Process for removing hydrogen sulfide from gas streams which include or are supplemented with sulfur dioxide

[0002] Applicant claims the benefit of its U.S. Provisional Application numbers 60/120,390 filed Feb. 17, 1999 and 60/156,545 filed Sep. 29, 1999.

[0003] This invention relates generally to processes and systems for removing hydrogen sulfide from a gaseous stream. More specifically the invention relates to improvements in a known process and system wherein hydrogen sulfide is removed from a gaseous stream, using a nonaqueous scrubbing liquor in which are dissolved sulfur and a reaction-promoting amine base. In a first aspect of the invention sulfur dioxide is added as an oxidizing gas to the sulfur-amine nonaqueous sorbent (or advantage is taken of SO₂which may already be present in the gas stream) to obtain better H₂S removal, lower chemical degradation rates, and lower rates of formation of byproduct sulfur salts. In a further aspect of the invention the gas to be treated is mixed with oxygen and passed through an oxidation catalyst reactor to either effect oxidation of part of the H₂S to form the required amount SO₂for reaction with the remaining H₂S, or to effect partial oxidation of the H₂S in the feed gas to form elemental sulfur, or to form various combinations of products as desired for the application, prior to scrubbing with the nonaqueous solvent.

[0004] Conventional liquid redox sulfur recovery processes use a redox couple dissolved in water to scrub hydrogen sulfide from a gas stream and convert it to sulfur. The redox agent is reduced by the hydrogen sulfide and then is regenerated by contacting with air in a separate vessel. One of the main problems with such processes is dealing with the solid sulfur product, which is formed in an uncontrolled manner. The sulfur formed from aqueous solution is notorious for plugging the absorber or other vessels which is passes through, and it is generally hard to separate and handle. Sulfur formed from nonaqueous solvents has much better handling properties. However, most nonaqueous redox systems have certain disadvantages such as sluggish sulfur formation kinetics or difficulties in regenerating the sorbent with air. In aqueous systems, contact of polysulfides with air primarily produces sulfates and other undesired sulfur oxyanion byproducts which are difficult to purge from the system.

[0005] The present inventor's U.S. Pat. No. 5,738,834, the entire disclosure of which is hereby incorporated by reference, discloses a process which uses a sulfur-amine nonaqueous sorbent (SANS) and operating conditions under which sulfur itself can convert hydrogen sulfide to polysulfides which are nonvolatile but which can be readily transformed to sulfur by reaction with an oxidizing agent. This is done in a solvent with a high solubility for sulfur so that solid sulfur formation does not occur in the absorber or in the air-sparged regenerator. Solid sulfur formation can be initiated in process equipment designed to handle solids and can be done under well-controlled conditions. In the SANS process, the sour gas is fed to an absorber (typically countercurrent) where the H₂S is removed from the gas by a nonaqueous liquid sorbing liquor which comprises an organic solvent for elemental sulfur, dissolved elemental sulfur, an organic base which drives the reaction converting H₂S sorbed by the liquor to a nonvolatile polysulfide which is soluble in the sorbing liquor, and an organic solubilizing agent which prevents the formation of polysulfide oil—which can tend to separate into a separate viscous liquid layer if allowed to form. The solubilizing agent is typically selected from the group consisting of aromatic alcohols and ethers including alkylarylpolyether alcohol, benzyl alcohol, phenethyl alcohol, 1-phenoxy-2-propanol, 2-phenoxyethanol, alkyl ethers including tri(propylene glycol) butyl ether, tri(propylene glycol) methyl ether, di(ethylene glycol) methyl ether, tri(ethylene glycol) dimethyl ether, benzhydrol, glycols such as tri(ethylene) glycol, and other polar organic compounds including sulfolane, propylene carbonte, and tributyl phosphate, and mixtures thereof. The sorbing liquor is preferably essentially water insoluble as this offers advantages where water may be condensed in the process. It is also preferable for water to be essentially insoluble in the solvent. The nonaqueous solvent is typically selected from the group consisting of alkyl-substituted naphthalenes, diaryl alkanes including phenylxylyl ethanes such as phenyl-o-xylylethane, phenyl tolyl ethanes, phenyl naphthyl ethanes, phenyl aryl alkanes, dibenzyl ether, diphenyl ether, partially hydrogenated terphenyls, partially hydrogenated diphenyl ethanes, partially hydrogenated naphthalenes, and mixtures thereof. In order to obtain a measurable conversion of sulfur and hydrogen sulfide to polysulfides, the base added to the solvent must be sufficiently strong and have sufficient concentration to drive the reaction of sulfur and hydrogen sulfide to form polysulfides. Most tertiary amines are suitable bases for this use. More particularly, tertiary amines including N,N imethyloctylamine, N,N dimethyldecylamine, N,N dimethyldodecylamine, N,N dimethyltetradecylamine, N,N dimethylhexadecylamine, N-methyldicyclohexylamine, tri-n-butylamine, tetrabutylhexamethylenediamine, N-ethylpiperidine hexyl ether, 1-piperidineethanol, N-methyldiethanolamine, 2-(dibutylamino)ethanol, and mixtures thereof are suitable for use in the said process. It should be noted that while the solvent utilized in the process requires the addition of a base to promote the reaction of sulfur and hydrogen sulfide to form polysulfides, the base and the solvent may be the same compound.

