Related Art As one result of concern with air pollution, environmental regulators have severely reduced and are continuing to reduce allowable emissions of sulfur oxides and hydrogen sulfide. As a result, a variety of technologies have been developed and are continuing to be developed for use in flue gas desulfurization (FGD). Internal combustion engines are one source of the emission of such sulfur compounds because of sulfur-containing compounds in fuel.

It is highly desirable to effectively remove sulfur compounds at process conditions in a selective manner, i. e., remove sulfur species without interference from organic or other combustion species, and with a physically durable system with low resistance to flow. Such technology would be useful for removing sulfur compounds from the exhaust gas recirculation loop of mobile combustion engines and the exhaust of mobile and stationary combustion engines.

Requirements for NOx reduction from diesel engine exhaust can be met by using exhaust gas recirculation (EGR). Unfortunately, the high levels of sulfur to be recycled can increase the rate of corrosion of engine components. It would therefore be desirable to remove the sulfur containing compounds.

Current technologies capable of removing sulfur compounds from combustion exhaust include the use beds of activated carbon, potassium permanganate, wet scrubbers or dry scrubbers. Each of these technologies has limitations, not the least of which are enumerated below : 1. Activated or impregnated carbon cannot be used at high temperatures (>300°F) in an oxidizing environment due to carbon combustibility. Carbon is a soft, weak material with poor durability.

2. Potassium permanganate is effective at removing sulfur compounds but has poor selectivity and is consumed by other exhaust constituents that can be oxidized, such as carbon monoxide and unburned hydrocarbons.

3. Wet scrubbers typically use hydrogen peroxide, ozone and/or caustic solution. The exhaust gas must be cooled to present compatibility with aqueous scrubbing solutions. The process must be carefully controlled to maintain temperature stability and chemical make up of scrubber solutions.

4. Dry scrubbing of flue gas involves contacting the effluent with a solid material which chemically reacts with the sulfur component and forms a compound. The system can be a fixed bed such as zinc oxide pellets that are used to react with H2S to form zinc sulfide. The zinc sulfide pellets must be removed and replaced after saturation. The dry scrubber can also be in powder or particulate injected into the stream followed by a baghouse or electrostatic filtration to remove the reacted product. An example of powder would be limestone, which would react after injection into the exhaust stream to form calcium sulfate and/or calcium sulfite hemihydrate sludge.

The material is typically delivered as a wet slurry, which dries in the exhaust and reacts with the sulfur oxides. This material is then removed and typically disposed of in a landfill.

Campbell et al. in U. S. Pat. No. 5,451,558 describes a regenerable catalyst/sorber using precious metal oxidation catalysts combined with alkali metal oxides, hydroxides, and carbonates to oxidize and capture by absorption the oxides of nitrogen and sulfur. Further patents by Campbell et al. such as U. S. Pat No. 5,650,127, describe ways to capture the oxides of nitrogen without the deleterious effects of capturing sulfur compounds. Campbell et al U. S. Pat. No 6,037,307 describes a regenerable system in which sulfur compounds are oxidized and captured by chemisorption and then regenerated in situ to drive off a higher concentration of gaseous sulfur compounds (SO2, H2S). In these processes it is disclosed that materials must be periodically regenerated.

The present invention focuses on maximizing the sulfur species capture capacity resulting in a system which efficiently removes sulfur compounds from an exhaust until the system is depleted. The system would then be removed and replaced with subsequent off-line recovery or regeneration of sulfur. The technical

targets of such a system are different from the prior art. First, there is no need for in situ regeneration. Second, reusability of the system is desirable but not required.

These considerations open opportunities in system design. For example, oxidation catalysts and sorbents need not withstand the rigors of oxidation and reduction. Solid state reactions between sorbent and catalyst supports which would have damaged performance caused by cyclic oxidation-reduction and temperature are no longer damaging but can be beneficial for increased capacity.

The use of an oxidation catalyst/sorber catalyst in an EGR system is not disclosed in the prior art.

