170
Ind. Eng. Chem. Res. 1991,30, 170-176
Analysis of a Fixed-Bed Gasifier IGCC Configuration L a r r y A. Bissett* and L a r r y
D.S t r i c k l a n d
United States Department of Energy, Morgantown Energy Technology Center, Morgantown, West Virginia 26507-0880
Integrated coal gasification combined cycle (IGCC) systems offer substantial advantages for power generation. This paper analyzes a proposed configuration in which regeneration gases from a zinc ferrite based, fixed-bed, hot-gas desulfurization system are recycled to a fixed-bed gasifier for capture and disposal of the sulfur by a calcium-containing material (e.g., limestone) added t o the coal. The minimum required sulfur capture in the gasifier for technical feasibility is analyzed and compared to test results from a 1.05-m-i.d., fixed-bed gasifier. Other integration issues involving zinc ferrite regeneration requirements and system operability are also discussed. This study concluded that the proposed system is less attractive than originally thought, primarily due to the low sulfur capture efficiency of a fixed-bed gasifier, the uncertain composition of the bottom ash, and the regeneration characteristics of zinc ferrite in a fixed-bed reactor. Introduction
regeneration gases from a hot-gas desulfurization system are recycled to a gasifier for capture by a calcium-containing material (e.g., lime or limestone) added to the gasifier feedstock. This concept, with its projected ultimate disposal of the coal sulfur as calcium sulfate, was viewed as a potentially attractive solution to the regeneration gas treatment and integration issues. The US. DOE initiated activities in 1986 to further develop and analyze this concept. Fixed-bed desulfurization (as opposed to moving bed) using zinc ferrite was selected on the basis of its more advanced status a t that time, and fixed-bed gasification was selected on the basis of perceived favorable characteristics for power generation. These include high thermal efficiency, good load following capability, stable operation, and product gas temperatures compatible with existing hot-gas desulfurization and control valve technologies. With these process selections, the degree and form of sulfur capture in the gasifier, and various system integration and control aspects were identified as the main technical issues. Process studies, supported by some experimental activities, were used to address these issues and to guide development activities.
Integrated coal gasification combined cycle (IGCC) systems are a promising new technology for power generation. In these systems, the fuel gas produced by reacting coal with air (or oxygen) and steam is burned and expanded through a combustion turbine, and thermal energy is recovered from the turbine exhaust gases by generating steam, which can be used to drive a steam turbine or be injected into the gas turbine to boost its output. To meet environmental emission regulations and protect the gas turbine, sulfur and other contaminant species released during the gasification of coal must be removed from the fuel gas stream. To realize the full benefits of IGCC systems, it is typically desirable to keep the fuel gas stream hot from the gasifier to the gas turbine. This is referred to as hot-gas cleanup in general and hot-gas desulfurization in particular when sulfur species are the primary contaminant being removed. As pointed out by Wieber and Halow (1987), IGCC systems offer superior environmental performance, repowering options, the opportunity for staging capacity additions in small but economical increments, and the potential for future improvements. For these and other Sorbent Chemistry reasons, the United States Department of Energy (DOE) embarked on a program in 1981 to develop simplified Some knowledge of zinc ferrite chemistry and its charIGCC systems. As outlined by Pitrolo and Bechtel(1987), acteristics when used in a fixed-bed reactor is necessary emphasis was placed on the development of hot-gas defor understanding the analysis of the proposed concept. Zinc ferrite is formed by heating an equal molar mixture sulfurization and particulate cleanup, sulfur capture directly in the gasifier, and improved performance of the of zinc oxide (ZnO) and iron oxide (Fe203)to approxipower island. mately 850 "C. The use of finely divided catalytic-grade oxides has given the best reactivity. Formulations are The US. DOE has been developing zinc ferrite (ZnFe204)as an advanced, regenerable (i.e., reusable), hot-gas typically made into extrudates (cylindrical) or pellets desulfurization sorbent since 1980. Sorbent formulations (spherical) with small quantities of inorganic and organic judged to be sufficiently reactive and durable for fixedbinders added for strength and porosity enhancement, and moving-bed reactors have been developed, and acrespectively. ceptable operating parameters and general operating Zinc ferrite possesses the favorable kinetic and capacity characteristics have been determined. When desulfurifeatures of iron oxide while simultaneously