RELATED APPLICATIONS This application is a continuation-in-part of application serial number 08/170,580 filed December 20, 1993 by Copeland for "High Temperature Regenerable Hydrogen Sulfide Removal Agents With Refabrication" which is, in turn, a continuation-in-part of application serial no. 07/844,829 filed March 3, 1992 by Copeland for "High

Temperature Regenerable Hydrogen Sulfide Removal Agents" (now U.S. Patent No. 5,271,907). All of these patents and applications are assigned to TDA Research, Inc., the assignee of the present invention. The disclosures of each of these patents and applications are hereby incorporated herein by this reference.

FIELD OF THE INVENTION

This invention relates to high temperature hydrogen sulfide removal from coal-derived fuel gases, including the use of a new structure for carrying a regenerable sorbent agent for absorbing hydrogen sulfide while recovering elemental sulfur. The new structure is identified as a "geode", more particularly, the geodes of this invention provide a very long-lived sorbent for multiple regeneration cycles under the conditions expected in a high temperature plant. The geodes of this invention have particular utility for the removal of sulfur contaminants from the gaseous product of coal fueled power plants and coal synthetic fuels plants.

BACKGROUND OF THE INVENTION

The development of an advanced clean-coal technoloσv will permit the use of coal to replace oil and gas while

eliminating the environmental penalty now associated with sulfur-containing coal. One way of producing fuel and power from coal is by way of pulverized coal combustion. Another, more efficient, way of doing so is by gasification of coal. Coal gasification produces a gas stream suitable for the production of electrical power, gaseous and liquid fuels, or other products. Before this can happen on a large scale, a low-cost, clean coal gas must be croduced.

Coal gasification plants carry the premise of highly efficient utilization of coal. Electrical energy, for example, can be generated by the partial oxidation of coal in a gasifier/molten carbonate fuel cell system (MCFC) or in an integrated gasification combined cycle (IGCC) plant. An IGCC plant generates power by the direct contact of hot coal-derived gases with turbine blades, and is one of the most promising new technologies for the production of base-load electric power from coal.

While coal is the most abundant energy resource in the United States, and IGCC and MCFC plants have good generating efficiencies, current coal-based power generation imposes greater environmental burdens than oil or natural gas. Coal-derived gases contain particulates, tars, ammonia, alkali metals, and sulfur. These materials are not only pollutants, but can cause corrosion, erosion or deposition on the turbine blades of a power plant.

Coal-derived gases contain significant levels of sulfur contamination. When coal is gasified, most of the total sulfur content is converted to hydrogen sulfide (H-S) . The hydrogen sulfide concentration in the coal gas depends on the amount of sulfur initially present in the

coal and on the nature of the coal gasification process used. Gas-phase concentrations of hydrogen sulfide on the order of several thousand parts per million (ppm) are typical, and 2,000 to 10,000 ppm is not unusual. Sulfur contamination of the coal gas is an environmental problem and also an operational problem. Because sulfur is a useful chemical, its recovery is also worthwhile economically.

In the environment, the un-removed sulfur species present in coal gases can react with oxygen and atmospheric water vapor to produce sulfuric acid and can contribute to the problems of acid rain. The United States Environmental Protection Agency standards of October 1, 1985 (40 C.F.R. Part 60, Subpart LLL) limit natural gas processing plants and petroleum refineries to sulfur emissions on the order of less than 90 parts per million (ppm) , requiring sulfur removal in the range of 99% efficiency.

The European Community will require a minimum of 98.5% sulfur recovery rates by 1992, and the Federal Re¬ public of Germany's regulations currently require up to 99.5% recovery. The United States New Source Performance Standards require at least 90% removal of sulfur for most new plants. IGCC plants, however, have the promise of greater than 99.8% removal of sulfur, and would largely eliminate the environmental penalties of coal use.

Economically, sulfur is valuable as a constituent of sulfuric acid, the largest single chemical consumed in the United States (over 11 million long tons of sulfur consumed in 1988) . In the United States, elemental sulfur is typically recovered by steam injection from

underground deposits, but this is thermally inefficient. Natural gas and petroleum processing is another large source, but these show signs of decline in the United States. Accordingly, if elemental sulfur were to be re- covered as a by-product of the desulfurization of coal gas, the recovered sulfur would have a ready market.

To recover sulfur from the coal gas stream and to minimize the emission of sulfur compounds, an IGCC plant typically operates with a reaction step and a separate sulfur removal step. During the reaction step, coal is converted to product gas (synthesis gas, or "syngas") at high temperature. During the sulfur removal step, physi¬ cal solvents are generally used to remove sulfur products and other contaminants from the crude syngas.

