Movement of IGT'S Hot Gas Cleanup Sorbent Towards the IGCC Demonstration Stage

FINAL TECHNICAL REPORT

November 1, 1999, through October 31, 2000

Project Title: MOVEMENT OF IGT’S HOT GAS CLEANUP SORBENT TOWARDS THE IGCC DEMONSTRATION STAGE

ICCI Project Number: 99-1/1.3A-1

Principal Investigator: Rachid B. Slimane, Gas Technology Institute

Other Investigators: Brett E. Williams, Gas Technology Institute, Javad Abbasian and Shahryar Rabiei, Illinois Institute of Technology

Project Manager: Ken K. Ho, Illinois Clean Coal Institute

Abstract

The overall objective of this program was to advance GTI’s most promising regenerable metal oxide-based sorbent towards commercial demonstration at the Piñon Pine entrained-bed high temperature coal gas desulfurization unit.

The copper-based IGTSS-326A sorbent was originally selected and produced via spray drying by United Catalyst, Inc. Evaluation tests showed the commercially produced formulation possessed improved attrition resistance and only slightly lower effective sulfur capacity than the IGTSS-326A produced at GTI. Unfortunately, fluidization of the commercially produced sorbent also proved to be difficult and it was concluded it would not be possible to demonstrate the suitability of this sorbent for the transport reactor application according to the test protocol provided by DOE/NETL. An alternative leading metal oxide sorbent was selected for demonstration instead. This zinc-based sorbent, designated as IGTSS-360, was developed based on a novel sol-gel processing technique in a recent DOE-sponsored program at GTI.

In this program the sol-gel technique was simplified considerably to render it more economical, less time consuming, and easily adaptable to large-scale production, without compromising sorbent properties. In addition, this novel sorbent preparation technique was adapted to commercial practice in coordination with a suitable commercial sorbent manufacturer (Chemat Technology, Inc.). Several variations of the leading zinc-based sorbent formulation (IGTSS-362) were successfully produced commercially. One superior zinc-based formulation (IGTSS-365B) was evaluated under DOE/NETL Test Protocol for qualification of candidate sorbents towards commercial demonstration at Piñon Pine IGCC facility.

To bridge the gap between bench-scale testing and future large-scale demonstration, evaluation of GTI’s leading sorbent in DOE/NETL’s facilities is needed. Furthermore, based on the results obtained in this program, it is highly recommended that additional regeneration studies, involving mixtures of fresh and partially sulfided sorbent samples, be undertaken to closely simulate regeneration conditions at Piñon Pine.

Executive Summary

Under the sponsorship of DCCA/OCDM as well as DOE/NETL, extensive research at GTI showed that only sorbents based on copper oxide possessed the best combination of high attrition resistance, sulfur removal efficiency, and extent of pre-breakthrough conversion at moderate desulfurization temperature. The durability of one leading sorbent formulation (IGTSS-326A) was demonstrated in over 50 sulfidation/regeneration cycles. However, work at GTI cast doubt on the feasibility of formulating an effective bulk sorbent based on zinc oxide, using conventional sorbent preparation techniques, such as co-precipitation and solid oxide mixing followed by granulation or spray drying, for a demanding commercial desulfurization application, such as the transport reactor. These techniques require very high thermal treatment temperatures to impart physical strength, and as a result, they often produce sorbents with low reactivity. It was shown that even when chemical reactivity was compromised significantly, the minimum attrition resistance requirement for the transport reactor could not be met. Because sorbents based on ZnO are best suited for desulfurization, especially at moderate temperature, an alternative sorbent preparation approach was deemed necessary.

The objective of this program was to bring GTI’s leading metal oxide sorbent one step closer to commercialization by producing sufficient quantities of the leading formulation with the required specifications by a commercial sorbent manufacturer. The resulting sorbent was then evaluated according to a test protocol provided by DOE/NETL to demonstrate the suitability of the sorbent and the compatibility of the sorbent’s optimum operating conditions for use in the Piñon Pine plant hardware as it currently exists.

Initially, ten pounds of the leading IGTSS-326A copper-based sorbent was prepared via spray drying by United Catalyst, Inc. (UCI). Following evaluation tests, this commercially produced sorbent was found to exhibit improved attrition resistance and comparable sulfidation reactivity to those previously obtained with the IGTSS-326A sorbent produced at GTI. Similar to this latter sorbent, however, fluidization of the spray-dried sorbent in reducing atmospheres, under bubbling fluid-bed conditions, proved to be difficult. It was concluded it would not be feasible to evaluate this sorbent under the experimental conditions specified in the DOE/NETL Test Protocol. To a large extent, this protocol is tailored to zinc-based sorbents, which require lower fluidization velocities than copper-based sorbents.