retaining the zation sorbents are regenerated, the captured sulfur is superior sulfur removal capability of zinc oxide. Numerous released, typically in the form of sulfur oxides, into the studies have shown that zinc ferrite is regenerable and, in regeneration gas. This regeneration gas must then be a fixed-bed reactor at approximately 600 O C , can reduce further treated to remove the sulfur and convert it to an both hydrogen sulfide and carbonyl sulfide to concentraacceptable form for either disposal as a waste stream or tions less than 1 ppmv. Since these are the major sulfur for sale as a byproduct. Effective integration of regenerspecies produced by coal gasification, their nearly complete ation gas treatment into an IGCC system has been one of removal typically enables the total concentration of all the the main impediments to hot-gas desulfurization techlighter gaseous sulfur species to be reduced to under 20 nology. ppmv. However, zinc ferrite does not effectively remove During sorbent development activities, an IGCC consulfur associated with the heavier tar compounds, and thus figuration was proposed (see Grindley (1988a)) in which the cleaned product gas from a fixed-bed gasifier may still This article not subject to U.S.Copyright. Published 1991 by the American Chemical Society
Ind. Eng. Chem. Res., Vol. 30, No. 1, 1991 171 contain several hundred parts per million by volume total sulfur. Ideally, the sorbent reduction (l),sulfidation or absorption (2), and regeneration (3) reactions proceed as follows: 3ZnFe204+ H2 3Zn0 + 2Fe30, H 2 0 (1)
-
3Zn0
+ 2Fe30, + 9H2S + 2H2
-+
3ZnS
3ZnS
+ 6FeS + 1502
-
+
6FeS
3ZnFe20,
-
+ 2S02 + O2
ZnO
+ SO3
-
1Scad
= 100 Moles
n /,,)-
1 ,>
timestoneor time
{%
S,,,.-193
........
Sulfur
- Laam Product Gas
+ 9S02
2ZnS0,
-
Air. Steam
ZnSO,
Zinc Ferrite Regenerator
\
..,S = 188 Moles Rsgenrratioo Gas I H,O. N,. SO,)
(3)
(4) (5)
Zinc sulfate formation is favored by high pressure, high concentrations of oxygen and sulfur dioxide, and low temperature. The formation of iron sulfates has not been observed in fixed-bed reactors. On the basis of test data in the 1.4-2.1-MPa range, Grindley (1986) suggests that about 15% of the original sulfide profile remains as a zinc sulfate profile after oxidative regeneration. The optimum regenerant inlet temperature appeared to be approximately 550 O C , and to keep the reaction zone peak temperature from exceeding approximately 800 “C due to the highly exothermic reaction, steam was used to dilute the air so that the oxygen concentration in the regenerant gas was less than 4 vol %. If an oxidatively regenerated bed containing zinc sulfate is returned directly to sulfidation service, a sharp exotherm rapidly traverses the bed, potentially damaging the sorbent, product gas heating value drops precipitously, creating flame-out concerns in the downstream combustor, and sulfur species are undesirably released into the product gas stream. This is mostly due to the highly exothermic
/Zinc Ferrite Absorber
n
/‘l,-33r{Fixed. Bed Gasifier
+ l l H 2 0 (2)
Desulfurized Product Gas
Moles
‘lo. 85%
In a fixed-bed desulfurization system, there must be at least two reactors so one can be on-line in a sulfidation mode while the other is off-line being regenerated. Sulfidation is typically performed by passing the sulfur-laden product gas upward through the bed until sulfur breakthrough occurs, which by definition is when the total sulfur concentration in the clean gas equals an acceptable limit from an environmental or equipment perspective. Operation in this manner results in the sorbent being progressively saturated with sulfur from the bottom up. Since the sorbent generally becomes stronger as it sulfides, this places the more durable material on the bottom where crushing forces from the weight of the bed are greatest. A t breakthrough, an S-shaped sulfide profile exists in the bed and the sulfur content of the bed will be something less than the ideal 3 mol of captured sulfur per mole of zinc ferrite. If the bed contains only 2 mol per mole, the sorbent is 67% utilized. As space velocity is increased, the S-shaped profile elongates and utilization decreases. However, since the upper part of the bed remains low in sulfur content and serves as a polishing zone, sulfur removal capability (efficiency) is relatively unaffected and remains high. Regeneration is conducted in the reverse direction to keep the upper part of the bed comparatively sulfur free to preserve its polishing potential. The rate at which oxidative regeneration is carried out in a fixed bed is determined by the rate of air addition. Actual oxidative regeneration is much more complicated than reaction 3. A wave of hydrogen sulfide and elemental sulfur is initially released, sulfur trioxide is also produced, and some of the captured sulfur remains in the sorbent as zinc sulfate, perhaps due to these reactions: 2Zn0
R-
cad
Ash
Sulfur Balances (Steady State)
Sulfur Efficiencies 1%) In. Bed Capturs.?), =
sco.*s“~”o*-SA.*S.., s,
-
Sm.
-s..,
O”er*R.mo”ill.tf.
