5676
Ind. Eng. Chem. Res. 2002, 41, 5676-5685
New Zno-Based Regenerable Sulfur Sorbents for Fluid-Bed/ Transport Reactor Applications Rachid B. Slimane* and Brett E. Williams† Gas Technology Institute, 1700 S. Mount Prospect Road, Des Plaines, Illinois 60018-1804
A new sorbent synthesis technique has been developed at GTI, based on sol-gel processing of inorganic and organic precursors, for the preparation of ZnO-based regenerable sulfur sorbents with unique properties that could not be attained with “conventional” sorbent preparation methods such as coprecipitation or solid-oxide mixing followed by extrusion, granulation, or spray drying. Unlike these techniques, which require very high calcination temperatures to impart physical strength to sorbents, sol-gel processing offers the unique capability of producing sintering at exceptionally low temperatures. As a result, materials with high surface areas, small pore sizes, and very high resistance to attrition can be synthesized. Sorbents prepared with this new technique demonstrated attrition indices that are well below the stringent requirement of the transport reactor of ≈4% and effective capacity for sulfur absorption exceeding 8 g of S/100 g of sorbent. In addition, these sorbents demonstrated regenerability at temperatures that are lower than those required by typical zinc titanates. On the basis of chemical analysis, physical characterization, attrition resistance determination, and evaluation of chemical reactivity and regenerability, the exceptional characteristics of this new class of zinc-based sorbents are demonstrated in this paper. Introduction Zinc-based sorbents have emerged as the leading candidates for the removal of hydrogen sulfide (H2S) from coal-derived fuel gases at high temperatures and pressures in integrated gasification combined cycle (IGCC) processes. These sorbents have resulted from numerous investigations, mainly at temperatures g500 °C. Significant advances have been made; however, for desulfurization in fluid-bed and/or transport reactors, there are still some concerns regarding the ability of these sorbents to resist attrition because of the mechanical forces as well as the chemical transformations they undergo in such applications. These sorbents have generally been prepared in bulk form using “conventional” techniques, such as solid-oxide mixing or coprecipitation followed by extrusion, granulation, or spray drying, with physical strength imparted to them through high-temperature treatment, commonly referred to as calcination or induration. Because of sintering effects, the desirable sorbent characteristics, such as surface area, porosity, and pore size distribution, are diminished with increasing calcination temperature, adversely affecting the reactivity of the sorbent. Moreover, because the reactivity of a sorbent undergoes an Arrhenius-type decrease with decreasing temperature, this trade-off between sorbent mechanical strength and chemical reactivity represents more of a concern in the moderate temperature range of 343-538 °C that is currently of industrial interest. Over 100 zinc-based sorbents were prepared and evaluated in a recent study at GTI for the development of attrition-resistant sorbents that are sufficiently reactive in the moderate temperature range.1-4 Figures 1-3
Figure 1. H2S breakthrough curves for the IGTSS-314B sorbent.
Figure 2. H2S breakthrough curves for the IGTSS-325A sorbent.
* To whom correspondence should be addressed. Tel: (847) 768-0606. Fax: (847) 768-0600. E-mail: rachid.slimane@ gastechnology.org. † Currently with Hatco Corporation in Sturgeon Bay, WI.
report the results obtained with two selected zinc-based sorbents as well as those obtained with a commercial zinc titanate sorbent (UCI-4169) manufactured by Su¨d
10.1021/ie011019t CCC: $22.00 © 2002 American Chemical Society Published on Web 10/22/2002
Ind. Eng. Chem. Res., Vol. 41, No. 23, 2002 5677
Figure 3. Desulfurization performance of baseline UCI-4169 zinc titanate sorbent.