In a typical, "cold gas" sulfur removal process, the removal step reaction cannot take place at the high temperatures encountered in the reaction step. Thus, a cold gas approach requires (a) cooling of the hot (500 - 800°C) syngas to the relatively low temperatures commonly needed for physical solvents, and (b) subsequent reheating of the cleaned syngas prior to its introduction into the gas turbine. These cooling and heating phases tend to increase capital costs and operating costs.

An IGCC plant has the potential for higher conversion efficiency, lower capital costs, and lower pollution impacts than pulverized coal-fired combustion even when used with cold gas cleanup systems. For economically more viable conversion of coal to gas without significant loss of thermal energy, there is a need for a "hot gas" cleanup system, capable of removing sulfur from the coal gas stream at high temperatures, in

the range of 500 - 800°C.

The use of hot gas cleanup can reduce capital costs and improve overall conversion cycle efficiency by elim- inating the need to cool and reheat the gasifier outlet gases. It can also reduce wastewater disposal costs. Other coal gasification technologies besides IGCC and MCFC applications that would significantly benefit from hot gas cleanup include gasifier/diesel engine combina- tions, and processes for producing synthetic fuels from coal.

Many commercial processes are available for cold gas cleanup, but advanced hot gas cleanup systems are just now being tested at the pilot scale. Over the past decade, the United States Department of Energy and its Morgantown Energy Technology Center have made extensive efforts to develop high temperature regenerable desulfurizing agents. Successful sorbents should absorb sulfur so as to provide efficient desulfurization, and should be long-lived or regenerable.

Current hot gas sulfur removal research is focused on regenerable, metal oxide sorbents that remove sulfur from the coal gas, and are then regenerated with air. Some of the metal oxides that have been tried include zinc ferrite, copper zinc oxide, and cuprous oxide.

The most developed candidate is zinc ferrite (ZnFe ₂ 0 ₄ ) : which reacts as follows with the hydrogen sul¬ fide (H ₂ S) contaminant of coal-derived gases to form zinc and iron sulfides (ZnS and FeS) :

ZnFe ₂ 0 ₄ + 3 H ₂ S + H ₂ → ZnS + 2 FeS + 4 H ₂ 0

The zinc and iron sulfide products of the absorption of hydrogen sulfide, when reacted with air, will regenerate the zinc ferrite starting material producing a sulfur dioxide (S0 ₂ ) by-product:

ZnS + 2 FeS + 5 0 ₂ → ZnFe ₂ 0 ₄ + 3 S0 ₂

Sulfur dioxide is a contaminant that must then be disposed of itself. The standard recovery method is to react the sulfur dioxide with limestone, producing ash.

This process incurs significant costs for the purchase of limestone (ranging from $7.00 to $30.00 per ton) and for the disposal of the ash (ranging from $4.50 to $15.00 per ton) . Disposal costs may be expected to rise as the number of available landfill sites is reduced. Another type of sulfur dioxide recovery method (known as the Direct Sulfur Recovery Processes or DSRP) reacts the sulfur dioxide with carbon monoxide (CO) and hydrogen gas (H ₂ ) in a sidestream of hot coal gases to produce water C0 ₂ , and elemental sulfur. The DSRP method incurs costs because the use of the coal gases as a reducing agent decreases the overall energy available from the gasifier by about 4%. Current hot gas cleanup technology, involving regenerable zinc ferrite (ZnFe ₂ 0 ₄ ) and follow-on removal of the sulfur dioxide by-product with limestone or DSRP methods can cost approximately $425.00 per ton of sulfur. This amounts to about 5.7 mills per kilowatt hour, or as much as 9% of the busbar (ideal rated capacity) electrical cost of an IGCC. Clearly, an improved method of hot gas recovery could significantly reduce the cost of the electricity from an IGCC. Full realization of the tremendous commercial potential of coal gas fueled power plants and related technologies awaits the development of an inexpensive and reliable hot

gas clean up method for the removal of sulfur contaminants from the coal derived gas stream.

A successful sorbent must, therefore, be able to remove sulfur so as to leave sulfur levels in the gas stream of 20 ppm or less (a recovery rate greater than 99.8%) ; and it must also have physical and chemical stability in gas atmospheres of 500°C and above. A sorbent pellet will be reused in successive absorption cycles.

Accordingly, and for the sake of economic efficiency, the pellet must be long-lived or, if short¬ lived, must be easily refabricated. In addition to its chemical characteristics, the sorbent's physical characteristics affect its suitability for use in high temperature desulfurization. Among the relevant characteristics are durability, temperature stability, life span, and rate of utilization. Sorbent pellets are subject to physical and chemical degradation over successive process cycles: they may be broken by mechanical transport, fractured by multiple chemical reactions, and contaminated by gasifier ash, which is not removed by upstream filtering.

Although zinc ferrite, and other metal oxides and mixed metal oxides have had some success in high temperature desulfurization of coal gases, they have limitations.