=
*
S,, zSlxr x 100 x 100
Figure 1. Recycle concept and relationships.
reduction of zinc sulfate to zinc sulfide, during which hydrogen and carbon monoxide are nearly quantitatively reacted to steam and carbon dioxide according to the following ideal reactions (written only for hydrogen): ZnSO, + 4H2 ZnS 4H20 (6) ZnSO,
+ H2
-
-
ZnO
+
+ SO2 + H 2 0
(7)
In practice, Grindley (1988b) has shown that a complex mixture of sulfur compounds rather than just sulfur dioxide is released and that about 85% of the residual sulfate sulfur remains in the bed as zinc sulfide. This translates into 2.25% of the captured sulfur being potentially released into the product gas. To counter this characteristic of the sorbent in a fixed-bed reactor, diluted reducing gas must be passed through the bed in a controlled manner before the bed can be returned to sulfidation service. This is referred to as a reductive regeneration step.
Details of the Concept Figure 1 schematically depicts the proposed configuration and defines two sulfur capture efficiencies based on steady-state sulfur balances around the system and around the gasifier. The overall sulfur removal efficiency is important from the perspective of environmental emission regulations and equals the percentage of coal sulfur that is discharged in the gasifier bottom ash, optimally as CaS0,. However, the sulfur captured in the gasifier bed relative to the total input of sulfur to the gasifier with the coal as well as with the regeneration gas stream is a measure that is more useful for analyzing and characterizing this concept. This parameter is defined as the in-bed sulfur capture efficiency. In a system without recycle, these two efficiencies would be numerically equal. To further illustrate the concept, sulfur flows based on 100 mol of sulfur input with the coal are included in the figure for a system having 95% overall sulfur removal and 33% in-bed capture. With these efficiencies, 1.88 mol of sulfur would be recycled to the gasifier per mole of sulfur in the feed coal. This points out an important aspect of the system; namely, that “low” in-bed sulfur capture efficiencies lead to sulfur recycle ratios above 1 and increase the load on the desulfurization system compared to a configuration without recycle. In the recycle configuration, the steam used to prevent excessive regeneration temperatures replaces all or a portion of the normal steam fed to the gasifier. This “double” usage of the steam and the transfer of regeneration heat to the gasifier are potentially key advantages
172 Ind. Eng. Chem. Res., Vol. 30, No. 1, 1991
of the concept. However, the gasifier steam demand must be less than or equal to the regeneration steam requirement. Since the regeneration steam requirement increases as the sulfur recycle ratio increases, a minimum level or efficiency of sulfur capture must be obtained in the gasifier bed for successful integration and technical feasibility.
Process Studies Analysis of Required Capture Efficiency. The minimum in-bed sulfur capture efficiency necessary before the regeneration steam requirement exceeds the gasifier steam demand is potentially a function of the eight parameters listed in Table I. The variable names given in the table are used in the equations below. Definitions are clarified as necessary in the discussions that follow. The numerical values were used in a design study for a system at 4.14 MPa and served as baseline values for a sensitivity analysis. Steam-coal weight ratios of 1.20 and 1.68 corresponding to gasifier steam-air weight ratios of 0.5 and 0.7 were also considered in the study. The following equations give the relationship between the minimum required gasifier in-bed sulfur capture efficiency and these parameters: C = 3U[l - FATE(1 - FREL)] (8) SOR = 3U(1- FATE) OOR =
1
+ 29U + 9U(FATE) 6
SRR = 3U(FATE)(FREL)
[
(9) (10)
[
(SOR)(SC)(l - N ) D(OOR) 2 . 6 7 5 5 ( F 8 C ) ( D(O0R) F S T ) ] ~(12)
(1 - N)(SRR)(l - FST - N ) -
(FR)(C)N
SO + SR = C
Parameter: Gasifier Steam-CoalWeight Ratio
Sulfur Content of Coal (weight K)
Figure 2. Effect of sulfur content of coal on in-bed capture requirement. Table I. Parameters Affecting In-Bed Sulfur Capture Requirement and Baseline Values Used for This Study baseline value parameter (name) 1.44 gasifier steam-coal weight ratio (SC) 0.0212 weight fraction of sulfur in coal (FSC) 5 oxidative regeneration steam-air molar ratio (D) 0.064 98 fractional conversion of coal sulfur to tar sulfur (FST) sulfate formation fraction (FATE) 0.15 0.1547 reductive regeneration cycle fraction (FR) fractional utilization of zinc ferrite ( W 0.6667 0.15 sulfate release fraction (FREL)
(11)
SO = 2.6755(FSC)(1 - N ) +
SR = (1 - N ) +
60