SUMMARY OF THE INVENTION Briefly, the present invention is the use of an oxidation catalyst/sorber in a gas exhaust recirculation (EGR) system or loop of an internal combustion engine.

The oxidation catalyst/sorber is positioned within the EGR loop in a form to allow rapid circulation through and in contact with the oxidation catalyst/sorber. The oxidation catalyst is preferably deposited on a high support area which is preferably intimately associated with the sorber. The oxidation catalyst/sorber is preferably deposited on a carrier.

BRIEF DESCRIPTION OF THE DRAWING The figure is a schematic diagram of an EGR loop using an oxidation catalyst/sorber to remove SOx from exhaust gas.

DETAILED DESCRIPTION In a preferred embodiment of the present invention, the oxidation catalyst component is dispersed on a high surface area support. An inorganic compound which will react with SOx compound, preferably inorganic compounds which will react or combine completely with SO2 and/or S03 to form sulfites or sulfates, is added to the catalyst component. In a more preferred embodiment, the inorganic compounds are dispersed uniformly and in close proximity to the catalyst component and more preferably the two are intimately associated, preferably by coating the sorber over the oxidation catalyst or supported oxidation catalyst. The oxidation catalyst/sorber can be coated onto a ceramic or metal substrate or formed, for example, into a pellet, sphere or polylobe. The significance of the shape is to optimize gas contact.

The inorganic compounds may be deposited in micropores on the high surface area support. During reaction with sulfur oxides, it is highly desirable that relatively large volumes of sulfur oxides be reacted and also that the products of reaction not expand to constrict the porosity or cause stress and breakage of the high surface area support. Table 1 list several examples of inorganic compounds which are suitable for the present invention.

Table 1 Inorganic Compounds Useful for Reacting with Oxides of Sulfur SORBER MW MW Density Density Density Density Sulfate SO2 % SO2 Carbonate Sulfate Oxide Hydroxide Li2CO4 109.95 64 58.2% 2.110 2.210 2.013 1.450 Na2CO4 142.04 64 45.1 % 2.540 2.700 2.270 2.130 K2CO4 174.26 64 36.7% 2.290 2.660 2.350 2.044 MgCO4 120.37 64 53.2% 3.050 2.660 3.600 2.370 CaCO4 136.14 64 47.0% 2.830 2.960 3.340 2.200 SrCO4 183. 68 64 34.8% 3.500 3.960 5.100 3.625 BaC04 233.39 64 27.4% 4.287 4.490 5.720 3.743 Fe2 (CO4) 3 399.88 192 48.0%----3.100 5.250 3.120 Precursor Precursor Precursor Precursor MW MW Volume Volume Oxide Hydroxide SORBER Sulfate SO2 % SO2 Carbonate Sulfate Li2co4 109.95 64 58.2% 0.474 0.452 0.497 1.379 Na2CO4 142.04 64 45.1 % 0.394 0.370 0.441 0.939 K2CO4 174.26 64 36.7% 0.437 0.376 0.426 0.978 MgCO4 120.37 64 53.2% 0.328 0.376 0.278 0.422 CaCO4 136.14 64 47.0% 0.353 0.338 0.299 0.455 SrCO4 183.68 64 34.8% 0.286 0.253 0.196 0.276 BaC04 233.39 64 27.4% 0.286 0.223 0.175 0.267 Fe2 (CO4) 3 399.88 192 48.0%----0.323 0.190 0.641 The oxidation catalysts may be any of those known in the art to oxidize organic sulfur compounds to the oxides, such as those disclosed in U. S. Pat. No.

5,451,558, which is incorporated herein in its entirety and U. S. Pat. No. 6,037,307, which is incorporated herein in its entirety. Suitable oxidation catalysts include the noble metals, e. g., Pt, Pd, Rh, Ru, Ni, Fe, Cu, Mo, Co, Mn and mixtures and

combinations thereof. The oxidation catalysts may be prepared in any manner known in the prior art. The noble metals, particularly Pt are preferred.