Thus, it can be seen that there is a need for an efficient high temperature desulfurization process that will remove as much as 99.8% of the hydrogen sulfide contaminants of the coal gasification stream. The

desired process would use regenerable sorbents. The de¬ sired process would also consume unwanted by-products of the absorption/regeneration reactions so as to minimize the need for separate recovery and disposal of such by- products. The desired process would recover elemental sulfur in a usable form for resale.

Because the desired process would subject the sorbent pellets used in the system to conditions of heat, chemical reaction and pulverizing forces, which tend to degrade the pellets, there is an additional need for a suitable pellet. If a long-lived pellet is not commer¬ cially feasible, the desired pellet must be one that is short-lived. The desired short-lived pellet must be ca- pable of being refabricated. Accordingly, a method for the inexpensive recovery and reuse of the tin (or other metal species) from the degraded sorbent pellets is de¬ sirable. The desired method would involve the periodic removal of degraded pellets, the chemical recovery of the metal species from the degraded pellet, and the re- fabrication of the high surface area tin oxide (or other metal oxide) in a new pellet.

United States Patent No. 5,271,907 of Copeland, assigned to the assignee of this invention, describes a regenerable sorbent system that removes sulfur contami¬ nants from hot coal gasifier-derived gases to the level of 20 parts per million or less on the absorption side of the process. On the regeneration side of the process, the sorbents are regenerated together with elemental sulfur, which can be recovered and resold. The regenera¬ tion process includes two stages, with the otherwise undesirable sulfur dioxide (S0 ₂ ) by-product of the first stage regeneration reaction being consumed in the second

stage regeneration reaction.

To produce this absorption/regeneration cycle, the system of United States Patent No. 5,271,907 uses two sorbents, one of which is stannic oxide (tin oxide,

Sn0 ₂ ) , and the other of which is an air regenerated metal oxide, such as zinc ferrite (ZnFe ₂ 0 _« ) , producing S0 ₂ during its regeneration.

A unique chemical feature of the system of United States Patent No. 5,271,907 is that, after the stannic oxide and zinc ferrite absorb the sulfur contaminants from the hot gas stream, the two sorbents can be regen¬ erated in two stages. In the first stage of regeneration reaction, the zinc ferrite is regenerated, forming a sulfur dioxide (S0 ₂ ) by-product. In the second stage re¬ generation reaction, the tin oxide is regenerated. Because sulfur dioxide is one of the required species for the regeneration of the tin oxide, this otherwise unde- sirable species is consumed as part of the reaction, producing the regenerated tin oxide and elemental sulfur, which can be reclaimed and resold. A system for refabricating the sorbent pellets is also described in United States Patent No. 5,271,907.

United States Patent Application No.08/170,580 of Copeland, assigned to the assignee of this invention, describes certain additional processes for refabrication of a short-lived pellet suitable for use in the system of United States Patent No. 5,271,907.

The instant invention relates to solving the problems of a long-lived pellet suitable for use in the system of United States Patent No. 5,271,907, and this

invention should be understood with reference to the prior disclosures of United States Patent No. 5,271,907 and of United States Patent Application No. 08/170,580, both of which are incorporated herein by reference.

SUMMARY OF THE INVENTION

This invention includes a new sorbent fabrication method. The new method creates a heterogeneous mixture of two components: a chemically active material (stannic oxide, zinc oxide, zinc ferrite, or other first or second sorbents) in the preferred embodiment of this invention within the system of United States Patent No. 5,271,907 and an inert binder material.

According to this invention, a "geode" pellet is formed. The geode has an inner core and an exterior shell. The inner core of the geode is constituted by the sorbent (that is, the chemically active material) and the exterior shell is constituted by the support material (that is, the binder) . The chemically active material needs to carry little or no structural load because it is supported by the shell. The exterior shell is porous (to allow gases to diffuse to the inner core of chemically active material) , inert (to retain its mechanical strength through multiple cycling) , durable, and strong.

According to a first embodiment, the chemically active material is shaped and formed into a sphere or like shape, and the shell made of inert material is applied as a coating around the inner core.

According to a second embodiment, the chemically active material is mixed with an oil (or water) and the inert material is mixed with water (or oil) and the two

materials are then mixed together. Since the oil and the water will not mix, two phases form in the final mixture which yields a sorbent that is strong and has small regions of chemically active material surrounded by a matrix of the strong inert material.

It should be understood that the geodes of this invention can produce a sorbent with high strength and durability of a supported sorbent, accompanied by a relatively higher chemical content (in the inner core material) and relatively low costs. The geode sorbents of this invention should have a useful life of 200 to 300 cycles within the conditions in the system for high temperature sulfur removal system by regenerable agents as described in United States Patent No. 5,271,907.

DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic view of the first embodiment of a geode produced according to this invention. FIG. 2 is a chart showing experimental results of testing on a second embodiment of a geode produced according to this invention. A 275.5 g sample was tested at 130 psia at a temperature of 573° -121° with 78% of the H ₂ S removed and a sulfur loading at 8.4% at the end of the test.