(14) where, in addition to the parameters defined in Table I, N = minimum required gasifier in-bed sulfur capture efficiency (fractional), C = zinc ferrite (ZF) sulfidation capacity (kmol of S/kmol of ZF), SOR = sulfur released during oxidative regeneration (kmol of S/kmol of ZF), OOR = oxygen used during oxidative regeneration (kmol of 02/kmol of ZF), SRR = sulfur released during reductive regeneration (kmol of S/kmol of ZF), SO = sulfur captured in the absorber during oxidative regeneration (kmol of S/kmol of ZF), and SR = sulfur captured in the absorber during reductive regeneration (kmol of S/kmol of ZF). These equations, which must be solved iteratively for N , were derived from material balances and zinc ferrite stoichiometry with the following three assumptions: (1) the gasifier in-bed capture efficiency is the same during both the oxidative and reductive regeneration cycles, (2) the total time necessary to complete both oxidative and reductive regeneration cycles of the off-line zinc ferrite bed is equal to the sulfidation time of the on-line bed when the oxidative regeneration is carried out at the maximum rate set by the gasifier steam demand, and (3) the capability of zinc ferrite to remove sulfur species to very low levels
prior to sulfur breakthrough and sorbent utilization at breakthrough are independent of space velocity over a limited range. With regard to the first assumption, it is likely that in-bed sulfur capture will depend on the rate of sulfur recycle. Since more sulfur is released from zinc ferrite during oxidative regeneration compared to reductive regeneration, the gasifier in-bed sulfur capture during recycle of the oxidative regeneration gases will likely be greater on an hourly basis, but less on a percentage basis. Thus, although impossible to quantify without a sufficient database, the first assumption is probably optimistic and leads to an underprediction of the minimum-required capture efficiency. The same is true of the second assumption since, for safety reasons, additional time will be required to purge the zinc ferrite reactors between cycles, which in practice will shorten the available sulfidation time and increase the in-bed capture requirement. With regard to the third assumption, the absorber space velocity varied less than 7% during the entire sulfidation cycle at the baseline condition used for the analysis. Experimentation (see Underkoffler (1986)) has shown that sulfur removal remains very good over a much broader range. In addition, the derivation of these equations showed that the in-bed capture requirement is independent of the space velocity used in the zinc ferrite reactors (as long as cycle times are kept reasonable). Therefore, in light of these considerations, the impact of the third assumption on the calculated in-bed capture requirement is viewed as being negligible. As shown by Figure 2, the required capture efficiency is very dependent on and progressively increases with the sulfur content of the coal. For a coal with 2-3% sulfur, the required in-bed capture efficiency is in the 25-40% range. The gasifier steam-coal ratio, shown by the plot parameter, also affects the required capture since it impacts the time required to complete oxidative regeneration. The explanation for this is as follows. As gasifier steam
Ind. Eng. Chem. Res., Vol. 30, No. 1, 1991 173 40
I-
Parameter: Gasifier Steam-Coal Weight Ratio
I
c!
1 20
'
4.0
95-
I
,
I
Parameter: Gasifier Steam-Coal Weight Ratio
1
f
4.5 5.0 5.5 Oxidative Regeneration Steam-Air Molar Ratio
6.0
Figure 3. Effect of regenerant steam-air ratio on in-bed capture requirement. Parameter: Gasifier Steam-Coal Weight Ratio u)
Parameter: Gasifier Steam-Coal Weight Ratio
I
f
;
20 0.05 201 0.00
I
I
1
0.02
0.04
0.06
I
0.08
0.10
Sulfur Conversion to Tar (weight fraction)
Figure 4. Effect of sulfur conversion to tar on in-bed capture requirement.
demand increases, higher steam flow rates can be used during oxidative regeneration. For a constant steam-air regenerant ratio, this translates into higher regeneration air flow rates and less time required to complete regeneration. This in turn means that faster sulfur breakthrough in the on-line absorber can be tolerated. For a fixed volumetric input (Le., space velocity), this translates into a higher allowable concentration of sulfur in the entering gas and, therefore, a lower required in-bed sulfur capture efficiency. This relationship will hold up until velocity or pressure drop limitations are encountered. As Figure 3 shows, the regenerant steam-air ratio has a similar effect for the same reason. Even at the baseline value of 5 mol of steam per mole of air in the regenerant, the calculated bulk temperature rise in the regenerator is approximately 250 O C . Since local sorbent temperatures could be higher, some sintering and loss of sorbent reactivity may occur during each cycle, and it seems likely that a steam-air molar ratio of 5 will be a practical minimum for a fixed-bed desulfurizer. Experimental evidence to date indicates that zinc ferrite does not effectively remove sulfur associated with tar compounds. Unless provisions are made to crack the sulfur out of the tar prior to hot-gas desulfurization, this sulfur, which may be up to about 8% of the coal sulfur in a fixed-bed gasifier, leaves the system as an emission. Although detrimental to overall system sulfur removal efficiency, this eases in-bed sulfur capture requirements for integration since less sulfur must be captured by the sorbent, and as shown by Figure 4, this is equivalent to feeding a lower sulfur coal. A t the higher pressures of advanced IGCC systems, tar yields will be reduced and there may be a
0.10
0.15
0.20
0.25
Reductive Regeneration Cycle Fraction
Figure 6. Effect of reductive regeneration time on in-bed capture requirement.
