5. Thus, the capacity of the sorbent extrudates can be increased via multiple loadings of the metal salts. Sorbents SP2C5M5 and SP2C9M1 were also compared for their ability to remove H2S from the simulated coal gas mixture. The second and third sulfidations of each sorbent are shown in Figure 11. The SP2C9M1 sorbent appears to have greater capacity than SP2C5M5 for H2S removal. The surface area of the sorbent was determined before and after the reaction by using the N2 adsorption method, with a maximum error of 11.5% (Micromeritics, ASAP 2400). The fresh SP-115 zeolite without metal coating had an average surface area of 482 m2/g. The unreacted SPlC5M5 had an average surface area of 434 m2/g, with a metal loading of 1.82 wt % Cu and 1.64 w t % Mn. This suggests that the presence of the metal has caused some pore blockage. After single sulfidations at 1000 O F (538
OC), 1250 OF (677 "C), and 1600 O F (871 "C), the surface area decreased to 379, 370, and 237 m2/g, respectively. The loss of surface area under reaction conditions may be due to coalescence of pores of the zeolite or, to a certain extent, to silicic acid formation by reaction of the zeolite with high temperature steam. For a five-cycleset of experiments on sorbent SP2C5M5, the sorbent utilization at breakthrough for each of the first through fifth sulfidations was respectively 20 % , 7 % , 7 % , 13%, and 12%. Sorbent utilization was defined as the ratio of weight of sulfur reacted to total weight of sulfur that was expected to react for complete sulfidation of the metal oxides based on stoichiometry. These percent utilizations were calculated using t h e following assumptions: (1)the H2S breakthrough point was 200 ppmv H2S in the outlet gas. (2) The metal composition of the sorbent was 4.39 w t % Cu and 6.11 wt % Mn, as determined by atomic absorption analysis. (3) The concentration of H2S in the inlet gas was 2000 ppmv. (4) The sulfided metals in the sorbent were present only in the form of CuzS and MnS. The decrease in sorbent utilization after the first sulfidation was due to incomplete regeneration. The increase between 7 % and 12-13% may have been due to random variations. It is possible that the relatively low utilization of the sorbent was caused by ineffective coating of the zeolite pores, i.e., the metals were deposited in several atomic layers on the outer surface of the extrudate rather than inside the pores. Or, even if the pores were effectively coated, there may have been significant resistances to diffusion of H2S into the pores in the zeolite and probably in the small amount of binder (zeolite SP-115 contains Si02 >99% by wt). However, for a three-cycle set of experiments on sorbent SP2C9M1, the sorbent utilization was 66% and 58% for the second and third sulfidations, respectively. Six different samples of the bed material from the fivecycle set performed with SP2C5M5 at 1250 O F (677 OC) and the three-cycle set run with SP2C9M1 at 1250 O F were analyzed using atomic absorption. There was no conclusive evidence for loss of Cu or Mn. Thus, these metals do not appear to undergo appreciable volatilization under these conditions of sulfidation and regeneration. The crush strength of the sorbent extrudates is an important consideration for practical operation. Crush strength was determined by taking the average of the force per unit length of extrudate required to crush 15 extrudates; the range of error for these tests is f20-35 % ,within the first standard deviation. The fresh SP-115 zeolite without metal coating had an average crush strength of 6.58 kg/cm. The crush strength for fresh SP2C5M5 and SP2C9M1 were 29.6 and 11.1 kg/cm, respectively. The metal coating makes the extrudates much stronger; particularly the Mn appears to promote increased resistance to crushing. The average crush strength of six samples from the SP2C5M5 bed, after five cycles, decreased by only 6 %. The average crush strength of five samples from the SP2C9M1 bed, after three cycles, increased by 3% . Therefore, the crush strength of the zeolite was increased by the metal loading and essentially unchanged by reaction. Conclusions For the temperature range of 1000-1600 O F (538-871 "C), the SP-115 zeolite-supported Cu/Mn oxides sorbent Environ. Sci. Technol., Vol. 27. No. 7, 1993
1301
1200
-3rd
12JO0F (677OC)
S u l f . , SPICSMS
U 3 r d Sulf., S P 2 CJMJ
-5
1000
Y
0
0
W
P
5 v)
i
z
8 4
c W 2
I-
0 3
TIME f m i n )
Flgure 10. Hydrogen sulfide breakthrough curves for the third sulfldatlon of sorbents SPlC5M5 and SP2C5M5 at nominal 1250 the effect of the number of metal loadings.