FIG. 3 is a chart showing experimental results of additional testing on the geode of FIG. 2 at a 0.4763% per cycle mass loss rate.

FIG. 4 is a schematic cross-sectional view of a portion of an alternate embodiment of a geode produced according to this invention.

DETAILED DESCRIPTION OF THE INVENTION New Structure for Hiσh Capacity and Hiσh Strength Sorbent

Adding chemically active material to a performed inert pallet, or extrudate, or agglomerate is a well established procedure for both catalyst and chemical sorbent. (The result produced by this known procedure will be referred to as a "supported sorbent".) The performed inert material supports the mechanical loads on the sorbent and eliminated any need for the chemically active material to contribute to the mechanical strength of the sorbent. Thus, the sorbent can change radically during application without affecting the life of the sorbent pellet.

While the supported sorbent structure previously proposed in U.S. patent no. 5,271,907 and in U.S. patent application serial no. 07/844,829 is very good, the loadings of chemically active material are low and expensive to apply, especially for tin based compounds. The inventor applied stannous sulfate to preformed structures of alumina, silica, and alumino-silicates. It was found that many applications were required to obtain a high stannic oxide loading and that frequent handling

weakened some preformed structures. Other structures were not able to accept a high loading of stannic oxide. Still others could be loaded with high quantities of stannic oxide and remain strong, but the pores were too large to retain the stannic oxide when handled (e. g. , during shipping the stannic oxide would fall out) . All of these supported sorbents were relatively expensive and the process to load the stannic oxide was even more expensive.

The sorbent fabrication method of this invention produces sorbent with the strength and durability of a supported sorbent but with the higher chemical content (i.e., stannic oxide) and lower costs associated with a sorbent formed by mixing metal oxides before firing. The new fabrication method of this invention provides a structurally durable support, using a shaped heterogenous mixture of two components: a chemically active material (stannic oxide) and an inert material (binder) .

According to this invention, the chemically active material [stannic oxide (Sn0 ₂ ) or zinc oxide (ZnO) ] is shaped and formed into a soft or weak inner core. The inner core carries little or no structural load. The binder material is shaped and formed into an exterior shell, which is porous, chemically inert, and strong. The exterior shell provides the structural strength to support the sorbent. Given that the structural and chemical aspects of the sorbent pellet are separated, the new method produces sorbent pellets with very long life. In the remainder of this disclosure, sorbent made by this method is called a "geode", since natural geodes have similar aspects, i.e., a hard, strong outer shell with a hollow interior. The same geode technique can be applied

to zinc oxide, zinc ferrite, copper oxide, and other sorbents for H ₂ S and other gases ( e . g. , HC1) for gasified coal, biomass, and other applications.

Geodes were formed by two different methods: (a) Type 1 -- coating a formed inner core of chemically active, but weak sorbent (stannic oxide) with an exterior shell of a strong, porous, inert material and firing the composite to a strong form and, (b) Type 2 -- mixing the chemically active material with an oil (or water) and the inert material with water (or oil) and then mixing the two materials together; since the oil and water will not mix, two phases form in the final mixture, and firing that mixture then produces a sorbent which is strong with small regions of chemically active material surrounded by a matrix of the strong inert material.

The Type 1 geode forms an exterior shell around a previously formed inner core pellet or spheroid of chemically active material. The general features of this new method can be understood with reference to Figure 1. The coating or exterior shell 12 must be porous to allow gases to diffuse to the interior and it must be chemically inert to retain its mechanical strength with multiple cycling. In one embodiment, the inner core 14 is a weak preformed stannic oxide pellet or spheroid with or without a binder used in its preparation. Since it is desirable to maximize the stannic oxide content of the geode, a 100% stannic oxide pellet or extrudate or agglomerate fired at low temperature (i.e., 900°C or lower) to maximize surface area of the sorbent is preferred.

In a preferred embodiment, structurally weak inner

core spheroids 14 formed by agglomeration (2.6 lb _f crush strength with 100% Sn0 ₂ ) are coated with an outer shell 12 of bentonite (a low cost mineral) forming the strong, chemically inert, and porous outer shell. When coated, the resulting geode pellet was fired at 900°C containing 44% Sn0 ₂ by weight, and had crush strengths > 3 lb _f and >30% void volume.

The Type 2 method as illustrated in Fig. 4 does not have a single soft or weak inner core of chemically active material (e.g., stannic oxide Sn0 ₂ ) , but rather has a large array of inner cores 16 formed at one time and a matrix of the inert material 18 (e.g., bentonite, or talc, etc.) forms the structure of the shell, like a honeycomb surrounding the inner cores. The method produces a structure that is rather similar to a honeycomb panel but much smaller and three dimensional. The Type 2 geode inert shell forms the webs of the panel and the voids are filled by the sorbed material.