greater tendency for desulfurization sorbents to crack the tars, but to what extent is presently unknown. Sulfate formation during oxidative regeneration increases the amount of oxygen required to complete the regeneration. For a constant steam-air regenerant ratio and with steam at the maximum flow rate consistent with the gasifier demand, regeneration will take longer to complete. This means that the absorber must also stay on-line longer before sulfur breakthrough occurs. For a fixed volumetric input, this translates into a lower allowable concentration of sulfur in the gas entering the absorber. Thus, as shown by Figure 5, a higher in-bed sulfur capture efficiency is needed. Here, the sulfate formation fraction is defined as the proportion of captured sulfur in the zinc ferrite that remains as sulfate sulfur. If only zinc sulfate forms, the highest sulfate formation fraction for stoichiometric zinc ferrite is 0.33. This also implies that, for the baseline formation fraction of 0.15, approximately half of the zinc sulfide regenerates to zinc oxide (reaction 31, and the balance resorts to zinc sulfate (reactions 4 and 5 ) . The maximum rate (and hence minimum time) for acceptable reductive regeneration is related to reactivity and heat-transfer effects. Thus, the amount of residual sulfate present and the reactor design parameters are both important. As shown by Figure 6, the required in-bed sulfur capture efficiency increases as the duration of reductive regeneration increases. This means that the technical viability of the concept improves as both the amount of sulfate formed and the time required for reductive regeneration are minimized. For this analysis, the reductive regeneration cycle fraction is defined as the proportion of total cycle time required to conduct reductive regeneration. About 15% of the cycle time was left for completing reductive regeneration in this design study when oxidative
174 Ind. Eng. Chem. Res., Vol. 30, No. 1, 1991
regeneration was carried out in the minimum time. With 50% steam dilution of the reducing gas as typically used in lab-scale tests, the resulting space velocity was about 2II2 times higher than what has been demonstrated by Grindley (1988b) as being acceptable (but not necessarily limiting) at about 0.3 MPa. Thus, reductive regeneration at higher pressures and space velocities was viewed as a technical uncertainty that required further experimental verification. Methods for conducting reductive regeneration are discussed later in this paper. The sensitivity analysis showed that the other two parameters, sorbent utilization and the fraction of sulfate sulfur released during reductive regeneration, had very little effect on required in-bed capture. Both are mathematically similar to space velocity and affect sorbent capacity and hence cycle time, but not in-bed capture requirements. The feasibility of the concept can also be expressed in terms of the maximum amount of sulfur that can be recycled per unit of steam fed to the gasifier. The following equation gives this relationship: FSC(1- FST - N ) SRSR = (15) N(SC) where SRSR = the maximum recycle sulfur to gasifier steam weight ratio (kilograms/ kilogram). Although not obvious from this equation because of the convoluted dependency of N on other parameters (eqs 8-14), the ratio is independent of the sulfur content of the coal and the gasifier steam-coal ratio. Of the remaining parameters listed in Table I, only three have a significant effect: the oxidative regeneration steam-air ratio, the sulfate formation fraction, and the reductive regeneration cycle fraction. Decreases in any of these parameters cause a fairly significant and desirable increase in the maximum recycle to gasifier steam ratio. This facilitates system integration. Figure 7 illustrates this for two of the parameters. The relationship with sulfate sulfur is similar. Experimental Support. In 1986 and 1987, the Morgantown Energy Technology Center (METC) of the United States DOE conducted concept feasibility tests with a 1.05-m-i.d., fixed-bed gasifier a t nominal pressures of 0.8 and 1.5 MPa. Seventeen test periods employing various mixtures of bituminous coals, limestones, and hydrated lime were completed. Sulfur dioxide, typically at a rate consistent with the steam integration limit discussed above, was injected into the bottom of the gasifier to simulate the recycle of regeneration gases from an external desulfurization system during 10 of the test periods. The primary purpose of the tests was to determine if sufficient in-bed sulfur capture could be achieved with respect to process steam integration. Secondary objectives included determinations of the forms of sulfur in the bottom ash and system response characteristics. Details and results of these tests have been presented by Reuther et al. (1987), Bissett et al. (1988), and Bissett and Strickland (1989). The average in-bed sulfur capture efficiency for the 8 test periods with representative sulfur dioxide injection rates was 33%. The averaged minimum in-bed capture efficiency required to technically demonstrate the recycle concept for these same periods was 38%. Thus, insufficient sulfur capture was achieved relative to the process steam integration limit. However, the results were encouraging and did seem to indicate that success could likely be achieved with further optimization, such as through the use of better limestones, higher calcium dosages, or different operating conditions. The tests also suggested that good sulfur capture in the bottom of the gasifier is a key requirement for technical success since some of the sulfur
.-2
.os Parameter: Reductive Regeneration Cycle Fraction
2.-
W
a P
5
.E
4.0
I
I
I
4.5
5.0
5.5
J
Oxidative Regeneration Steam-Air Molar Ratio
Figure 7. Effect of regenerant steam-air ratio on maximum sulfur recycle.