Efforts were directed to pursue the movement towards demonstration of an alternative leading formulation from a new class of zinc-based sorbents that were recently developed in a DOE-sponsored program at GTI. These new zinc-based sorbents were prepared using a modified procedure based on sol-gel processing of organic and/or inorganic precursors. This procedure departs significantly from “conventional” sorbent preparation approaches, and was shown to produce sorbents with the highly desirable combination of high reactivity (desirable pore size distribution and high surface area), high attrition resistance, and regenerability at lower temperatures. These highly desirable characteristics proved unattainable using “conventional” sorbent preparation techniques, especially for the case of sorbents based on ZnO. Selected formulations from this new class of sorbents include IGTSS-353, IGTSS-354, and IGTSS-360.

Initial efforts in this program focused on the simplification of the sol-gel-based sorbent preparation procedure to minimize the number of steps involved, thereby reducing processing times and sorbent cost. A formulation equivalent to the IGTSS-360 sorbent was prepared using a simplified approach that eliminated a time-consuming step (refluxing) from the original procedure. This sorbent, designated as IGTSS-362, was evaluated for its attrition resistance, chemical reactivity, effective sulfur capacity, and regenerability. It was determined that the sorbent preparation steps eliminated had no adverse effects on the chemical reactivity or attrition resistance of the resulting sorbent.

Following the encouraging results obtained with the IGTSS-362 sorbent, arrangements were made for the commercial manufacture of this formulation by Chemat Technology, Inc., using the simplified sol-gel procedure. Additional modifications in the procedure were sequentially made during the commercialization process as recommended by Chemat, resulting in the IGTSS-362C, IGTSS-362E, and IGTSS-362F sorbents, versions of which were made both at GTI and at Chemat. A final step in optimizing the sol-gel procedure was to investigate the possibility of increasing the effective sulfur capacity, by increasing the ZnO content of the prepared sorbent beyond the nominal 40 wt% ZnO used in the IGTSS-362 sorbent series. Sorbents containing approximately 45% ZnO (IGTSS-364), 50% ZnO (IGTSS-365), and 60% ZnO (IGTSS-363) were prepared at GTI. Each of these sorbents was evaluated for its attrition resistance, chemical reactivity, and effective sulfur capacity. Based on the results obtained, a sorbent (IGTSS-365B) with the optimum ZnO content of 50% was selected for commercial production and evaluation.

Physical characterization and packed-bed reactor evaluation indicated the GTI-produced sorbents had very similar properties to their commercially produced counterparts. It was evident that none of the modifications that were systematically implemented during the sol-gel preparation step compromised the effectiveness of these sorbents for H₂S removal from a simulated fuel gas mixture. In addition, with the exception of the sorbent containing 60.4% ZnO (i.e., IGTSS-363), all of the sorbents had attrition indices that ranged from 1.8 to 2.6%, well below the transport reactor requirement of » 4%. It appears that attrition resistance depends on the ZnO content of the sorbent and is adversely affected when the ZnO content exceeds 50% by weight. Furthermore, additional amounts of the reactive component ZnO did not appear to increase the effective capacity of the sorbent for sulfur absorption. A ZnO content of about 50% by weight appears to be optimum for maximizing the effective capacity of these sol-gel derived sorbents, while still achieving desirable attrition indices.

Comparison of physical characterization results indicates the mercury pore surface areas of the sol-gel-derived sorbents are one to two orders of magnitude higher than that of the IGTSS-326A copper-based sorbent. This high surface area should be attributed to significantly smaller pore diameters in these sorbents. Results show the median pore diameters of the sol-gel sorbents are about 16 times smaller than those of the IGTSS-326A sorbent. This highly desirable combination of high porosity, high surface area, and small pore diameters resulted in high chemical reactivity and high attrition resistance for these zinc-based sorbents.

All four commercially-produced sorbents were subjected to a 5-cycle test in the packed-bed reactor for the purpose of screening the superior sorbent formulation for bench-scale testing in the HPTR unit. All sorbents demonstrated a high H₂S removal efficiency (< 1 ppmv H₂S in the cleaned gas) and reasonably maintained an acceptable effective sulfur capacity (i.e., > 8 g S/100 g of sorbent) throughout the five cycles completed. During the first sulfidation all sorbents achieved similar effective sulfur capacities, ranging from 11 to 12 g S/100 g of sorbent. However, only the IGTSS-365B sorbent maintained this level of performance by the fifth cycle, while the effective sulfur capacities of the other three commercially produced sorbents declined to levels ranging from 8 to 9 g S/100 g of sorbent. Based on these results and favorable attrition resistance and physical properties, the commercially produced IGTSS-365B sorbent was selected for evaluation in the bench-scale HPTR unit according to the DOE/NETL Test Protocol. This protocol was designed for qualification of candidate sorbents specifically for the Piñon Pine application.