The requirements for the binder material to be used as the shell for both types of geodes are defined as follows. To assure that the binder permits transfer of the molecules, the void fraction of the binder must be moderately high (e.g., >20% of the volume occupied by the inert) . Since the binder must have high strength, it must be fired to sinter it into a load bearing matrix; however, if the chemically active material were also sintered at the same time, the sorbent would lose its surface area and become inactive (thus, for stannic oxide the temperature must be less than 1,000°C, 1,832°F, to prevent sintering) . Firing temperatures for the binder must also be significantly higher than the expected operating temperature of the application (e.g., 570°C,

1,058°F for hot gas cleanup of gasified coal). If the binder continuously sinters during use, it would lose its porosity and prevent reactants from reaching the chemically active inner core material (stannic oxide) . Thus, and to summarize the foregoing requirements, in a stannic oxide inner core, the required firing temperature of the binder to be used as the shell is above 800°C (1,472°F) and below 1,000°C (1,832°F).

In addition, the chemically active inner core material must represent a significant fraction of the mass of the geode ( e. g. , >25% stannic oxide by weight) . Other sorbents (e. g. , ZnO, Fe ₂ 0 ₃ , CuO, MnO, V ₂ 0 _s or mixtures of these metal oxides with Sn0 ₂ ) could be used with lower costs due to the higher loadings of chemically active material. Requirements of geodes made with zinc oxide, copper oxide, etc., would be somewhat different and would vary with the operating temperature of the application, such requirements could be readily ascertained by one skilled in the art without undue experimentation.

Experimental Testing of Geodes

Given that the geodes of this invention are a new type of sorbent structure, development of the geode structure was conducted in two steps, a) first, the inert binder shell material, firing temperature and porosity were developed, and b) second, the heterogenous geode structure, including the commercially reactive inner core and the inert shell was tested. Based on the requirements for geodes (previously given) , binders were tested and selections of preferred materials made. Bentonite (HPM 20)™, Polargel T™, and Volclag™ had a high crush strength but a low porosity. Graphite added to the inert

shell material before mixing with the chemically active material (stannic oxide) increases the porosity of the inert materials with high strength. Two types of bentonites (HPM-20 and Polargel) were tested both with and without graphite. Adding graphite to the binder increased porosity; nevertheless, a high crush strength can still be retained as shown below:

binder graph. wt % porosity crush cc/g strength lb/mm

HPM 20 0 <0.001 >45

HPM 20 5 0.018 15.1

HPM 20 10 0.027 16.0

Polargel 0 <0.001 >45

Polargel 10 0.063 26.0

Other suitable binders include borax, silica, alumina, alumina-silicates, sodium silicates, and mixtures of these in any proportion.

Preparation and Testing of Type 1 Geodes

A number of samples were fabricated using pre-formed 0.1" diameter spheroids of Sn0 ₂ as the inner core material. These spheroids were sized by passing them through a 4 X 8 mesh screen (nominal opening of 0.22" by 0.11") . These were pressed in a pellet die that is 0.125" in diameter. The crush strength and composition of these "Type 1 geodes" are shown below:

Sample Binder Graphite SnO,/ crush strength: Porosity No. % binder' l /mm cc/g Y1222-1 polargel 10 0.43 14.9 0.109

JV1222-2 polargel 20 0.42 9.1 0.136 Y1222-3 polargel 30 0.43 £.1 0.182

JY1222-5 polargel S> 10 0.5S 2.6 0.143 Silica Gel

" Weight of SnO, divided by the weight of binder.

It was found that the crush strength decreases with the increasing quantities of graphite, and the porosity increases with the increasing quantities of graphite. Compared to Y1222-1 (a polargel binder composition with 10% graphite), sample no. JY1222-5 had a higher porosity but a reduced crush strength, when some crushed Silica Gel was added to the binder.

Preparation and Testing of Type 2 Geodes

Several geodes were prepared and tested for 10 complete absorption and regeneration cycles and shown to have high H ₂ S absorption capacity, high porosity, and regenerate to the original stannic oxide. Three geodes, identified as JY0121-1, Y0121-3, and Y0121-5 demonstrated significant promise. The procedures to prepare the geodes are described in the following paragraphs.

The materials required to prepare the above Type 2 geodes are: graphite powder, an oil mix (kerosene mixed with or without a surfactant, mixed in the oil) , talc, and the high surface area chemical in reactive inner core material (stannic oxide) or chemical precursor (stannous sulfate) . Several different batches of the same formulation were prepared. The batches were given

different prefixes to identify the batch i.e., JY0121-1 and BW0121-1 have the same formulation but were prepared by different persons.