The oxidation catalyst is preferably deposited on a high surface area support. The high surface area supports are typically alumina, silica, silica- alumina, titania or zirconia, and combinations thereof. Alumina is a preferred high surface area support. Generally, the surface area of high surface area supports is in the range of 40 to 350 m2/g, preferably 100 to 325 m2/g and more preferably 200 to 300 m2/g.

The sorber material may be any material which will react with or otherwise combine with SOx (SO2 and S03) to form stable compositions under the condition of the EGR loop. Included among these materials are the alkali metals, alkaline earth metals, Ti, Zr, Hf, Ce, Al, Si, Fe and mixtures and combinations thereof.

Preferred sorbers include Li, Na, K, Mg, Ca, Sr, Ba, Fe and mixtures and combinations thereof. The sorber may be supplied to the oxidation catalyst as a variety of precursor, but is preferably converted to or deposited as the oxide, hydroxide or carbonate for use.

The supported oxidation catalyst/sorber may be extruded in a suitable form such as beads, spheres, loops, polylobes and the like or applied to a carrier. The carrier may be honeycombed ceramic, or expanded (foamed) metal generally known as monoliths. The carrier is preferably a macroporous material having from 64 to 600 cells (pores) per square such (CPS) which is about 25 to 80 pores per linear inch (ppi) although lower pore count materials are suitable.

The oxidation catalyst comprises from 0.25 to 50 wt%, preferably 0.5 to 20 wt% of the combination of oxidation catalyst and high surface area support. The sorbent comprises 1 to 50 wt%, preferably 10 to 30 wt% of the combination of catalyst, sorbent and high surface area support. When the oxidation catalyst/sorber is applied to a carrier, it comprises 1 to 50 wt %, preferably 10 to 30 wt% of the combination of oxidation catalyst/sorber and carrier.

The rate of flow of the gas in EGR loop will depend on many factors such as the sulfur compound content in the gas, the effectiveness of the oxidation catalyst/sorber, the temperature of the contacting, degree of contact between the gas and the oxidation catalyst/sorber to list a few. However, the preferred rate of contact is at a GHSV (gas hourly space velocity) in the range of 5,000 to150, 000

hr~1, more preferably 20,000 to 60,000 hr', at temperatures in the range of 300 to 1000°F, more preferably 750 to 900°F.

Referring now to the figure, the internal combustion engine 10 is shown attached via exhaust 12 to controller 18 of the type known in the art to selectively recycle a portion of the treated exhaust via line 24 to an exhaust treatment zone 14 in which two oxidation catalyst/sorber units are disposed and hence via line22 to the engine for use in the combustion chamber as appropriate and to direct a portion into exit line 20. The arrows indicate the direction of gas flow. The units in this embodiment are ceramic monoliths coated with the oxidation catalyst/sorber.

There may be only one such unit or two or more. The composition of the oxidation catalyst/sorber on each unit may be the same or different. Furthermore, the composition of oxidation catalyst/sorber may vary within a single unit.

The treated exhaust gas in exit line 20 may be passed to the atmosphere or to other treatment zones, such as conventional catalytic converts and/or additional oxidation catalyst/sorber treatment zones (not shown). Moreover, two or more exhaust treatment zones may be arranged in parallel for alternate use when the oxidation catalyst/sorber is totally used up and SOx occurs.

In the preferred embodiment the high surface area material is sacrificed to sulfate formation after the inorganic precursor is consumed. The latent reactivity toward sulfur compounds is critical to achieving high sulfur oxide removal capacity.

Example 1 A washcoat solution of Al203/Pt is prepared. This is accomplished by the incipient method technique. This technique involves the addition of an aqueous platinum solution to the dry Al203 until reaching incipient wetness. The concentration of the aqueous platinum solution was adjusted to produce 1.6% platinum in the dried Al203/Pt. This solid is then calcined at 500°C for 1 hour.

After cooling to room temperature, 7% acetic acid is added to the Al203/Pt. After ball mill overnight, the resulting washcoat solution contains 45% A1203/pt with a particle size between 3 and 4 um.