captured in the upper portion of the gasifier is subsequently released while passing downward through the hotter combustion zone. The nature of the captured sulfur in the gasifier bottom ash is a key waste-disposal issue. For the bottom ash to be considered nonhazardous in this regard, the releasable sulfide level, as determined by standard Resource Conservation and Recovery Act (RCRA) methods, must be less than the current Environmental Protection Agency (EPA) action level of 500 ppm. Since some sulfur is captured as calcium sulfide (Cas) in the reducing zone of the gasifier, it must be further oxidized to calcium sulfite (CaSO,) or calcium sulfate (CaS04) as it passes through the combustion zone. Since this reaction becomes severely diffusion-limited as a relatively nonporous oxidized outer layer forms, this could result in the persistence of an unreacted calcium sulfide core. During the feasibility testing, analyses using the RCRA method consistently showed that nearly all of the bottom ash samples were well below the EPA action level. However, further investigations of ash chemistry were conducted after the test program was completed. During these subsequent investigations, the amount of releasable sulfides in a bottom ash sample was found to be 10 times greater than what was determined by the RCRA method when the strength of the acid used to contact the sample was increased and when a more rigorous procedure for the analysis of the evolved hydrogen sulfide was adopted. Apparently, the high buffering capacity of ash due to the addition of calcium-based sorbents to the coal had resulted in only the partial release of the sulfide present when the RCRA method, which employs a weaker acid, was used. The significance of this finding is that it demonstrates that an environmentally benign bottom ash cannot be guaranteed and that further ash processing may be required when in-bed sulfur capture and sulfur recycle are employed. Reductive Regeneration Methods. In addition to affecting the in-bed capture requirements, the need for controlled, reductive regeneration of zinc ferrite when used in a fixed-bed reactor introduces other complications. Four approaches were evaluated in the design study and are presented below. Figure 8 illustrates a method in which a portion of the "clean" (i.e., comparatively sulfur-free) gas from the on-line absorber is used to carry out the reductive regeneration. This has the advantages of maintaining total steam integration with the gasifier and keeping the plant mass output relatively constant. However, as the absorber approaches sulfur breakthrough, a situation could arise where there is insufficient time to complete the regeneration of the off-line reactor. This situation would be further aggravated
Ind. Eng. Chem. Res., Vol. 30, No. 1, 1991 175 Product Gas
Product Gas
1
Gasifier
t -_
1 1 I
Absorber
lu
I I t
Regenerator
1 1
I I
I
-
Figure 9. Reductive regeneration using product gas with recycle to absorber. Pr luct IS
by any sulfur species “trickling” out of the absorber due, for example, to impending sulfur breakthrough, channeling, or sorbent deterioration. Since any sulfur escaping the absorber would tend to be captured in the top of the regenerating bed, the sulfur capture capacity and polishing potential would both decrease during the subsequent sulfidation cycle. This would contribute to a progressively worsening condition and hinder efforts to get the system resynchronized with sufficient slack times in the cycles. This could possibly lead to a “process tailspin’! from which a relaxation of emission standards, reduced output, or shutdown might be necessary for recovery. Another concern with this approach is the fate of the tar in the “clean’! gas used for regeneration. Depending on temperatures, this tar can potentially condense in lines, on heat exchanger surfaces, or with the gasifier bottom ash, transforming the ash into a hazardous material. Finally, this approach requires some way to boost the pressure of the “clean” gas to overcome the system pressure drop for injection of the regeneration gases into the gasifier. If ejectors (as depicted by “E” in the figure) are used, pressure drops in particulate removal equipment and/or the absorber may begin to dictate cycle frequency rather than in-bed sulfur capture and sorbent performance. In a variation of this approach, shown in Figure 9, the tar concerns are eliminated and the pressure drop problem is lessened by recycling the reductive regeneration gases to the inlet of the on-line absorber rather than back to the gasifier. In concept, the recycle of the sulfur to the gasifier would then be completed during the following oxidative regeneration cycle. Since the dilution steam needed for reductive regeneration is not directed back to the gasifier, there is not total steam integration with the gasifier and this configuration would likely have higher steam requirements and result in greater steam dilution of the product gas. Also, the recycle would tend to increase the absorber space velocity in the latter stages of sulfidation, which would likely accelerate sulfur breakthrough. Thus, this concept would exacerbate any tendency for process tailspin as regeneration cycle times approach sulfidation time. Figure 10 represents an approach where the regeneration gases are injected downstream into the product gas rather than being recycled. This completely eliminates the pressure boost problem and still allows recycle of most of the sulfur to the gasifier during oxidative regeneration. The penalities for this approach are higher steam usage and lower overall system sulfur removal efficiency. This approach could still achieve sufficient sulfur removal to meet the requirements of the New Source Performance Standards (NSPS) if the amount of sulfur associated with tars and the residual sulfate content of the zinc ferrite are low enough. Even though this configuration gave a pro-
Steam
Absorber
Gasifier
Regenerator
t Figure 10. Reductive regeneration using product gas with downstream injection instead of recycle. Product Gas
I
Air
I
LS J I Air Natural
Figure 11. Reductive regeneration using partially combusted natural gas with recycle to gasifier.