All of the four above Type 2 geodes were prepared with the same mixture of stannic oxide and kerosene (mixed with surfactant) . The following discussions on the preparation of the stannic oxide mix for the inner core and the preparation of each geode, including the shell, describe the procedures given the relative proportions of the constituents. To prepare the stannic oxide and oil mixture for the inner core, mix 3 kg {6 .6 lbs) of stannic oxide were mixed with 0.34 kg (0.75 lb) of the kerosene. The solid and liquid are mixed well to coat the stannic oxide with a layer of the organic coating.

The Y0121-1 is made by first mixing Polargel-T with graphite; 500 g of Polargel are mixed with 56 g of graphite making a 10% by weight mixture of two dry powders. 62 g of water are added, making a 10% by weight mixture with the mixed dry powders. To the 618 g of Polargel mixed with graphite and water, 1,235 g cf stannic oxide mixed with kerosene are added. Due to the presence of the graphite, the mixture will exhibit a predominantly black color but small areas of the white stannic oxide may be seen. The mixing continues until there is no further change in the appearance of the mixture. Water and Methocel™ are added to the composite mixture to help it extrude. When extruded (i.e., green), the water and kerosene do not mix and a phase separation occurs. The stannic oxide appears as a number of white spots in the black matrix of the graphite mixed with Polargel.

Drying and firing are very important to the integrity of the resulting geode. The green extrudates are dried at 125°C for 2 hours or until all water and kerosene have evaporated. The next step is to slowly oxidize the graphite at 600°C in air for 6 hours to insure that all of the graphite burns out. It was found that pellets containing carbon and fired directly at >900°C fractured due to rapid internal gas and heat generation by the reaction of carbon and oxygen (i.e., C + 0 ₂ -> CO + (l-a)C0 ₂ + heat); at temperatures over 900°C, and before oxidation is complete, the rapid generation of gas and heat destroys the pellet. After oxidation at 600°C, the extrudates are fired at 900°C; extrudates of the JY0121-1 had approximately equal properties upon firing for 2, 6, or 10 hours.

Y0121-3 and JY0121-5 geodes have the same formulations and preparation procedures through the green state but are fired at different temperatures, 900°C and 950°C respectively. No graphite was used in the preparation of this material, since the talc (3MgO*4Si0 ₂ *H ₂ 0) produces sufficient porosity due to the water contained within the crystal. When fired, the talc produces 3 moles of MgSi0 ₃ with 1.0 mole of excess Si0 ₂ and this insures that all of the MgO is locked up with silica during regeneration. Equilibrium calculations show that no side reactions should be executed to occur in either absorption or regeneration with the MgSi0 ₃ or Si0 ₂ .

The preparation of these geodes with talc is as follows. To 500 g of talc, 56 g of water are added and well mixed giving 556 g of damp powder. To the talc plus water mix, 556 g of the stannic oxide mixed with kerosene

were added. The 1,112 kg mixture is pressed with added water and Methocel™ as required to extrude a good green pellet. When pressed, small regions of white stannic oxide forming the inner cores within the matrix of the slightly darker talc shell may be seen. The pellets are dried at 125°C for several hours or until the water and kerosene evaporates. The JY0121-3 pellets are fired at 900°C; JY0121-5 is fired at 950°C. No intermediate firing at 600°C is required with these formulations since no graphite is present.

Breakthrough Characteristics of SnO _? Sorbent

Breakthrough characteristics of a Type 2 geode were measured at high pressure and high temperature. This geode was identified as BW0121-1 and is made in the same manner as JY0121-1, but by a different person.

275.5 grams were tested. The temperature was controlled to 573°C inside the reactor near the inlet and the oven temperature was 727°C. The pressure was 130 psia. The breakthrough characteristic of this material is shown in Figure 2 for the 65 ^th absorption cycle with simulated KRW gases.

A very low H ₂ S concentration (78% of the H ₂ S was removed) was sustained for a very long period of time. During the test, H ₂ S levels at the inlet were measured twice: once after 180 minutes to 210 minutes, and again at the end of the test. The calculated sulfur loading was 8.4%. The data demonstrate that very high loadings can be achieved with this material in high temperature use, even after many previous cycles have been accumulated on the same batch of pellets.

Life Testin of Geode One Type 2 geode, BW0121-1, was cycled through

several absorption and regeneration cycles. The reactor contained 400 grams of the material and was tested both at high space velocity (3450 h" ¹ ) with simulated KRW gases and at lower space velocities (2000 h" ¹ ) with simulated Texaco gases, diluted with steam (lower space velocity and higher H ₂ S concentration) to load the sorbent to the maximum extent, simulating the inlet region of a fixed bed reactor. The inlet region of the reactor was simulated because it has been shown to have the highest sulfur loading, which creates the greatest expected potential for damage to the sorbent with cycling.