[0006] As it is removed, the H₂S thus reacts with elemental sulfur and a tertiary amine, both dissolved in the sorbent, to form an amine polysulfide. One of the polysulfide-formation reactions in the absorber may be depicted as follows (where B stands for the amine, HB⁺is the protonated amine, g denotes the gas phase, and l denotes the liquid phase).

[0007] H₂S(g)+S₈(l)+2 B(l)⇄(HB)₂S₉(l)   (1)

[0008] The stoichiometry shown in this equation is representative, although polysulfides of other chain lengths may be formed, and varying degrees of association of the amine and polysulfide may occur, depending on the specific solvent chemistry and operating conditions. The primary solvent is selected to have a high solubility for sulfur (as well as for the amine) so that the sorbent circulation rates can be low, producing small equipment sizes for both the H₂S absorber and the solution regenerator. Another ingredient is normally added to the sorbent to solubilize the amine polysulfides which might otherwise separate. The sweet gas from the absorber exits the process. The rich sorbent from the absorber may be passed through a reactor to allow further time for polysulfide formation reactions to occur, if desired. The sorbent is flashed down to near atmospheric pressure in one or more stages, producing a small flash gas stream that can either be recycled or used as fuel for local power generation. The sorbent is then contacted with an oxidizing gas such as air in the regenerator to oxidize the polysulfide to elemental sulfur, which remains dissolved in the solvent. This reaction, which also frees the amine for the next sorption cycle, can be depicted as follows.

[0009] (HB)₂S₉(l)+½ O₂(g)⇄S₈(l)−H₂O(g)+2 B(l)   (2)

[0010] Under the proper chemical and physical conditions, the efficiencies for simple air regeneration are unexpectedly high and the rates of the air oxidation reaction to sulfur are unexpectedly fast for a nonaqueous system. Tertiary amines produce high regeneration efficiencies. Spent air from the oxidizer contains the product water. The sorbent stream from the oxidizer is cooled in a heat exchanger and fed to the crystallizer where the cooling causes the formation of crystalline sulfur. The sorbent is cooled to a sufficiently low temperature to crystallize enough solid sulfur to balance the amount of hydrogen sulfide absorbed in he absorber. This produces the same overall reaction as in other liquid redox sulfur recovery process.

[0011] H₂S(g)+½ O₂⇄⅛ S₈(s)+H₂O(g)   (3)

[0012] The solvent generally can have a solubility for sulfur in the range of from about 0.05 to 2.5, and in some instances as high as 3.0 g-moles of sulfur per liter of solution. The temperature of the nonaqueous solvent material is preferably in the range of about 15° C. to 70° C. Sulfur formation is obtained, when desired, by cooling the liquor proceeding from the air-sparged regenerator. This can for example be effected at a sulfur recovery station by cooling means present at the station. The solvent is thereby cooled to a sufficiently low temperature to crystallize enough solid sulfur to balance the amount of hydrogen sulfide absorbed in the absorber. The solubility of elemental sulfur increases with increasing temperature in many organic solvents. The rate of change of solubility with temperature is similar for many solvents, but the absolute solubility of sulfur varies greatly from solvent to solvent. The temperature change necessary to operate the process will vary primarily with the composition of the sorbent, the flow rate of sorbent, and the operating characteristics of the recovery station. For most applications, a temperature difference of 5° C. to 20° C. is appropriate as between the temperature of the solvent material at the absorber/reactor and temperature to which the said solvent is cooled at the sulfur recovery station; The regenerated sorbent from the crystallizer is recycled back to the absorber. The slurry of crystalline sulfur from the crystallizer is thickened and fed to a filter that produces a filter cake of elemental sulfur for disposal or sale.

[0013] The reaction between sulfur and H₂S to form polysulfide is chemically reversible and the physical solubility of H₂S in the sorbent is high. The equilibria are such that at low inlet H₂S concentrations it becomes difficult to achieve high H₂S removals at acceptably low liquid flow rates due to the “back-pressure” of H₂S.