jected overall sulfur removal of 8870, which is slightly better than the NSPS requirement, it was not seriously considered in the design study because it did not demonstrate the full potential of this new technology. A fourth approach, illustrated in Figure 11, is the use of partially combusted natural gas to supply the reducing gas needed for reductive regeneration. This approach maintains total sulfur recycle and steam integration with the gasifier, decouples the regenerator from the absorber and thus eliminates pressure boost, sulfur trickle, and tar carry through concerns, and should permit more flexibility in managing reductive regeneration cycle times. The natural gas used is heat-integrated in the process and ultimately adds to the mass flow and energy output of the gas turbine. For the design study, the mass output of the plant was 9% higher during reductive regeneration and the natural gas input energy was about 2% of the coal input energy. If necessary, the mass output of the plant could be lowered during reductive regeneration by partial turndown of the gasifier. The major disadvantages of this approach are that the process would no longer be exclusively coal-fired and sorbent durability could worsen if soot formation occurs during partial combustion of the natural gas. However, relative to the other approaches, it was concluded that the potential benefits of this scheme outweighed the disadvantages and should be seriously considered if further development activities were pursued. Other Integration Considerations. The study also
176 Ind. Eng. Chem. Res., Vol. 30, No. 1, 1991
raised concerns with respect to system operability, design issues, and the possible detrimental behavior of trace species. One of the more favorable characteristics that a fixedbed gasifier offers to power generation is the ability to turn down all the way to a banked, or stand-by, condition. Normally, as long as combustion zone bed temperatures stay high enough for ignition, a banked gasifier can be easily and quickly returned to service. However, the METC feasibility tests showed that, when a calcium-based sorbent is added to the feed coal, the lower portion of the bed tends to "set up" like hardening cement during stand-by. This resulted in considerable ash discharge and bed stirring problems during restart. Thus, from this viewpoint, in-bed sulfur capture compromises a very desirable gasifier characteristic. Another hallmark of fixed-bed gasification is stable operation. However, in the proposed configuration, the fixed-bed desulfurization system is never truly at steadystate conditions because temperatures, flow rates, compositions, and sulfur release vary inherently during a particular regeneration cycle and are notably different between the oxidative and reductive regeneration cycles. The recycle loop provides a feedback to the gasifier and propagates these transients to the rest of the system. The sulfur transients are effectively dampened by the sink capacity of the gasifier and the high removal efficiency of the absorber; however, the product gas composition and flow rate to the power island are always in a state of flux. Therefore, the cyclic nature of this configuration reduces the stable operating characteristic typical of other fixedbed gasifier IGCC systems. The addition of calcium-based sorbents to the gasifier and the recycle of regeneration gases will impact the design of the gasifier and increase its cost. It is important to note that the METC feasibility tests used a water-cooled stirrer that rotated and traversed vertically through the coal bed over a span of about 0.15-1.22 m above the ash grate. Although no firm conclusions could be drawn from the test data, there is the possibility that the improved gas-solid contacting that occurs with deep-bed stirring may be beneficial, and perhaps even crucial, for obtaining better in-bed sulfur capture and low sulfide levels in the bottom ash. Although deep-bed stirring is an undesirable complexity for a commercial unit, it may be necessary for this reason as well as to help overcome the restart difficulties discussed above. The stirrer and drive unit will also need to be stronger since the addition of sorbents seemed to make the bed stiffer and harder to stir. In addition, other costs will increase since extensive corrosion to the bottom of the gasifier from the injection of sulfur dioxide and additional ash lock-hopper valve problems due to sorbent addition to the feed coal were experienced in the METC tests. Finally, in any recycle system, various physical or chemical species can potentially be concentrated to levels that become detrimental to the process or the equipment. In the present concept, trace species in the coal may possibly accumulate in the recycle loop by being captured by zinc ferrite during the sulfidation cycle and subsequently released during regeneration. Experimental results from sidestream tests at METC have shown that significant amounts of mercury and chloride removed from fixed-bed gasifier product gas are released during oxidative regeneration. It is likely that other species will exhibit similar behavior, although definitive results are not available. These species must leave the system in the product gas, in the spent zinc ferrite material, or preferably in the gasifier bottom ash. Gasifier operating conditions,
such as an upper limit on the steam-air blast temperature, may have to be adjusted to encourage capture and removal of these species in the ash. Concentrations beyond presently unknown limits could adversely affect the chemical and physical durability of the zinc ferrite, and accumulations in the sorbent itself could lead to the spent material being classified as hazardous.