Cvclic Testing of BW0121-1

Periodically, the geode pellets were removed from the reactor and the mass of whole geode pellets remaining was weighed. Some whole geode pellets were removed and analyzed for the properties of the sorbent. The remaining geode pellets were again weighed before returning them to the reactor. The properties of the BW0121-1 with cycling are given below:

Cycle Number Porosity ¹ Crush Strength Mass Loss Rate cc of water/g lb/mm %/cycle ² of sorbent initial 0.150 4.6 -

2 - - 0.33 regenerated

10 0.147 5.3 0.35 regenerated

22 0.139 6.2 0.35 regenerated

36 0.138 4.3 ¹ 0.92 ¹ regenerated

43 - - 0.0082 regenerated

63 sulfided 0.136 5.7 0.469

*At end of 36" cycle, cooling of the reactor caused all of the geode pellets in the reactor to be flooded with water; the rapid cooling associated with a water quench would not be experienced in practice. This quench may have weakened some geode pellets or caused acturing with others. The weight of the material that was removed from the reactor is subtracted from that which was added to- calculate the loss of material. By dividing by the number of cycles the average loss per cycle is then calculated (i.e., the reciprocal of the cycle life) during each period.

We subtracted the mass of pellets that are used for sample taking and fitted a curve to the experimental data; Figure 3 illustrates the curve fit to the data. The average loss of material through 63 cycles is 0.476%/cycle implying a cycle life of 210 cycles.

The loss rates for a zinc titanate sorbent are also

shown in Figure 3; the data are for TRZ-21 as reported by Jung et al . ¹ . The zinc titanate sorbents were tested under simulated inlet conditions and Jung et al . report that the zinc titanate is rapidly destroyed under these ₅ conditions (2.5% per cycle) . By comparison, the stannic oxide sorbent losses were very low (0.48%/cycle) , which is most probably a direct result of the geode structure of this invention.

₁₀ These experiments tested geode pellets in conditions that are modelled after those expected at the inlet of a fixed bed reactor. In those conditions the sulfur loading should be the highest, creating the greatest changes in the chemical sorbent and the greatest τ_5 anticipated stresses on the sorbent. However, the bed average sulfur loading would be less than these test conditions, and the bed average stresses en the sorbent should be less, so that the predicted life of a whole bed would be greater than the life of the geode pellets, o which are at the inlet. Accordingly, with the geodes of this invention, over 300 cycles for the bed average conditions is anticipated.

Different Sources of Stannic Oxide 5 The source of the stannic oxide can affect both the cost and properties of the sorbent. By starting with a material that is partially refined and contains significant quantities of impurities, the cost of the

¹ Jung, D.Y., J.S. Kasεman, T.F. Leininger, J.K. Wolfenbarger, and P.P. Yang (1992) . "Integration and Testing of Hot Desulfurization and Entrained Flow Gasification for Power Generation Systems," Proceedings of the Twelfth Annual Gasification and Gas Stream Cleanup Systems Contractors Review Meeting, editors R.A. Johnson and S.C. Jain, Morgantown, West Virginia: U.S. Department of Energy, MorgantownEnergyTechnology Center, Report No. DOE/METC-92/6128.

stannic oxide (or other chemical sorbent) can be greatly reduced (by avoiding the costs for the purification) . The impact on the properties of the sorbent can also be favorable. A geode was fabricated in the same manner as BW0121-1 but using an impure source of stannic oxide (76% Sn0 ₂ by weight) . That geode had a strength of 19.7 lb/mm versus 4.6 lb/mm for the same geode made with a pure source of stannic oxide. Several different sorbents were then fabricated with that material and the properties of new and sulfided pellets are shown below:

Pellet Porositv: cc/g Crush Strength: ID. No. lb _f /mm

Initial Initial

Sulfide Sulfide

BC0121-1 0.101 not 19.7 not tested tested

LP 004 0.139 0.098 9.8 9.1

LP 005 0.154 0.126 5.7 4.8

LP 006 0.150 0.129 6.1 4.4

Testing of Stannic Oxide Extrudates Additional stannic oxide extrudates were prepared and tested; one of which was a mixed metal oxide supported structure, and another of which was incorporated in a type 2 geode of this invention. These extrudates were 1529-40 (a mixed metal oxide) and 1529- 48-0 (a geode) . The 1529-48-0 geode has the same composition as the BW0121-1 type 2 geode (and JY0121-1 type 2 geode) ; however, the extruded geodes have greater strength (9.4 lb/mm versus 4.6 lb/mm in the BW0121-1) and

higher porosity (0.168 cc/g versus 0.150 cc/g for the BW0121-1) . The following data present the properties of these extrudates in both the initial (new) state, and after the first sulfiding of the materials:

Pellet ID . No . Porositv: cc/g Crush Strength:

Ib _f /mrn

Initial ιϋ Initial HI

Sulfide Sulfide

1529-40 0 .131 0 .086 8 .5 8 . 8

1529-48 -0 0 . 166 0 .139 9 .4 7 . 9

1529-48-H2R Above 0 .161 Above K ₂ 6 . 9

(1529-48 -0 H ₂ H ₂ reduced reduced) reduced

The reduced strength and porosity are typical of the sulfided stannic oxide and other tests have shown that the material regains its strength when regenerated.