Chemie (formerly United Catalyst, Inc.) Although these sorbents exhibited modest reactivities toward H2S and appeared to require activation,4 their attrition resistance properties were not sufficiently high to meet stringent criteria that have been set for the transport reactor in the Pin˜on Pine Clean Coal Demonstration Project.5,6 Attrition jet indexes (AJI), as determined by the ASTM D 5757-95 method, were found to be in the range of 75-85% for these zinc-based sorbents.4 To improve chemical reactivity in the moderate temperature range, research at GTI investigated alternative support materials for zinc oxide (ZnO). The rationale for this approach was that the addition of titania (TiO2) to form zinc titanate might not be the best option to pursue because it was used in the past to reduce the vapor pressure of zinc, thereby improving its stability at temperatures >500 °C. The compounding of ZnO, however, is inevitably accompanied by reduction in chemical reactivity. The results obtained indicated nontitania sorbents based on ZnO showed no improvement over zinc titanate. Alternative sorbents, based on the oxides of iron, copper, and manganese, were also pursued at GTI to determine the best material for striking an acceptable balance between chemical reactivity and attrition resistance. Some success was achieved with copper and manganese, but not with iron.4 Manganese-based sorbents combine the advantages of high sulfur capacity and high reactivity in the moderate temperature range, without any requirement for sorbent preconditioning or activation. One leading manganese-based sorbent developed in this study (i.e., IGTSS-057) was found to achieve an effective sulfur capacity of over 20 g of S/100 g of sorbent at 450 °C. However, a temperature of at least 750 °C was required for oxidative regeneration of this manganese-based formulation. This could not be readily accommodated by existing desulfurization systems requiring regeneration ignition temperatures of about 550 °C. Further research on alternative regeneration approaches was recommended to reduce the disparity between temperatures for sulfidation and regeneration for manganesebased sorbents. Sorbents based on copper oxide were found to possess the best combination of high attrition resistance, chemical reactivity, sulfur removal efficiency, and prebreakthrough conversion in the moderate temperature range of 343-538 °C.7 One of these sorbents (i.e., IGTSS-179) was shown to have excellent H2S removal efficiency and an effective sulfur capacity approximating 7 g of sulfur/100 g of sorbent at 450 °C, as shown in
Figure 4. H2S breakthrough curves for the IGTSS-179 sorbent.
Figure 4, which corresponds to a sorbent conversion of ≈70%. This sorbent formulation was also found to have an attrition loss that is over 8 times lower than that of UCI-4169 zinc titanate sorbent. Moreover, despite its significantly high attrition resistance, this sorbent did not require an activation step to enhance its chemical reactivity in the moderate temperature range, unlike the majority of zinc-based sorbents. Unfortunately, despite the excellent attributes of the Cu-based sorbents, the leading IGTSS-179 formulation could not be evaluated according to DOE/NETL’s Test Protocol for qualification of candidate sorbents for demonstration.8 It proved very difficult to sustain good and uniform fluidization with Cu-based sorbents in the reducing fuel gas environment during the desulfurization step in a fluidized-bed reactor. Very strong attractive interparticle forces appeared to develop, resulting in sluggishness and ultimately complete defluidization of the sorbent bed. Several “remedies” were attempted to overcome this defluidization phenomenon; however, none provided a long-term solution.9 Given the above results, it became evident that, for desulfurization applications relying on fluid-bed and/ or transport reactor technology at moderate temperature, “conventional” sorbent preparation techniques, such as coprecipitation or solid oxide mixing followed by extrusion, granulation, or spray drying, may not be suitable for preparing sorbents with attrition resistance, chemical reactivity, and fluidization properties that meet the stringent requirements of these reactors. An alternative sorbent preparation approach was deemed necessary. In addition, because sorbents based on ZnO are best suited for desulfurization, particularly