[0014] Now in accordance with a first aspect of the present invention, it has been found that in the process of the 5,738,834 patent, the addition of SO₂to the absorber produces a more complete chemical conversion of H₂S thus reducing the equilibrium back-pressure of H₂S and allowing much better removals to be obtained. The SO₂is thereby used as an or the oxidizing gas referred to in the 5,738,834 patent. A major reaction appears to be between SO₂and the amine polysulfide formed in the initial SANS reaction:

[0015] 2 (HB)₂S₉(l)+SO₂⇄19/8 S₈(l)+2 H₂O(g)+4 B(l)   (4)

[0016] Combining this with equation 1 gives the overall reaction:

[0017] 2 H₂S(g)+SO₂(g)⇄⅜ S₈(l)+2 H₂O(g)   (5)

[0018] This is the same as the well-known Claus reaction which is usually practiced in the gas phase at elevated temperatures using a catalyst. The gas phase Claus process is highly exothermic and equilibrium limited in the temperature range where it is normally practiced. Many attempts have been made at devising a liquid phase Claus reaction-based H₂S removal process to circumvent the equilibrium limitation. Most of these attempts have been plagued by the sluggish kinetics of the direct low temperature reaction of H₂S and SO₂and by formation of intractable mixtures of polythionates (Wackenroder's solution) and other undesirable byproducts. Use in the present invention of a hydrophobic solvent having a high sulfur dissolving ability removes these defects. The high concentration of sulfur promotes the formation of polysulfide when reacted with H₂S and the polysulfide reacts rapidly and completely with SO₂.

[0019] The process can be operated with an excess of H₂S over the classical “Claus” mole ratio of 2 mole H₂S per mole SO₂where air is used as an oxidant in addition to SO₂. This is illustrated by Example 1 (see below) which used an input mole ratio of 3 mole H₂S per mole SO₂. Under these conditions, reactions (2) and (4) apparently operate in parallel to form the product sulfur. During the 44 hour run under these conditions, the sulfur oxyanion byproduct make rate was only 1.92% (total moles sulfur in oxyanion byproducts per mole sulfur absorbed). This is somewhat lower than the value usually obtained in the SANS process in the absence of SO₂, and illustrates that the conventional view that mixtures of sulfur and SO₂necessarily produce a lot of thiosulfate does not apply to the process conditions described here. A significantly lower than normal amine degradation rate of 0.28% (mole amine degraded per mole sulfur absorbed) was also observed in this run. This is to be compared with the value of about 1.0% (mole amine degraded per mole sulfur absorbed) previously observed in runs done with no SO₂. This increase in amine stability could be due to the oxidation of polysulfide with SO₂being less harsh and generating less reactive oxidizing intermediates than the oxidation of polysulfide with oxygen.

[0020] When operated without any air regeneration, the SO₂-enhanced SANS process operates at a nominal H₂S:SO₂mole ratio of 2:1 in accordance with the usual overall Claus reaction. However the sorbent composition, particularly the high elemental sulfur concentration, provides a buffering effect which allows extended operation under “off-ratio” inlet gas conditions. This is important for obtaining stable operation and for meeting environmental regulations which may limit short term peak emissions. During episodes of higher than stoichiometric addition of SO₂, the high elemental sulfur concentration in the solution allows the formation of thiosulfate.

[0021] SO₂(l)+⅛ S₈(l)+2 B(l)+H₂O→(HB)₂SO₂O₃(l)   (6)

[0022] which can later be converted to product sulfur by reaction with H₂S or a polysulfide.

[0023] (HB)₂S₂O₃(l)+2 H₂S(l)→½ S₈(l)+2 B(l)+3 H₂O   (7)

[0024] (HB)₂S₂O₃(l)+2 (HB)₂S₉(l)→5/2 S₈(l)+6 B(l)+3 H₂O   (8)

[0025] Thus the elemental sulfur serves as a redox buffer for both of the main reactants, reacting either with excess H₂S to form polysulfide, or with excess SO₂to form thiosulfate. Importantly, the products of these reactions can then be efficiently converted to elemental sulfur when the H₂S:SO₂ratio swings back the other way, and byproduct formation is minimal.

[0026] The process of the foregoing aspect of the present invention can be applied to many different sulfur recovery situations with the addition of SO₂being accomplished by several possible means including (1) obtaining SO₂as a gas or liquid from an independent source and injecting it into the inlet gas or absorber; (2) burning product sulfur and injecting the resultant SO₂as in the (1); and (3) converting a portion of the H₂S or other sulfur species in the inlet gas to SO₂by passing a portion of the inlet gas stream along with air or oxygen through a catalyst bed or other device which will provide the desired amount of SO₂.

