LARGE PLANT SULFUR RECOVERY PROCESSES STRESS EFFICIENCY

B. Gene Goar, Elmo Nasato
Goal, Allison & Associates Inc,
Tyler, Tex.

Natural gas processing in the future will encounter significantly more raw sour gas, i.e., gas containing 15 20 mol % H2S or greater.

Deciding whether to make the significant investment to build a sour gas treating and sulfur recovery plant involves many considerations.

An operator selecting the optimum gas treating, sulfur recovery, and tail gas cleanup processes must choose each carefully because each upstream step can affect design and operation of subsequent downstream steps.

Reviewed here are current as treating and sulfur recovery processes in use today. For this purpose, plants that process sour gas and recover 50 long tons/day (ltd) or more of sulfur will be considered. Sample plants employing major technologies are listed in Table 1.

Such other sulfur recovery, removal processes as Lo Cat, SulFerox Stretford, Iron Sponge, SulfaTreat, and Sulfa Scrub are available to industry. But these processes are normally considered only when the total sulfur to be handled is 15 20 ltd or less. Additionally, many are troublesome to operate and have relatively high operating costs.

GAS TREATING

Selecting the best gas treating process for a given application must include consideration of the following:

The several viable processes available
The sales gas specifications to be met
Possible LPG liquids recovery from the gas
The need to keep energy consumption to a minimum
The effect on the downstream sulfur recovery unit (SRU)
The desired simplicity of operations
The need to hold capital investment and operating costs to a minimum as much as practical to yield the best economics for the project.

Major factors that affect the selection of an overall gas-processing scheme are the following:

Plant design capacity and expected throughput
Concentration of sour components in the raw gas
Gas contactor treating pressure
Ratio of H2S/CO2
Permissible CO2 content in the sales gas
Presence of trace sulfur compounds (COS, CS2, mercaptans, etc.)
Tonnage of sulfur to be recovered
Overall sulfur recovery efficiency (%) to be achieved to meet local environmental regulations.

The gas treating processes normally considered for major sour-gas treating facilities may be divided into three general categories:

Chemical solvent processes using alkanolamines (monoethanolamine-MEA, diethanolamine DEA, diglycolamine DGA, and methyldiethanolamine MDEA).
Chemical solvent processes using alkaline salt solutions (Hot Pot, Catacarb, and Benfield).
Physical solvent processes (Sulfinol, Selexsol, and Sepasolv).

Following are some general guidelines for selecting the best gas-treating process.

MEA was the workhorse of the industry until about 20 years ago. It has lost appeal in recent years because of its relatively large circulation rates, high energy consumption, high solvent vaporization losses, and susceptibility to degradation by trace sulfur compounds often found in sour gas.

DEA initially gained popularity because of the development of "high-load" DEA technology, which yields much lower circulation rates, lower energy consumption, lower vaporization losses, and resistance to degradation from trace compounds.

DEA is probably the most widely used gas treating solvent, today.

DGA has a special area of application for gas treating.

It is well suited for gas processing in very hot climates (Middle East and similar desert areas) and when the gas contains appreciable trace sulfur compounds.

In many applications, it is very energy efficient.

Disadvantages of DGA are the high cost of the solvent, the relatively high viscosity of the solution (effect on heat exchange), the need to reclaim the solvent, and its tendency to absorb heavy or aromatic components from the sour gas stream.

MDEA has gained popularity in recent years because it exhibits significant selectivity for H2S in the presence Of CO2. It is also very energy efficient.

If the H2S/CO2 ratio in the sour gas is relatively low (say, less than 1:3), MDEA can be used to good advantage to absorb essentially all of the H2S and only a small portion of the CO2. This can yield a better acid gas feed to the Claus SRU.

If too much CO2, is "slipped" in the treating step, however, the sales gas' CO2 specification can be exceeded. If so, a second treating step can be used to remove the CO2; but, this is usually very expensive.

Use of MDEA should be evaluated carefully for each application.

The alkaline salt solution processes (Hot Pot, Catacarb, and Benfield) appear to have lost popularity for major gas processing projects in recent years.

These processes are probably best suited for bulk removal of sour components, where applicable, with final cleanup or polishing by a final alkanolamine step. Some alkanolamine and physical solvent processes offer several advantages over alkaline salt processes.

The Shell Sulfinol process appears to be gaining popularity in recent years.

If the sour gas to be treated is at a high pressure (for example, 800 1,000 psig), if the H2S/CO2 ratio is 1:1 or greater, and if the acid gas partial pressure is 110 psia or greater, Sulfinol may be an excellent choice.

It is very efficient for removing trace sulfur compounds to low levels, and it is very energy efficient.

Some disadvantages of Sulfinol are the high cost of the solvent and the absorption of heavy and aromatic components from the sour gas (problems for Claus SRU).

The Selexol and Sepasolv processes resemble Sulfinol in many ways: absorption of heavy, aromatic components can present a serious problem, and the solvents are expensive.

Their process flow schemes are more complex than the scheme for Sulfinol. Selexol appears to have lost popularity in recent years.