Scoping test results showed that the IGTSS-365B sorbent was capable of being cyclically loaded to effective sulfur capacities exceeding 8 g S/100 g of sorbent, with residual H₂S concentrations below 20 ppmv in the effluent gas. In addition, with breakthrough arbitrarily defined at 100 ppmv H₂S, this sorbent is capable of achieving an effective sulfur capacity approximating 12 g S/100 g of sorbent by the fourth cycle, similar to the results obtained during testing at ambient pressure. There appears to be a gradual decline in sorbent efficiency with cycling, as demonstrated by higher and higher H₂S concentrations in the effluent gas as the breakthrough point is approached. This decline may be due to sorbent exposure to temperatures exceeding 650°C during regeneration, as indicated by the measured temperature profiles.

Results further show that the IGTSS-365B sorbent could be successfully regenerated at a starting temperature of 538°C with a feed gas containing 7 vol% O₂. However, regeneration with 14 vol% O₂ in the feed gas led to temperature excursions exceeding 825°C, which caused severe overheating and sintering of the sorbent, as was substantiated by its poor performance in the subsequent sulfidation test. It appears there is an optimum combination of O₂ content in the feed gas and starting regeneration temperature capable of creating suitable conditions to initiate and sustain sorbent regeneration.

During extended multiple-cycle testing, the regeneration conditions were further refined to minimize the amount of heat generated while still allowing complete regeneration to occur within a reasonable amount of time. The conditions used were a starting regeneration temperature of 565°C and a feed gas containing 2 vol% O₂. The results indicate both sorbent efficiency for H₂S removal and the effective sulfur capacity gradually and steadily decline over the 8-½ cycles completed. Regeneration test results indicate the sorbent was not exposed to temperatures higher than 650°C. Therefore, this gradual decline is probably due to sorbent sintering effects upon prolonged exposure to elevated temperature.

To assess the degree of fluidization and to determine the effects of cyclic testing on sorbent properties, post-test samples, from the top and bottom portions of the sorbent bed, from the various testing campaigns, were physically and chemically analyzed. The results obtained appear to confirm the conclusions drawn previously. First, none of the reacted samples showed any significant amount of sulfate formation, indicating that cyclic decline in sorbent performance during any of the tests was not due to incomplete regeneration because of sulfate formation. Second, the amounts of sulfide sulfur measured by chemical analysis for each sample are in agreement with the final sulfur loading attained in each scoping test, as estimated based on the breakthrough curve. Third, given the consistent zinc contents, loss of sulfur capacity due to zinc volatilization can be ruled out, as should be the case given the moderate operating temperature for sulfidation (510°C). Fourth, the effects of the high temperature excursions are evident by an enormous change in the median pore diameter accompanied by a considerable reduction in porosity and surface area. This is especially true for samples where the sorbent was exposed to temperatures as high as 825°C, leading to dramatic declines in sorbent performance. Finally, the negligible degree of discrepancy between the measured properties of the samples taken from the top and bottom portions of the sorbent bed after each test is an indication of the high degree of fluidization that was consistently maintained during HPTR testing.

The results discussed above in connection with evaluation of the IGTSS-365B sorbent according to the DOE/NETL Test Protocol are encouraging. It is evident there is no sulfate formation upon regeneration of the IGTSS-365B sorbent, even at a starting regeneration temperature as low as 538°C (1000°F). This is an indication of improved regenerability for this sorbent compared to other zinc-based sorbents developed using “conventional” preparation techniques, for which sulfate formation was reported to be a major concern. The slight decline in the IGTSS-365B sorbent performance during the extended cycle test is solely due to temperature excursions during sorbent regeneration, which adversely affected sorbent properties. However, as discussed earlier, in the Piñon Pine transport reactor system, only a slipstream of partially sulfided sorbent is circulated between the desulfurization and regeneration vessels. Therefore, the sorbent in this reactor system is always in dilute phase, and temperature excursions are completely avoided, even when neat air is used for regeneration. Under these conditions, the results obtained in this program indicate the IGTSS-365B sorbent can be regenerated successfully, thereby maintaining its high effective capacity for sulfur absorption in a cyclic hot gas cleanup process.