in the moderate temperature range of interest, significant efforts were devoted to identifying a suitable approach for the preparation of effective Zn-based sorbents. These efforts led to the development of a proprietary technique based on sol-gel processing of inorganic and organic precursors. Sol-gel processing is defined broadly as the preparation of ceramic materials by preparation of a sol, gelation of the sol, and removal of the solvent.10 The sol is a colloidal suspension of solid particles in a liquid and may be produced from inorganic or organic precursors. For example, common precursors for aluminum oxide include inorganic salts such as Al(NO3)3 and organic compounds such as Al(OC4H9)3. The latter is an example of an alkoxide, the class of precursors widely used in sol-gel research because they react readily with
5678
Ind. Eng. Chem. Res., Vol. 41, No. 23, 2002
water (i.e., hydrolysis reaction). Gels have a huge interfacial area, typically 300-1000 m2/g, and this enormous area serves as a driving force to bring about sintering at exceptionally low temperatures, compared to ordinary ceramic materials. Therefore, a considerable amount of the extraordinary properties (high surface areas and small pore sizes) of unfired gels are retained following calcination, offering the possibility of using the gel as a substrate for chemical reactions. These properties, characteristic of inorganic gels as a result of lowtemperature processing, are unattainable by “conventional” sorbent preparation techniques and, to the best knowledge of the authors, have not been exploited in hot gas desulfurization applications. These characteristics were exploited to synthesize regenerable ZnOTiO2 (i.e., zinc titanate) sorbents with the highly desirable combination of excellent microstructural control (high surface areas and small pore sizes) and very high attrition resistance that well exceeds the stringent requirement of the transport reactor desulfurization application. Experimental Section Sorbent Preparation. The original sorbent preparation technique relies on sol-gel processing, where a stable sol is first prepared from a titania precursor (titanium(IV) isopropoxide (C12H28O4Ti, Alpha Aesar, 97% purity)), to which the required amount of a zinc precursor (zinc nitrate hexahydrate (N2O6Zn-6H2O, ACROS, 98% purity)) is then added to bring the zinc oxide content to the desired level. The resulting solution is well-mixed to disperse the reactive oxide, dried, and finally calcined at the desired temperature (450 °C) for a predetermined period of time (≈4 h). Granules from the calcined material are then produced in the desired size range of 40-425 µm, by grinding and sieving, for evaluation in the attrition resistance measurement unit and in the ambient pressure packed-bed reactor. This novel sorbent synthesis technique has been developed further and a number of preparation steps minimized to reduce processing times and sorbent cost. In addition, we worked closely with a commercial sorbent manufacturer, which led to further simplification of the sorbent preparation procedure, without compromise of the desirable sorbent properties obtained with the original procedure. The nominal ZnO content of the sorbents prepared using the original procedure ranged from 10 to 40 wt %. The test work reported in this paper is directed only to formulations IGTSS-353 and IGTSS-354, which were prepared using the original sol-gel procedure, to illustrate the unique characteristics of this new class of zinc-based sorbents. Chemical and Physical Characterization. Fresh as well as reacted samples from selected sorbents were subjected to chemical analysis and physical characterization, including bulk density, particle (Hg) density, Hg pore volume, porosity, surface area, and median pore diameter, according to ASTM procedures. Chemical analyses on the metal oxide sorbent samples were performed as follows: Samples for metal analysis were prepared using acid digestions and/or sodium tetraborate fusion and were analyzed using a Thermo Jarrell Ash Atomscan 25 inductively coupled plasma atomic emission spectrophotometer (ICP/AES). Samples for sulfide sulfur and sulfate sulfur analysis were aciddigested, and in the case of the sulfide analysis, a closed system was used where the evolved gases were collected
Figure 5. Schematic diagram of the three-hole attrition resistance measurement unit.