O F
(677 OC):
*3rd Sulf.,SPZC5M5 3rd Sulf., SP2C9MI
0
50
too
I50
200
TIME (min)
Flgure 11. Hydrogen sulfide breakthrough curves for the second and third sulfidatlons of sorbents SP2C5M5 and SP2CQM1at nomlnal 1250 O F (677 OC): the effect of nominal molar ratios of Cu:Mn of 5:5 versus 9:l.
had more total capacity at 1250 O F (677 "C) and 1600 O F (871 O C ) than at 1000 O F (538 "C). Regeneration with a 50 mol % steam/50 mol % air mixture was preferred over air regeneration, due to decreased sulfate formation in the presence of steam. Further studies would be required to optimize regeneration conditions so that minimal sulfate is formed during oxidative regeneration. Appreciable amountsof COS were formedat 1250 O F (677 "C), probably due to reactions catalyzed by the zeolite; however, at 1500 O F (816 "C) the production of COS greatly decreased. The sorbent containing Cu:Mn molar ratios of 9:l had a greater 1302
Envlron. Scl. Technol., Vol. 27, No. 7, 1993
capacity for HzS than those containing equimolar ratios of the metals. Sorbent utilization with repeated sulfidatiodregeneration cycles was low (513 %) for SP2C5M5 but higher (58436%)for SP2C9M1. Surface area of the sorbent decreased with metal loading and with reaction temperatures between 1250 O F (677 "C) and 1600 O F (871 "C). It is uncertain whether the zeolite pores had been effectivelycoated or plugged by the metal acetate solutions during preparation. There was little or no volatilization of the metals at 1250 O F (677 O C ) . The SP-115 zeolite will require further testing to determine whether it can
withstand the highest temperatures (1250-1600 871 "C)for multiple cycles.
OF)
(677-
Acknowledgments
A.T.A. acknowledges the financial support provided by the Oak Ridge Associated Universities. The views expressed in this paper belong to the authors. Literature Cited Kyotani, T.; Kawashima, H.; Tomita, A.; Palmer, A.; Furimsky, E. Fuel 1989,68, 74. Westmoreland, P. R.; Harrison, D. P.Environ. Sei. Technol. 1976, 10, 659. Westmoreland, P. R.; Gibson, J. B.; Harrison, D. P. Environ. Sei. Technol. 1977, 11, 488. Rao, T. R.; Kumar,R. Chem. Eng. Sei. 1982,37,987. Ranade, P. V.; Harrison, D. P. Chem. Eng. Sei. 1981,36, 1079. Grindley,T.; Steinfeld, G. Report DOE/METC/16545-1125, 1981. Focht, G. D.; Sa, L. N.; Ranade, P. V.; Harrison, D. P. Report DOE/MC/21166-2163 (DOE 86016041), 1986. Flytzani-Stephanopoulos,M.; Gavalas, R. G.; Johimurugesan, K.; Lew, S.; Sharma, P. K.; Bagajewicz,M. J.; Patrick, V. Proceedings of the Seventh Annual Gasification and Gas Stream Cleanup Systems Contractors Review Meeting; DOE/METC-87/6079; U.S.Department of Energy, Office of Fossil Energy, Morgantown Energy Technology Center:
Morgantown, WV, 1987; Vol. 2, p 726. (9) Gangwal, S. K.; et al. Bench-Scale Testing of Novel HighTemperature Desulfurization Sorbents. Proceedings of the Eighth Annual Gasification and Gas Stream Cleanup Systems Contractors Review Meeting; DOE/METC-88/ 6092; U.S.Department of Energy, Office of Fossil Energy, MorgantownEnergy TechnologyCenter: Morgantown,WV, 1988; Vol. 1, p 103. (10) Jalan, V.; Desai, M. Proceedings of the Eighth Annual Gasification and Gas Stream Cleanup Systems Contractors Review Meeting. DOE/METC-88/6092;U.S.Department of Energy, Office of Fossil Energy, Morgantown Energy Technology Center: Morgantown, WV, 1988; Vol. 1, p 45. (11) Chemistry of Hot Gas Cleanup in Coal Gasification and Combustion; MERC/SP-78/2; Morgantown Energy Research Center: Morgantown, WV, 1978. (12) Flytzani-Stephanopoulos, M.; Gavalas, R. G.; Johimurugesan, K.; Lew, S.; Sharma, P. K.; Bagajewicz, M. J.; Patrick, V. Final Report for the project DE-FC21-85MC22193,Oct 1987. (13) Desai, M.; Brown, M.; Chamberland, B.; Jalan, V. Prep. Pap. Am. Chem. Soc., Diu. Fuel Chem. 1990,35 (l),87-94. (14) Bissett, L. Proceedings of the Ninth Annual Gasification and Gas Stream Cleanup Systems Contractors Review Meeting; DOE/METC/PETC; U.S.Department of Energy, Office of Fossil Energy, Morgantown Energy Technology Center: Morgantown, WV, 1989.
Received for review August 5, 1992. Revised manuscript received January 4, 1993. Accepted February 11,1993.
Envlron. Sci. Technol., Vol. 27, No. 7, 1993
1303