[0027] To summarize, one major advantage of the aforementioned aspect of the invention is that better H₂S removal can be obtained if SO₂is present in the scrubbing liquor or in the inlet due to the more complete chemical reaction of H₂S in the absorber. Another advantage is that the H₂S can conveniently be converted to elemental sulfur in the absorbing vessel, which may be at high pressure, thus minimizing the flashing of H₂S when the liquid is “flashed” down to the lower pressure. This property of conversion to sulfur in one vessel also provides the opportunity to maintain the circulating liquid at operating pressure, thus eliminating the flashing of volatile hydrocarbons, as can happen if the liquid pressure is reduced prior to regeneration in an oxidizer operating near ambient pressure. Yet another advantage is the decrease in degradation of amine, apparently due to the less harsh oxidant conditions provided by SO₂compared to oxygen.

[0028] In a further aspect of the invention the gas to be treated is mixed with oxygen and passed through an oxidation catalyst reactor to either effect oxidation of part of the H₂S to form the required amount SO₂for reaction with the remaining H₂S, or to effect partial oxidation of the H₂S in the feed gas to form elemental sulfur, or to form various combinations of products as desired for the application, prior to scrubbing with the nonaqueous solvent.

[0029] In the drawings appended hereto:

[0030] FIG. 1 is a schematic block diagram of a test apparatus operating in accordance with a first embodiment of the present invention;

[0031] FIG. 2 is a more generalized schematic block diagram depicting the same embodiment in which the H₂S and SO₂are already in the correct ratio;

[0032] FIG. 3 is a schematic block diagram depicting a second embodiment of the invention, in which full oxidation of a sidestream of the input H₂S-laden gas is carried out in a sidestream upstream of the absorber followed by combining with the main inlet stream;

[0033] FIG. 4 is a schematic block diagram depicting another embodiment of the invention, in which both partial and full oxidation of the input H₂S are carried out on a sidestream of the inlet gas upstream of the absorber followed by combining with the main inlet stream and treatment in the absorber;

[0034] FIG. 5 is a graph illustrating the effects of SO₂, extinction on H₂S removal under various laboratory conditions;

[0035] FIG. 6 is a graph illustrating the effects of operating the byproducts removal feature in a system in accordance with the invention; and

[0036] FIG. 7 is a graph showing the effects upon H₂S removal achieved by adjusting the SO₂input in a system in accordance with the invention so as to achieve a specified concentration of thiosulfate.

[0037] The SANS process as described above can be applied to many different sulfur recovery situations with the addition of SO₂to the feed or sour inlet gas, or with SO₂already being present in the inlet sour gas. Certain gas streams of interest will already contain SO₂along with H₂S, and in those cases there may be no need to add additional SO₂. For these cases it may be desirable to adjust the conditions of the process which generates the gas to obtain the optimum mole ratio of 2:1 (H₂S:SO₂). Where such a manipulation of concentrations is not possible various amounts of SO₂can be added to the gas stream by the methods previously described. The SANS process will then remove both SO₂and H₂S to form sulfur according to the overall reaction (5) given above. In practice of the present invention the sorbent liquor used will unless otherwise indicated be comprised of components as has been above described in the referenced 5,738,834 patent. It is preferred that the sorbent have a low dissolving power for water so that any condensation of liquid water does not contaminate the sorbent in such a way as to necessitate a separate distillation or water stripping step. Similarly, the sorbent components should have a low solubility in water, such that they are not lost to a separate water phase which may be formed by condensation

[0038] An example of a gas stream which already contains SO₂is the off-gas from the catalyzed gas phase Claus process. In the Claus process, a portion of the H₂S in the sour gas stream is converted to SO₂and the mixture of H₂S and SO₂is passed across a catalyst at elevated temperature to cause the conversion to elemental sulfur. This process is highly exothermic and is equilibrium limited at the elevated temperature, so the off-gas (“Claus tail gas”) contains significant amounts of SO₂and H₂S. The Claus process can be operated so as to produce the appropriate ratio 2:1 mole ratio of H₂S:SO₂in the tail gas and the SANS process will then remove the remaining SO₂and H₂S from this tail gas. Both the addition of SO₂and the process configuration where SO₂was already present in the input “sour gas” were demonstrated experimentally as depicted in the system 2 schematically shown in FIG. 1. The dashed line 4 represents addition of SO₂to the system at some point other than the inlet. The solid line 6 joins the H₂S-containing gas before being admitted to the absorber 8, showing that both were contained in the inlet gas for many of the experimental cases. FIG. 2 is a more general presentation of this case and may be considered simultaneously with FIG. 1.