in an alkaline peroxide sorbent. The resulting solutions were analyzed by ICP for sulfur. Attrition Resistance Determination. IGTSS-353 and IGTSS-354 sorbent formulations were evaluated for their attrition resistance properties in accordance with the ASTM D5757-95 method.11 Their performance was compared against those of selected sorbents previously prepared using “conventional” techniques and that of a leading candidate zinc titanate sorbent (i.e., EX-SO3) for the Pin˜on Pine Demonstration Project. D 5757-95 is a standard ASTM test method for the determination of the relative attrition characteristics of powdered catalysts by air jets. This test method is capable of providing reliable information concerning the ability of a powdered material to resist particle size reduction (i.e., attrition) during use in a fluidized environment. Strictly speaking, this method is applicable to spherically or irregularly shaped particles ranging in size between 10 and 180 µm and having skeletal densities between 2.4 and 3.0 g/cm.3 Nevertheless, the information it provides can be particularly valuable for research and development efforts in the area of sorbent development for fluidized-bed as well as transport reactor applications, where sorbent durability during multicycling is crucial. An air jet attrition unit was constructed at GTI according to the ASTM method specifications. As indicated on the schematic diagram shown in Figure 5, this apparatus consists of an attrition tube, a settling chamber, a fines collection assembly, and a circular orifice plate containing three holes and is attached to the bottom of the vertical attrition tube within an air delivery manifold. During a typical test, a representative dry sample of the granular material (≈50 g) is subjected to attrition by means of three high-velocity jets of air. The fines generated during this process are continuously removed from the attrition zone by elutriation into the collection assembly. The percent attri-
Ind. Eng. Chem. Res., Vol. 41, No. 23, 2002 5679 Table 1. Chemical Analyses of Selected Sorbents sorbent designation IGTSS-353 IGTSS-354 IGTSS-314B IGTSS-325A UCI-4169 IGTSS-057 IGTSS-179
chemical analysis, wt % Zn
Cu
Mn
21.1 19.4 27.7 36.9 43.8
Fe
Ti
Al
39.8 30.0 38.9 2.99 28.1 23.5 72.5 35.83 11.3
Zr 13.7
2.3 17.9
sulfur capacitya (g of S/ 100 g of sorbent) 10.35 9.52 13.59 18.10 21.48 42.31 9.03
a Based on 2Cu + H S ) Cu S + H , ZnO + H S ) ZnS + H O, 2 2 2 2 2 and MnO + H2S ) MnS + H2O.
Figure 6. Comparison of attrition results of different materials at GTI and ANRF.
tion loss after 5 h is defined as the “air jet index” (AJI) and is calculated from the elutriated fines to give a relative estimate of the attrition resistance of the granular material. The amounts of fines generated at the end of the first hour, at the end of the second hour, and during the last 3 h are determined during each test. The performance of GTI’s attrition unit was verified, using four different materials, in collaboration with Akzo Nobel Research Facility (ANRF) in Pasadena, TX. ANRF is one of the laboratories that participated in establishing the ASTM procedure. The materials tested included FCC catalyst, dolomite 245, limestone 246, and the UCI-4169 zinc titanate sorbent. The FCC material is widely used in the petroleum industry and may be considered as a suitable standard against which sorbent performance in attrition tests can be compared. The results obtained are presented in Figure 6, indicating that the results obtained using the GTI attrition unit are well within the experimental error of the test method and are consistently slightly lower than those obtained at ANRF. A striking result is the one obtained with the UCI-4169 zinc titanate sorbent, showing an air jet index >75%. Chemical Reactivity and Regenerability. The granular sorbents were evaluated for their H2S removal efficiency and effective capacity for sulfur absorption from simulated fuel gas mixtures ranging from simple H2S-H2-H2O-N2 to compositions simulating the KRW commercial fuel gas composition (15% H2, 24% CO, 6% CO2, 6% H2O, 0.5-2% H2S, and 47-48.5% N2). Sorbent evaluation was carried out in a 2.5-cm-diameter quartz packed-bed reactor system at ambient pressure and at temperatures in the range of 350-550 °C, at a space velocity of about 2000 h-1 (at STP). The reactor system is configured for flowing gas downward for both sulfidation and regeneration tests. The volume of the sorbent bed was typically about 20 cm3. The particle size of the sorbents tested is in the range of 180-425 µm. Sulfided sorbents were regenerated also at a space velocity of about 2000 h-1 using O2-N2 and O2-H2O-N2 mixtures. Both 538 °C (1000 °F) and 593 °C (1100 °F) were investigated with dry and wet regeneration gases. The extent of desulfurization and regeneration was determined by analyzing the reactor exit gas for H2S and SO2, respectively, with a dedicated gas chromatograph. Further details about the experimental arrangement and procedure employed can be found in ref 12.