A 400 cpi honeycomb cordierite carrier is then dipped into the washcoat solution. After excess washcoat is blown off, the sample is dried at 150°C for 30 minutes and then calcined at 500°C for 1 hour. The washcoat loading is determined to be 2.5g/in3.

A 45% lithium nitrate solution is prepared. The Al203/PVhoneycomb sample is then dipped into the 45% lithium nitrate. After excess solution is blown off, the sample is dried at 150 °C for 30 minutes. The sample is then exposed to ammonia and carbon dioxide to convert the lithium nitrate to lithium carbonate, which has greater reactivity.

The oxidation catalyst/sorber is loaded into a laboratory reactor. The laboratory reactor simulates the exhaust gas of an internal combustion engine and consists of a sample tube furnace, compresses gas cylinders, steam tube furnace, flow controllers, flow traducers, chiller, stainless steel tubing, heating tape, and a Bovar sulfur dioxide analyzer interfaced to a PC for data acquisition.

The test conditions consist of a gas hourly space velocity of 10, 000/hr, 550°F, 200 ppm SO2, 12% 02, 5% H20, and 8% CO2. The higher than typical sulfur dioxide concentration, 200 ppm, is utilized to accelerate the test. Previous work has demonstrated that this modification to simulated conditions has no effect on capacity or performance. The data is summarized in Table 2 and demonstrates that the process efficiently removes the sulfur dioxide for 260 minutes. In the vicinity of 260 minutes, the sulfur dioxide begins to breakthrough.

Table 2.

Time (Minutes) SO2 (ppm) 020 000 044 000 060 000 080 000 100 000 120 000 140 000 160 000 180 000 200 000 220 000 240 000 260 3.66 280 8.54

Table 2 Continued 300 17.09 320 25.63 EXAMPLE 2 The carrier is 400 cpi honeycomb cordierite. A washcoat solution is prepared consisting primarily of high surface area alumina and deionized water.

The alumina washcoat is wet milled overnight to obtain a particle size between 3 and 4 microns. The washcoat is then analyzed for solids content and diluted to about 45% solids content.

The 400 cpi honeycomb is dipped into the alumina washcoat. The excess washcoat is blown from the cells with compressed air. The 400 cpi piece is dried at 150°C for 30 minutes, then calcined at 500°C for 1 hour. After calcination the alumina washcoat loading is determined to be 2.3 grams/cubic inch.

A 26% Co (NO3) 2 solution is prepared by dissolving 20.88 grams of Co (NO3) 2 6H2O into 29.12 grams of deionized water. The 400 cpi sample is then dipped into the 26% Co (NO3) solution. After the excess solution is blown from the cells, the sample is placed into a 150°C furnace for 20 minutes. Then the sample is calcined at 500°C for 1 hour. After calcination the elemental Co loading is determined to be 0.25 g/in. 3 present as Co203.

A 45% LiNO3 solution is prepared. The AI203/Co203/honeycomb sample is then dipped into the 45% LiNO3 solution. After the solution is blown from the cells, the sample is dried at 150°C for 30 minutes. The sample is then exposed to a gas stream containing anhydrous ammonia and carbon dioxide.

The catalyst/sorber is loaded into a laboratory reactor. The reactor simulated the exhaust gas of an internal combustion engine and consists of a sample tube furnace, compressed gas cylinders, steam tube furnace, flow controllers, flow transducers, chiller, stainless steel tubing, heating tape, and a BOVAR sulfur dioxide analyzer which interfaces to a PC for data acquisition which is reported in Table 3.

The test conditions consist of a space velocity of 50,000/hr, 1000°F, 200 ppmv SO2, 12% 02, 8% CO2, and 5% H20. The SO2 concentration of 200 ppmv is used to accelerate the test. Previous work has demonstrated that this modification to simulated conditions has no effect on capacity or performance. The data below

demonstrates that the base metal sulfur trap adsorbs 100% of the SOZ initially, but the removal efficiency slowly decreases after about 4 minutes exposure to the 200 ppmv SO2 at 50,000/hr space velocity. The base metal trap with Co203 converts only a minute amount of SO2 to S03, which can and does erode stainless steel components from the reaction of S03 with H20 to form sulfuric acid.