[0039] The input gases proceed upwardly in absorber 8 wherein contact is effected with the downwardly moving sorbing liquor as described in the aforementioned 5,738,834 patent, with the purified sweetened gas exiting at 10. As otherwise described in the referenced patent, the sorbent liquor and its several components exit absorber 8 and proceed through a plug flow reactor 12 to an oxidizer 14, thence through sulfur crystalizer 16 which is agitated by a stirrer 15 and cooled by cooling water 18. Crystalized sulfur 20 is removed at settler 22, with the regenerated sorbent liquor 24 being passed through a pump 26 and flow measuring transducer 28 before being recycled to absorber 8.

[0040] It is to be noted in FIG. 1 that the reactions occurring in absorber 8 are such that the sorbent liquor from the absorber need not pass via the oxidizer 14 as in the 5,738,834 patent; but instead as shown by dashed line 25 the sorbent flow may proceed directly to the sulfur crystalizer.

[0041] In the more generalized showing of FIG. 2, a system 30 is shown in which similar components to those in FIG. 1 are identified with the corresponding reference numerals. It is assumed here that the ratio of H₂S to SO₂in the input gas is at or brought to the classical 2:1 ratio preferred for the Claus reactions, as shown at 32. Several additional details are shown in FIG. 2, such as the loop from crystalizer 16 which includes a crystalizer pump 11 and cooling means 13. Similarly the sulfur removal means is shown as including a sulfur filter and washing means 17 with the filtrate from the latter being recycled via a filtrate pump 19. Also a heater 23 is shown for bringing the recycled sorbent liquor back to the appropriate temperature for the absorber.

[0042] A feature has been added to the system 30 which is otherwise disclosed in the present applicant's co-pending patent application filed as International Application PCT/US99/16,500, and published on Feb. 3, 2000 as WO 00/05171. As applicable to the system 30, it is found that a small fraction of the removed sulfur-containing gases are converted to sulfur oxyanion byproducts (such as sulfate) rather than to elemental sulfur. Although the fraction is small, if these compounds are not removed they can build up in concentration and cause operating problems or necessitate costly blowdown of the solution. The present sorbent system 30 allows the use of the unique method of the referenced application for removing these unwanted byproducts. The method is based on addition of gaseous ammonia 21 to the process solution. Surprisingly, it has been found that ammonium sulfate in particular is quite insoluble in the nonaqueous SANS solution, in contrast to its high solubility in aqueous solutions. Therefore, bubbling ammonia into a SANS solution containing these salts results in nearly instantaneous formation of solid ammonium sulfate which precipitates from the solution, thereby allowing its removal at separator 36 by settling, filtration, or other common solid/liquid separation methods. The reaction (for sulfate) can be written as follows, with B representing the amine (HB⁻is then the protonated amine):

[0043] HB HSO₄+2 NH₃→(NH₄)₂SO₄↓+B

[0044] The reactions appear to be essentially quantitative, and since ammonia is quite inexpensive, the removal process is economically favorable. Water-soluble components can be used in the sorbent without losing them to the water wash or using other separation steps to recover them from the wash water of the earlier method. Ammonium thiosulfate is also removed by this method if its concentration is high; however in the present system it can undergo further reaction with H₂S to form elemental sulfur and so its concentration is unlikely to rise to the point where a substantial amount of it is removed by ammonia addition. In general however, sufficient quantities of ammonia are used to bring the concentrations of sulfate and thiosulfate species below a predetermined point, usually to less than 0.05M.

[0045] In the system 40 schematically depicted in FIG. 3, a full oxidation catalyst is provided at reactor 42 upstream of the absorber to convert enough H₂S to SO₂to give the desired 2:1 ratio. Here the oxidation catalyst reactor 42 is seen to receive a portion 44 of the input sour gas 132 and effect the desired full oxidation to SO₂. The output 46 from the oxidation reactor 42 is then returned to the input stream to absorber 8. The remaining portions of system 40 are as in system 30 of FIG. 2, and corresponding elements are similarly identified.

[0046] Gas phase catalytic processes such as Claus and others form elemental sulfur in the gas phase which is normally condensed to liquid form for removal. Complete removal by condensation is expensive, and the tail gases from such processes often contain elemental sulfur vapor which may end up being converted to SO₂when the gas stream is incinerated, producing a source of pollution. The SANS process is useful for removing any such gas phase elemental sulfur, due to the high solubility of elemental sulfur in the SANS sorbent. Thus the SANS process can be effectively used to remove that sulfur as well as any remaining H₂S or SO₂in the tail gas stream. For those cases where SO₂is generated catalytically for mixing with H₂S prior to contacting with the SANS sorbent, it may also be desirable to operate the catalyst such that part of the H₂S is converted to elemental sulfur by the catalyst, since this elemental sulfur can be readily removed by the SANS sorbent followed by crystallization of solid sulfur from sorbent on cooling.