It should be noted that although these sorbents are intended for a fluid-bed desulfurization application, packed-bed testing is used merely as a convenient tool to screen the sorbents. This approach is routinely used by researchers engaged in similar work. In addition, to provide a basis for comparison, the UCI-4169 zinc titanate sorbent was included in the testing. This sorbent had been evaluated extensively at GTI in a bench-scale fluidized-bed reactor operating at about 550 °C.13 Results and Discussion Sorbent Characterization. To illustrate the exceptional characteristics of this new class of ZnO-TiO2 sorbents, two formulations will be used in this paper. These are designated as IGTSS-353 and IGTSS-354, and were developed based on an early version of the sorbent preparation procedure. Both sorbents had a nominal ZnO content of 30 wt %. IGTSS-354 contained an additive, zirconium oxide (ZrO2), at approximately the 10% level (nominal). This additive was inspired from the work of Sasaoka et al.14 in Japan, which will be discussed in a later section of this paper. The chemical analyses of these two sol-gel formulations as well as those of a selected number of “conventional” sorbents are reported in Table 1. It should be noted that the ZnO content of the IGTSS-353 and IGTSS-354 sorbents are about half that of the baseline UCI-4169 zinc titanate sorbent. The results of physical characterization of the IGTSS353 and IGTSS-354 sorbents as well as those of selected sorbents previously prepared by “conventional” techniques are summarized in Table 2. A comparison of these results indicates that although all these sorbents have similar densities and porosities, the mercury pore surface areas of the sol-gel-derived sorbents are 1-2 orders of magnitude higher than those prepared by other techniques. The high surface areas of the solgel-derived sorbents should be attributed to their significantly smaller pore diameters. As indicated, the median pore diameters of the sol-gel sorbents are remarkably smaller than those of the other selected sorbents shown in Table 2. For example, the IGTSS354 sorbent has a median pore diameter that is over 120 times smaller than that of the UCI-4169 zinc titanate sorbent. This desirable combination of high porosity, high surface area, and small pore diameters resulted in higher reactivity and much higher attrition resistance for the sol-gel sorbents compared to “conventional” sorbents. Comparison of Attrition Resistance. Figure 7 compares the results from attrition tests carried out on the sol-gel sorbents (i.e., IGTSS-353 and IGTSS-354) against those obtained previously with the commercial
5680
Ind. Eng. Chem. Res., Vol. 41, No. 23, 2002
Table 2. Physical Characteristics of Selected Sorbentsa sorbent designation
main reactive metal
particle (Hg) density (Fb) (g/cm3)
skeletal (He) density (Fa) (g/cm3)
Hg pore volume (cm3/g)
porosityb (%)
surface area (m2/g)
median pore diameter (µm)
IGTSS-353 IGTSS-354 IGTSS-314B IGTSS-325A UCI-4169 IGTSS-057 IGTSS-179
Zn Zn Zn Zn Zn Mn Cu
2.89 2.53 2.85 2.31 2.19 2.39 3.65
3.66 4.12 4.48 4.55 3.14 4.69 4.17
0.072 0.153 0.127 0.212 0.138 0.206 0.034
20.9 38.6 36.3 49.1 30.3 49.1 12.5
39.23 56.92 3.33 2.59 0.71 9.19 5.21
0.0068 0.0110 0.82 1.20 0.82 0.19 0.51
a Corrected for interparticle void. b Calculated based on corrected values as (1 - F /F ) × 100, or equivalently as F × (Hg pore volume) b a b × 100.
Figure 7. Attrition resistance of new and “conventional” ZnObased sorbents.
Figure 9. Attrition resistance of selected sorbents (US DOE/ NETL definition).
Figure 8. Attrition resistance of selected sorbents (second hour only).
Figure 10. Sulfidation (desulfurization) performance of IGTSS354 sorbent.