TABLE 3 Time (minutes) SO2 (ppmv) 2 0 5 7.32 10 41.5 15 101.32 20 119.63 30 137.94 40 144.04 50 147.71 60 157.47 80 164.80 100 168.46 120 173.34 140 180.66 160 183. 11 180 189.21 200 190.43 220 194.09 240 192.87 260 198.98 Example 3 A catalyst/sorber sulfur trap was prepared with the catalyst and sorber components on two separate carriers.

An alumina washcoat is prepared by adding enough alumina to an acidic solution to contain 45% solids. The washcoat is wet milled overnight until the particle size measured 3 to 4 microns in diameter. A 400 cpi ceramic honeycomb carrier is then dipped into the alumina washcoat. The excess washcoat is blown

from the cells with compressed air. Once the cells are clear, the sample is placed into a 150°C furnace for 30 minutes and calcined at 500°C for 1 hour. The alumina loading is determined to be 2.2 gr/in3.

After the sample has calcined, the sample is dipped into a platinum solution.

The concentration of the solution is adjusted to give a platinum loading equivalent to 1.66% of the alumina weight. The excess solution is blown from the cells. The sample is then dried at 150°C for 30 minutes and calcined at 500°C for I hour.

The sorber is prepared on a second 400 cpi honeycomb carrier by dipping the carrier in the 45% alumina washcoat. The excess washcoat is blown from the cells. After the excess washcoat is removed, the sample is placed in the furnace to dry at 150°C for 30 minutes and calcined at 500°C for I hour. The washcoat loading is determined to be 2,46 g/in3. The lithium sorber coating is dispersed on top of the alumina. A 45 % lithium nitrate solution is prepared and the sample placed in the solution for 10 minutes. After 10 minutes, the sample is removed from the solution and the excess blown from the cells. The sample is heated at 150°C for 30 minutes. The carrier is wrapped with a high-temp aluminosilicate rope and placed into a stainless steel tube. A gas containing anhydrous ammonia and carbon dioxide is passed over the sample for about 10 minutes to precipitate the carbonate of lithium. The carrier is dried again at 150°C for 30 minutes and calcined at 500°C for 1 hour. The carrier contains a lithium loading of about 1.4% of the total sorber/washcoat/carrier weight.

The dual catalysvsorber samples are placed inside a stainless steel tube reactor with the oxidation catalyst placed in front, upstream of the sorber trap. The sample is subjected to a gas stream which closely simulates an exhaust gas stream of an internal combustion engine. The gas mixture contains 12% 02, 8% C02, and 5% H20. The level Of S02 is 154 ppmv, which accelerates the point at which 100% breakthrough occurrs. Previous data showed there is no decrease in capacity or adsorption efficiency by increasing S02 levels. The gas space velocity is 50,000/hr and the temperature is1000°F. The S02 level is monitored using a BOVAR model 721-M Photometric S02 analyzer and the analyzer interfaced to a PC for data acquisition.

The data in Table 4 demonstrates the S02 capacity achievable by the dual catalyst/sorber design.

Table 4 Time S02 (minutes) (ppmv) 0 154 2 7.32 4 4.88 6 6.1 8 7. 32 10 8.55 20 17.09 30 25.64 40 32.96 50 41.5 60 50.05 70 54.93 80 62.26 90 69.58 The data shows that the dual catalyst/sorber design is very efficient for selective removal of sulfur dioxide in a gas stream laden with various other pollutants like CO, NOx, and hydrocarbons. The % S02 removal efficiency is 89.5% and the trap removes SOa equal to 10 times its own volume for the entire 90 minute experiment. The trap is still performing at 50% efficiency at the end of run.