Conclusions On the basis of the findings of this study, the proposed IGCC configuration does not appear to be as attractive as originally thought. The main difficulties arise due to the low sulfur capture efficiency of a fixed-bed gasifier and the need for reductive regeneration due to sulfate formation when zinc ferrite is used in a fixed-bed reactor. Both aggravate the steam integration constraint of this configuration. In addition, the low in-bed sulfur capture efficiency increases the load on the hot-gas desulfurization system, and the need for reductive regeneration contributes to the cyclic nature of the process, which complicates system operability and control. Although it is likely that better in-bed sulfur capture could be obtained with further development, certain commercialization obstacles would still be present, such as potentially unacceptable sulfide levels in the gasifier bottom ash. Therefore, although the United States DOE has not precluded recycle schemes, the near-term research plans are focused more on making improvements to the hot-gas desulfurization system by developing fluid-bed technology. It is presently felt that a fluid-bed desulfurization system will provide the potential for steady-state operation, eliminate the need for reductive regeneration, improve temperature control and heat management, and possibly ease steam integration limits if a recycle concept is pursued. Registry No. ZnFe204,12063-19-3.
Literature Cited Bissett, L. A.; Strickland, L. D. Aspects of Fixed-Bed Gasification/Fixed-Bed Zinc Ferrite Integration. Morgantown Energy Technol. Cent. [Rep.]; DOE/METC ( U S Department of Energy), 1989; DOE/METC-89/6107, Vol. 1 (DE89011706),pp 68-86. Bissett, L. A.; Reuther, R. B.; Strickland, L. D. Sulfur Capture in a Fixed-Bed Gasifier. Morgantown Energy Technol. Cent. [Rep.]; DOE/METC ( U S . Department of Energy), 1988; DOE/METC88/6092, Vol. 1 (DE88010253), pp 270-288. Grindley, T. The Effect of Temperature and Pressure on the Regeneration of Zinc Ferrite Desulfurization Sorbents. Morgantown Energy Technol. Cent. [Rep.], DOE/METC (US. Department of Energy), 1986; DOE/METC-86/6042 (DE86001088),pp 190-212. Grindley, T. Method for the Desulfurization of Hot Product Gases From a Coal Gasifier. US. Patent 4,769,045, 1988a. Grindley, T. Laboratory-Scale Study of Reductive Regeneration of Zinc Ferrite Sorbents. Morgantown Energy Technol. Cent. [Rep.]; DOE/METC (US. Department of Energy), 1988b DOE/METC88/6092, Vol. 1 (DE88010253); pp 58-82. Pitrolo, A. A.; Bechtel, T. F. Simplified IGCC: Coal's 'Adam Smith" Response to a Changing World. EPRI Report AP-6007-SR, 1988 p 5-1.
Reuther, R. B.; Bissett, L. A,; Strickland, L. D. A Process Scheme for HZS Removal From a Fixed-Bed Gasifier Gas Stream. Presented at the AIChE Annual Meeting, New York, 1987; Session 114g. Underkoffler, V. S. Summary and Assessment of METC Zinc Ferrite Hot Cas Desulfurization Test Program, Final Report. Morgantown Energy Technol. Cent. [Rep.]; DOE/METC (US. Department of' Energy), 1986; DOE/MC/21098-2247, Vol. 1 (DE87001073). Wieber, P. R.; Halow, J. S. Advanced IGCC Power Systems for the United States. Energy Prog. 1987, 7 (21, 119-125. Receiued for review January 2, 1990 Reuised manuscript received July 10, 1990 Accepted July 24, 1990