[0047] Accordingly, instead of operating the oxidation catalyst reactor in FIG. 3 to effect full oxidation of the H₂S, a suitable catalyst and operating conditions can be employed so as to effect a partial oxidation reaction (not illustrated) which converts the H₂S to elemental sulfur which can be partially removed upstream of the absorber as molten sulfur by cooling the gas stream. Where operated as such a partial oxidation reactor, a catalyst can be used which is active and selective for the partial oxidation of H₂S to sulfur. Catalysts and conditions appropriate to this reaction are well known, and references may be had to the disclosures of U.S. Pat. Nos. 4,623,533; 4,857,297; 4,552,7465; and 4,311,683, setting forth details of suitable catalysts and operating conditions which may be used for these purposes—typically temperatures in the range of from about 150 to 400 C. are used. The sulfur-laden gas exiting the reactor is conveyed to a sulfur condenser, where most of the sulfur is condensed and exits the process as a molten liquid.

[0048] With the type of arrangement depicted in system 50 in FIG. 4 three types of reactions can be brought about upstream of the absorber: conversion of H₂S to SO₂; conversion of H₂S to elemental sulfur; and partial conversions of the foregoing character, so that quantities of H₂S can slip through unconverted. All of these variations can occur at the same time depending on the temperatures, the type of catalysts, and the amount of air added at the reactor—ranging from substoichiometric quantities of air to excess air. To enable these several results the full oxidation reactor of FIG. 3 is shown as replaced by a partial and full oxidation reactor 49.

[0049] In the arrangement of FIG. 4 it may be desirable to employ suitable catalysts to convert as much of the inlet H₂S to elemental sulfur as possible (which can then be removed as means 51 upstream of the absorber as molten sulfur when the stream is cooled as at 47), and then convert enough of the remaining inlet H₂S to SO₂so that the resultant gas has the desired 2:1 ratio. By maximizing the amount of sulfur produced and removed ahead of the absorber, less sulfur remains or removal at the end of the SANS process which proceeds downstream of the absorber. The crystallization of sulfur is generally most costly to carry out than bulk condensation of sulfur upstream of the absorber gas phase conversion. However, complete condensation of elemental sulfur becomes expensive, and the SANS solvent will be highly effective for removing sulfur vapor. In addition, a practical limit exists respecting the amount of sulfur which can be formed in the upstream gas phase reactor or reactors when the gas is being treated at pressure. This occurs because at high pressure the gas will hold less sulfur than at atmospheric pressure. Accordingly any sulfur formed which is in excess of the amount the gas can hold can condense in the catalytic bed to cause plugging.

[0050] In the following Examples, the effects of adding SO₂to the SANS H₂S removal process was tested using a continuous-flow bench scale apparatus. A diagram of the testing apparatus is shown in FIG. 1. Except as otherwise noted, pure H₂S gas is mixed with pure CO₂gas and the mixture is injected at the bottom of the absorber which in the arrangement shown was a 1.0 inch inner diameter column packed with approximately 36 inches of Pro-Pak® distillation packing (0.24 inch square grids of 316 stainless steel). This inlet gas pressure is slightly above atmospheric pressure. The scrubbing liquid is introduced at the top of this column and flows down through the packing. The liquid exits the bottom of the column, passes through a plug flow reactor, then flows into the bottom of a liquid-filled vessel which is sparged with air. The liquid exits this oxidizer and goes into a stirred crystallizer vessel where the temperature is reduced. The solution/sulfur slurry then goes to a settler where the solid sulfur falls to the bottom of the vessel and the liquid drawn off the top, reheated to operating temperature, and pumped up via a flow transducer to the top of the absorber column to repeat the cycle.

[0051] The sorbent for these tests consisted of nominally 0.6 M elemental sulfur, 0.4 M decyldimethylamine (ADMA-10, Albemarle Corporation) and 2.0 M 1-phenoxy-2-propanol (DOWANOL Pph, Dow Chemical) dissolved in phenyl-o-xylylethane (PXE, Koch Chemical). The liquid flow rate was varied from 25 to 55 mL/min and the liquid inventory was 1500-4000 mL depending on equipment configuration.

[0052] A long term (44 hour) run was done under typical SANS process operating conditions except with the addition of SO₂. The average inlet H₂S concentration was 32.9% and the average inlet SO₂concentration was 10%, producing a total S gas content of 42.9%. The total gas flow rate was 350 cc/min. The liquid flow rate was 55 cc/min, and the average liquid inventory was 4000 mL. These conditions were maintained for 44 hours, during which time the average H₂S removal was 99.7%. This average removal is significantly better than that observed in a number of other runs done with 42.9% inlet H₂S but no SO₂, for which the usual average removal is about 95%. During this run the sulfur oxyanion byproduct make rate was 1.92% (total moles sulfur in oxyanion byproducts per mole sulfur absorbed). The major sulfur oxyanion was sulfate. No other peaks which could have indicated other anionic byproducts were seen in the ion chromatograms used to analyze for sulfate and thiosulfate. A significantly lower than normal amine degradation rate of 0.28% (mole amine degraded per mole sulfur absorbed) was also observed in this run with SO₂than previously observed in runs done with no SO₂, typically 1.0% (mole amine degraded per mole sulfur absorbed). The solid sulfur produced by the process during the run was rinsed, dried and weighed, showing a yield of 522 g. This compares extremely well to the sum of the total sulfur introduced to the absorber as H₂S (404 g as S) and as SO₂(123 g as S), or 527 g S added per gas measurements.