UCI-4169 zinc titanate sorbent and two selected bulk, zinc-based sorbents (i.e., IGTSS-314B and IGTSS-325A) developed earlier.4 Also shown in Figure 7 are the results from the baseline FCC material. The results reported in Figure 7 are in terms of the attrition jet index (AJI), as defined by the ASTM D5757-95 procedure, which corresponds to the 5-h loss. As shown in this figure, the new sol-gel-derived zinc-based sorbents achieved attrition losses that are about half that of the FCC material and about one-tenth of those of the “conventional” bulk, zinc-based sorbents. It should be noted that the FCC material, which has an AJI of about 5%, is quite suitable as a standard against which sorbent performance in attrition tests can be compared. It is generally accepted in the catalyst industry that materials with an AJI of 5 or below are suitable for use in a transport reactor.6
Figure 8 reports the attrition losses obtained with the IGTSS-353 and IGTSS-354 sorbents during the second hour only of the 5-h attrition test. This modified procedure of the ASTM D5757-95 method was reported to provide an adequate assessment of the relative attrition resistance of sorbent materials for the purpose of screening effective sorbents for the Pin˜on Pine application.6 Also shown in Figure 8 are the results obtained with the leading copper-based regenerable sorbent developed at GTI (i.e., IGTSS-179) and the results reported for EX-SO3 zinc titanate sorbent.6 As mentioned earlier, this latter is the only sorbent that has been considered for the Pin˜on Pine Clean Coal Demonstration Project. As shown, both IGTSS-353 and IGTSS-354 exhibit significantly lower attrition losses that are 3-4 times better than the EX-SO3 sorbent. Figure 9 compares the performance of IGTSS-353 and IGTSS-354 sorbents with those of other selected copper-
Ind. Eng. Chem. Res., Vol. 41, No. 23, 2002 5681
Figure 11. Sulfur balance in a typical sulfidation/regeneration cycle.
based (IGTSS-179), manganese-based (IGTSS-057), and zinc-based (UCI-4169) sorbents in the attrition unit. Attrition resistance in this figure is reported in accordance with DOE/NETL’s definition of (fifth hour minus first hour)/4, that is, the average per hour loss due to attrition with the first-hour loss not taken into consideration to eliminate any particle size effects on attrition.8 As shown, IGTSS-353 and IGTSS-354 distinctively demonstrated low attrition losses that significantly exceed the stringent requirement of the Pin˜on Pine application, as indicated by the solid horizontal line in Figure 9. Chemical Reactivity and Regenerability. The results from two consecutive sulfidation tests carried out on the IGTSS-354 sorbent at 450 °C are presented in Figure 10. As shown, in both tests this sorbent achieved an effective capacity for sulfur absorption approximating 7.5 g of S/100 g of material (taking 100 ppmv exit H2S concentration as a breakthrough point). This value corresponds to about 79% conversion of the ZnO component in the IGTSS-354 sorbent. Figure 10 also clearly shows the sorbent maintained its efficiency for H2S removal in addition to its effective capacity in the second cycle, an indication of successful regeneration following the first sulfidation test. The H2S and SO2 concentrations in the reactor product gas during the first sulfidation/regeneration cycle are presented in Figure 11. Both tests were carried out at the same space velocity (≈2000 h-1 at standard temperature and pressure). In addition, the feed gas during the sulfidation test contained 2% H2S, and the O2 content of the regeneration feed gas was regulated at 3 vol %, such that a 2 vol % SO2 concentration would result in the dry product gas, based on regeneration reaction stoichiometry and complete conversion of the O2 to SO2. The results obtained indicate an excellent balance between the amount of sulfur loaded during sulfidation and the amount of sulfur released during regeneration. Chemical analyses on sorbent samples taken from the gas inlet section indicated a sulfur content of about 9% (sulfide S) after sulfidation and 0.2% (sulfate S) after regeneration, which are consistent with the experimental results, indicating an average sulfur loading of about 7.5% (Figure 10) and complete regeneration (Figure 11).
Figure 12. Performance of IGTSS-354 sorbent in extended testing.