The testing of the sorbents has also used steam dilution to change the temperature and composition of the modelled coal gas stream. The un-diluted gas stream would reduce Sn0 ₂ to metallic tin (melting point of 232°C) , a liquid at the operating temperatures of an anticipated hot gas cleanup system {e.g. , 570°C) . The 1529-48-0 extruded geodes were tested with a gas composition, which simulated a low steam KRW gasifier. The gas was 57.5% N ₂ , 30.6% H ₂ and 11.9% C0 ₂ with no steam or CO or E ₂ S (calculating the reducing potential at 2.57 and equal to a 5% steam gasifier as reported by Gupta

(1991) ). These extrudates were exposed to the gas for 5 hours at 570°C. The pellets lost weight, by 17% since the extrudates are 67% stannic oxide by weight, the extrudates should theoretically lose 14% in forming metallic tin from the stannic oxide. Small pieces of metallic tin were observed on the surface on the extrudates. In the containing wire screen there were both small pieces of shinny metal (probably tin) and small pieces of inert materials, which had broken off from the extrudates; the greater than theoretical weight loss is most likely due to the loss of these pieces. Nevertheless, the extrudates were intact when removed, although some cracking was apparent. These H ₂ reduced extrudates (identified as 1529-48-H2R) were then sulfided. After sulfidation they were strong, porous, and had high sulfur loadings; again the bentonite sin¬ tered and had a surface area of 1.1 m ² /g.

Testing of Zinc Oxide Geode

A geode in zinc oxide was also prepared. The zinc oxide geodes used talc ³ as the inert binder and a mixture of ZnO and Sn0 ₂ as the . sorbent . This geode , identified as TZSN40-359 , contained 27. 1% ZnO, 8 . 67% Sn0 ₂ , 0 . 16% NiO

Gupta, R. P . ( 1991) . "Enhanced Durability of Desulfurization Sorbents for Fluidized-Bed Applications , " in Proceedings of the Eleventh Annual Gasification and Gas Stream Cleanup Systems Contractors Review Meeting, Morgantown, WV: U. S . Department of Energy, Morgantown Energy Technology Center, Report Number DOE/METC-91/6123 (CONF-9108116 ) .

Talc is 3MgO*4SiO _j *H ₂ 0 and forms 3MgSi0 ₃ + SiO- when fired at 900 °C . Talc makes an effective inert binder for SnO- . However, thermodynainic calculations indicate that MgSi0 ₃ is inert but that the Si0 ₂ will react and form ZnSi0 ₃ , a compound that does not react with H.S . The zinc silicate removes some of the capacity to remove H ^* -S and may have limited the conversion in this preliminary test .

enhancer, and 64.07% talc and other inerts. Theoretically, the geode could absorb up to 13.9% (wt.) S, if 100% of the sorbent is converted to the sulfide, of which 1.9% is due to the Sn0 ₂ and enhancer and 12.0% to ZnO. The geode was fired at 900°C (1,652°F), which is higher than most zinc oxide based sorbents. We are able to fire at lower temperatures. However, to decompose zinc sulfate, which will form during regeneration, the commercial sorbent will be heated to >.700°C at the end of each regeneration cycle. We fired at 200°C higher

{i . e. , 900°C) to assure that we had a sorbent that would not sinter during the high regeneration temperature.

The geode was exposed to simulated gases from a KRW gasifier containing 1.0% H ₂ S in H ₂ , CO, C0 ₂ , H ₂ 0, and N ₂ for 5 hours at 570°C. We measured the sulfur content of the sorbent after exposure to simulated KRW gases. On the first attempt, the zinc oxide geode absorbed 4.0% S (wt.) or 31.6% of theoretical. The crush strength was 3-5 lb/mm with a porosity of 0.16 cc (water) /g. We also tested the same geode at 500°C and 400°C; while the sulfur loadings were lower, sulfur was absorbed in both cases (3.3% and 1.0%) . The data demonstrate that a strong, porous, and chemically active zinc oxide geode can be made. Since previous work on zinc titanate has shown that zinc oxide can be converted virtually completely to ZnS, in the future we will add larger quantities of enhancers and can use titania, Ti0 ₂ , as the inert binder to avoid any problems of deactivating the sorbent with Si0 ₂ . Ti0 ₂ may also be mixed with the ZnO to form a zinc titanate sorbent, the addition of Ti0 ₂ stabilizes the ZnO, especially during high temperature operation (i.e., to minimize zinc loss by vaporization) .