[0053] For this run, the inlet concentration of H₂S to the absorber was reduced from its normal value of 42.9% to 21.5% and the liquid flow rate was also reduced by approximately a factor of two from 55 cc/min to 25 cc/min, with no SO₂added to the system. This produced conditions in which the removal of H₂S by the SANS sorbent is approaching equilibrium control and as a consequence the H₂S removal was only 64%. At this point, adding 2.0% SO₂to the gas stream and maintaining the same total sulfur concentration in the gas produced an H₂S removal of 72%. Similarly, adding 4.0% SO₂increased the H₂S removal to 80%, and adding 6.0% SO₂gave a removal of 90% of the H₂S. At this point, when the SO₂flow was turned off (without increasing the H₂S inlet to compensate), the H₂S removal decreased to a value of 59.4%, consistent with the low removal noted before starting SO₂flow, adjusted for the somewhat lower H₂S concentration.

[0054] To determine if the direct reaction of H₂S with SO₂occurs to an appreciable extent in this system, the bench scale unit was run under similar conditions to those in Example 2 except that the initial sorbent did not contain any elemental sulfur. At an average inlet H₂S concentration of 22%, SO₂concentration of 6%, and liquid flow rate of 25 mL/min, the outlet H₂S concentration was initially 10%, a removal of only 52%. The removal improved somewhat with time, probably due to the formation of some elemental sulfur either by the direct reaction of H₂S with SO₂or by air oxidation of sulfide in the oxidizer. Clearly, however, the direct reaction of H₂S with SO₂is not as facile as the reaction mediated by elemental sulfur, which evidently proceeds via the intermediate polysulfide formed by reaction of H₂S and S₈in solution.

[0055] The conditions were similar to Example 2 except that a higher gas flow rate was used (1075 cc/min) and thus the inlet H₂S concentration was decreased to about 7.0%. The initial solution contained 0.6 M (as S) elemental sulfur and 0.4 M ADMA-10 in addition to the DOWANOL PPh and PXE components and the nominal liquid flow rate was 25 cc/min. With no added SO₂, the H₂S removal was 32.3%. Keeping the volumetric flow rate of H₂S constant (at 75 cc/min), co-injection of 21 cc/min SO₂(a mole ratio of 3.6:1 H₂S:SO₂) resulted in an H₂S removal efficiency increase to 45%. Moving the point of injection of SO₂to just above the lower of the three 12 inch packed columns produced a further increase in H₂S removal to 50%. At this point the SO₂flow rate was increased to give a mole ratio of 2.0:1 H₂S to SO₂. The H₂S removal increased to 72.7%. Further increase of the SO₂flow rate to give a 1:1 mole ratio of H₂S to SO₂produced a further increase in H₂S removal to 99%. However, this last step also caused an increase in conductivity of the solution by more than a factor of ten, perhaps indicating that the SO₂was being introduced in excess of stoichiometric.

[0056] As an illustration of the ability of the reaction to proceed without the addition of oxygen the bench scale unit was modified by removing the oxidizer from the process loop (represented by the dashed line in FIG. 1). In the first run of this type, the plug flow reactor was also removed from the process loop. The sorbent solution was the same as that used in Example 4 except that the ADMA-10 concentration was increased to 0.8 M. The liquid inventory, 1500 mL, was lower than before due to the removal of the vessels. The unit was operated for 6.6 hour using an H₂S flow rate was 75 cc/min (21.3%), SO₂flow rate of 37.5 cc/min for 3.3 hours then 41 cc/min for 3.3 hour and CO₂added to give a total flow rate of 350 cc/min. At the initial H₂S:SO₂ratio of 2.0 the average removal was 99.6%. At the higher SO₂addition rate of 1.83:1 H₂S:SO2 the average removal was 99.96%. After the 6.6 hour run duration the CO₂flow rate was increased to give a total flow rate of 528 cc/min and inlet H₂S concentration of 14.2%. The SO₂concentration was set to give an H₂S:SO₂mole ratio of 1.9. The process was operated under these conditions for 1.7 hours while producing an H₂S removal efficiency of 99.9%.