The results of eight (8) consecutive sulfidation tests completed on the IGTSS-354 sorbent are presented in Figure 12, indicating that the effective sulfur capacity of this sorbent was maintained in the cyclic process. All sulfidation tests in this series were carried out at 450 °C using a simulated fuel gas consisting of 2% H2S, 10% H2, and balance N2. In cycles 1-5 the sulfided sorbent was regenerated at 593 °C (1100 °F) using 3% O2, 10% H2O, and balance N2 (this is referred to as “wet” gas). During regeneration tests 6-8, the effects of temperature and steam content of the feed regeneration gas were investigated. The sixth regeneration was carried out with the “wet” gas at 538 °C (1000 °F). The seventh regeneration was carried out at 593 °C, but this time using a “dry” regeneration gas consisting of 3% O2 in N2. The eighth regeneration was carried out using the “dry” gas at 538 °C. The effects of these changes in the operating conditions on sorbent regenerability and subsequent desulfurization performance are discussed below. The results from the fifth and seventh regeneration tests are compared in Figure 13. As shown, in both cases, the SO2 content of the regeneration product gas closely approaches the theoretical value with a sharp breakthrough when the sorbent is fully regenerated. Therefore, the IGTSS-354 sorbent can be regenerated successfully at 593 °C with or without steam in the feed gas. At 538 °C, however, the presence of steam in the
5682
Ind. Eng. Chem. Res., Vol. 41, No. 23, 2002
Figure 13. Effect of steam on regenerability of IGTSS-354 sorbent at 593 °C (1100 °F). Table 3. Physical Characteristics of Selected Fresh and Reacted Sorbent Samplesa IGTSS-354 IGTSS-354 IGTSS-354 (second (fifth (fresh) sulfidation) regeneration) particle (Hg) density (Fb), g/cm3 skeletal (He) density (Fa), g/cm3 Hg pore volume, cm3/g porosity,b % surface area, m2/g median pore diameter, Å BET N2 surface area, m2/g
Figure 14. Effect of steam on regenerability of IGTSS-354 sorbent at 538 °C (1000 °F).
regeneration gas is required to fully regenerate the sorbent. As shown in Figure 14, in the presence of steam at 538 °C, the theoretical amount of SO2 in the product gas is achieved during the first 90 min only of the regeneration test. It is also seen that an additional 1 h was required to complete the regeneration. However, because of the favorable results obtained during the seventh sulfidation test (following R6, see Figure 12), regeneration at 538 °C with wet gas can be deemed successful. At 538 °C with “dry” gas, the regeneration test was quite sluggish and did not complete following about 4 h on stream. Although not shown in Figure 12, the effective sulfur capacity achieved in the ninth sulfidation, following the eighth regeneration with a “dry” O2-N2 gas mixture at 538 °C, showed about 50% reduction. This is likely due to incomplete regeneration, which suggests steam is a necessary component in the regeneration feed gas for successful regeneration of the IGTSS-354 sorbent at 538 °C. The physical characteristics of reacted IGTSS-354 sorbent samples are compared with those of fresh sorbent in Table 3. In addition to the properties listed in Table 2, the nitrogen BET surface area values are also reported. As shown, following five sulfidation/ regeneration cycles, the surface area of the regenerated IGTSS-354 sorbent was reduced by about 50%, whereas its median pore diameter increased by about 86%. This decline in desirable properties are likely to be due
2.53
2.70
2.81
4.12
3.73
4.59
0.153 38.6 56.92 110 71.7
0.102 27.6 23.81 160 26.1
0.138 38.8 27.82 205 19.5
a Corrected for interparticle void. b Calculated based on corrected values as (1 - Fb/Fa) × 100 or equivalently as Fb × (Hg pore volume) × 100.
mostly to the sorbent exposure to temperatures as high as 650 °C during the exothermic regeneration step (see temperature profile shown in Figure 13). Although these effects resulted in a slight decrease in the effective capacity of the sorbent for sulfur absorption, they did not prove detrimental to the overall chemical reactivity of the sorbent. This indicates the sol-gel sorbents are capable of tolerating this high temperature; however, for the sorbent to retain as much as possible of its desirable properties during the cyclic process, it is probably best to avoid exposure to temperatures exceeding 635 °C (1175 °F). This can be accomplished by initiating regeneration at the lower temperature of 538 °C, for example. The results from two consecutive regeneration tests carried out on the IGTSS-353 sorbent using the “wet” gas first at 593 °C and then at 538 °C are reported in Figure 15. At 538 °C, this sorbent exhibited similar regenerability characteristics to those of the IGTSS-354 sorbent formulation. The same is true for regeneration at 593 °C, except for a less sharp breakthrough toward the end of the regeneration test. This is an indication the IGTSS-354 possesses better regenerability characteristics, possibly because of its ZrO2 additive. More importantly, however, because the regenerability characteristics of the two sorbents are not drastically different, the improved regenerability of this new class of Zn-based sorbents, compared with that of “conventional”
Ind. Eng. Chem. Res., Vol. 41, No. 23, 2002 5683
Figure 15. Regenerability of IGTSS-353 sorbent at different temperatures.