[0057] Some other processes which utilize the reaction of H₂S and SO₂are not able to sustain good removal efficiencies if the H₂S/SO₂mole ratio varies from the optimum value (2.0) for even short periods of time. Laboratory experiments of the SO₂-enhanced SANS process have demonstrated that a temporary loss of SO₂feed to the system will not cause an immediate reduction in removal. In fact, as shown with the laboratory data in FIG. 5, it typically takes several hours to observe even a small reduction in H₂S removal when the SO₂feed is shut off during some portions of the testing.

[0058] This particular test was run with 3.5 L of a sorbent similar in composition to the one in Example 5, nominally 3.2% H₂S and 1.6% SO₂in the inlet gas (balance CO₂), using a sparged column followed by a Snyder column both held at 180 F. for the absorber. In this particular test, the sorbent had been subjected to an excess of SO₂for six hours prior to turning off the SO₂. Roughly 80% of the H₂S reacted during the time period shown in the figure can be attributed to this dose of SO₂, while the remaining 20% can be associated with the decrease in thiosulfate concentration from 0.049 M to 0.015 M during the time that the SO₂was turned off.

[0059] Thus the sorbent has a buffering capacity in that it evidently can maintain a large concentration of dissolved and complexed SO₂along with thiosulfate and other intermediates which can react with H₂S even in the absence of SO₂in the gas phase. This allows the system to continue achieving good removal even without the addition of SO₂for significant periods of time. Similar buffering effects were noted during the pilot plant testing described below.

[0060] The SO₂-enhanced SANS process described here was operated at a pilot unit located in the Permian Basin of west Texas. This pilot unit processes a slipstream of a 300 psig sour CO₂stream from enhanced oil recovery operations. Design flow through the pilot unit is 0.1 to 1.0 MMscfd with sulfur production of 20 to 200 pounds per day. The sour gas contains 1800 ppmv H₂S along with approximately 80% CO₂and 10 to 11% methane, with the remainder being heavier hydrocarbons (including a relatively high level of aromatic compounds). Since there is no SO₂in the inlet gas, SO₂was supplied to the process by pumping liquid SO₂from an SO₂cylinder. The process was operated essentially as shown in FIG. 2 except that the SO₂was injected into one of three places: directly into the absorber, into the lean solvent stream, or into the inlet gas for testing purposes. Results were essentially independent of the location of SO₂injection.

[0061] The pilot unit was able to consistently reduce the H₂S content of the inlet gas from 1800 ppm to approximately 25 ppm, far exceeding the removal specification for the host site (100 ppm H₂S in the sweet gas). The pilot unit was operated for 1550 hours with no chemical addition. The concentration of the tertiary amine and solubilizer decreased only slightly. The measured chemical costs during this period were on the order of $250/long ton of sulfur (LTS), corresponding to a loss of approximately 0.011 gmole amine/gmole H₂S removed. This is considerably lower than other aqueous based technologies on both the pilot and commercial level. The pilot and commercial chemical losses reported in the literature for one common aqueous iron liquid redox technology are $2000/LTS and $500 to $700/LTS, respectively. The amine loss includes both chemical degradation and volatilization.

[0062] Only a limited amount of the absorbed H₂S was converted to byproducts during the operation of the SO₂-enhanced process at the pilot plant. The byproduct build up in the pilot unit was less than 0.2% mole byproduct/mole H₂S reacted. In comparison, 2 to 4% is typically considered a manageable cost for most applications.

[0063] The ammonia addition approach for on-line removal of the byproducts was successfully demonstrated at the pilot plant. A slipstream of the SANS sorbent was sent to a separate tank and ammonia was added to precipitate the solid ammonium solid state. The slipstream was pumped through a filter to remove the precipitated salts and then returned to the main solution stream. Using this method, byproduct levels were controlled while operating the unit, and solution purging was not required. Results of operating the byproduct removal system are shown in FIG. 6. During testing with SO₂addition, it was only necessary to operate the salts removal system three times (4% of the total elapsed run time).

[0064] It was observed during pilot unit testing that H₂S removal was correlated with the thiosulfate concentration in solution and that the H₂S removal could be adjusted by setting the SO₂input so as to achieve a certain concentration of thiosulfate. An example of this phenomenon is shown in FIG. 7. As the thiosulfate concentration was increased, the H₂S concentration in the outlet gas decreased proportionally. Such a control technique is a unique and significant benefit for commercial applications.

[0065] While the present invention has been set forth in terms of specific embodiments thereof, it is to be understood in view of the present disclosure that numerous variations upon the invention are now enabled to those skilled in the art, which variations are yet within the teachings of the invention. Accordingly the invention is to be broadly construed, and limited only by the scope and spirit of the claims now appended hereto.