Figure 16. Performance of IGTSS-353 sorbent following regeneration at different temperatures.
sorbents, can be attributed mostly to the method of sorbent preparation. This reasoning is supported further by the results reported in Figure 16 of the sulfidation tests following the regeneration tests presented in Figure 15. Again, successful regeneration of the IGTSS353 sorbent in the presence of steam at both 538 and 593 °C is confirmed. Figure 16 also indicates that, despite its higher theoretical sulfur capacity (see Table 1), the IGTSS-353 achieved a slightly lower effective
capacity than the IGTSS-354 sorbent, suggesting the possibility that the improved chemical reactivity of this sorbent can be attributed to its ZrO2 additive. The improved chemical reactivity toward H2S of the IGTSS-354 sorbent over that of the IGTSS-353 sorbent is attributed to the ZrO2 additive. As shown in Table 2, the addition of ZrO2 increased the surface area and decreased the particle density, leading to a more porous sorbent. The improved pore structure of the IGTSS-354 sorbent also explains its improved regenerability. As discussed in more detail below, these interpretations are consistent with similar arguments made by Sasaoka et al.14 More importantly, for regeneration, the ZrO2 additive serves to improve the tolerance of the sorbent pore structure for sintering, as the work of Sasaoka et al.14 also suggested. To assess the improved regenerability of this new class of Zn-based sorbents, developed based on a simplified sol-gel procedure, over that of a typical “conventional” Zn-based sorbent, the UCI-4169 zinc titanate sorbent was sulfided at 550 °C (to achieve a sulfur loading similar to that of IGTSS-354 at 450 °C, see Figure 3) and then regenerated under the same conditions as the IGTSS-354 sorbent. As shown in Figure 17, whereas it took only about 2 h to fully regenerate the IGTSS-354 sorbent, the UCI-4169 zinc titanate sorbent took an entire 14 h, and even after that period, the SO2 concentration in the regeneration product gas was still being measured at about 0.2% (i.e., 2000 ppmv SO2). Also, as seen in Figure 17, regeneration of the “conventional” UCI-4169 sorbent is quite sluggish, a clear indication of significant sulfate formation. It is wellestablished in the literature that zinc titanate cannot be efficiently regenerated at temperatures below about 625 °C.15 Therefore, the zinc titanates prepared using GTI’s proprietary technique have inherently lower regeneration temperatures than “conventional” zinc titanate sorbents. The above results underscore the positive role steam plays in ensuring successful regeneration of the solgel-derived sorbents at reduced temperatures. Furthermore, experimental data strongly hint to a positive role that ZrO2 also plays in achieving the same goal, in addition to improving chemical reactivity toward H2S at moderate temperatures. This new class of zinc-based
Figure 17. Regenerability of GTI’s zinc titanate sorbent vs that of UCI-4169.
5684
Ind. Eng. Chem. Res., Vol. 41, No. 23, 2002
sorbents possesses a unique combination of high chemical reactivity at moderate temperature, significantly high resistance to attrition, and regenerability at temperatures lower than required by typical zinc titanates. Therefore, these new sorbents appear to hold significant promise to satisfy the stringent requirements of the transport reactor desulfurization application, such as Pin˜on Pine. In this application, the sulfidation reactor operates as a dilute fluid-bed reactor, requiring a highly reactive sorbent to achieve the desired level of desulfurization efficiency (i.e.,