SULFUR RECOVERY OPTIONS

Process options for recovery of sulfur from sour gas all include some variation of the venerable Claus process.

THREE STAGE CLAUS

The original Claus process was patented and put in operation in 1883. I.G. Fabenindustrie patented and installed the "modified Claus process" in 1938.

Various schemes of this modified process have been used since. A[I versions of the basic Claus process are nonproprietary because it has been in use for so many years.

When the acid gas feed to a Claus sulfur recovery unit SRU contains 35-45 mol % H2S or greater, the once-through Claus process is normally used.

Three catalytic stages are normally used after the initial thermal stage. if an overall sulfur recovery of 95 97% is acceptable, then a tail gas cleanup unit (TGCU) probably will not be required.

In most U.S. states, an SRU of 20 ltd or larger will require some form of tail-gas cleanup. In Canada, SRUs of 50 ltd or larger normally require a TGCU process. The requirement for TGCU units for smaller SRUs results from more stringent environmental regulations that have emerged in the last 810 years.

In the once through Claus process, all acid gas is fed to the main burner on the reaction furnace (Fig. 1). Enough air is fed to the reaction furnace to burn a third of the H2S to SO2 and all hydrocarbons to CO2. Thermal conversion then occurs in the furnace, and the process gases are cooled in waste heat boiler.

The sulfur produced in the thermal step (typically 65 70%) is removed in the first sulfur condenser. From that point forward in the process, there is a repetition of three steps: reheating, catalytic conversion, and sulfur condensation.

The sulfur components in the final tail gas are normally oxidized to SO2 in a thermal incinerator and vented to atmosphere through a tall stack.

Various types of reheat schemes may be used, depending upon the availability of high pressure steam, richness of the acid gas feed, and overall sulfur recovery efficiency (%) required.

The Claus process has proven very reliable and is considered a mature technology.

Certain problems can be experienced, however, especially with older or poorly designed units.

One major problem can occur with the operation of the main burner (on reaction furnace). The main burner is probably the single most important piece of equipment in the Claus process; some designers nevertheless select the least expensive, most primitive burners on the market.

Use of modern, high intensity, efficient mixing main burners can result in a more stable flame, especially with leaner feeds; much better contaminant destruction for compounds such as hydrocarbons, NH3, mercaptans, etc.; reduced or nil oxygen breakthrough; improved Claus thermal conversion; and much wider turndown or turnup operations.

In addition to better burners, improvements are being made in better Claus catalysts and improved process controls.

The use of electronic instruments and "advanced control" can be very beneficial to the operation of Claus SRUs and TGCUs.

SPLIT FLOW PROCESS

When the acid gas feed contains 30-35 mol % H2S or less, the split-flow Claus process can be used (Fig. 2).

In this scheme, 40 45% of the acid-gas feed enters the burner on the furnace, and 55 60% of the feed bypasses the furnace. The split flow of the feed stream is used to achieve a hotter temperature and more stable flame in the furnace.

The bypassed portion of the feed can flow to the side of the furnace (poor design) or completely around the furnace and waste-heat boiler (proper design) and join the effluent from the No. 1 condenser. The mixed gases then flow to the first catalyst bed.

From this point, the flow scheme is essentially, the same as the once through process.

The sulfur recovery efficiency achieved for this process is normally 1 3% lower than for the once through process. Also, catalyst life in the first converter is typically shorter than in the once through process because of hydrocarbon contamination from the split flow stream.

From an operating standpoint, it is preferable to avoid the split flow process. Alternate techniques which can be considered are air and acid gas stream preheating, fuel gas spiking, and oxygen enrichment.

The use of modern, more efficient main burners will enhance the operation of all of these techniques, as well as the split flow scheme.

OXYGEN ENRICHMENT

If an operating company wants to construct an SRU of a certain size (100 ltd, for example) for air based operation but also to have the capability to operate the SRU at a much higher capacity (150 200 ltd, for example) for peak shaving, redundancy, or higher raw-gas production purposes, oxygen enrichment can be used to good advantage.

By substitution of pure oxygen for some or all of the air to the main burner, a significant increase in throughput can be achieved. Three levels of oxygen enrichment can be considered:

Low level enrichment: up to 28 mol % oxygen in the air stream
Medium enrichment: up to 35 40 mol % oxygen with a special burner designed for handling significant amounts of oxygen
High level enrichment: up to 6080 mol % oxygen with the COPE process (Fig. 3).

Naturally, the throughput of the SRU increases as the percentage of oxygen enrichment increases.

Oxygen enrichment offers several other benefits, among which are improved flame stability, better contaminants destruction, lower CS2 production, improved Claus conversion, and savings in fuel gas consumption in the thermal incinerator.

PLUS SUPERCLAUS

The Superclaus process, licensed by Comprimo b.v., Amsterdam, is a simple enhancement of the conventional Claus process that results in sulfur recoveries of 98.0 to 99.0%. The process can be easily retrofitted to an existing Claus plant, as well as incorporated in a new SRU design.

The Superclaus process was introduced commercially in 1988; currently,, 28 plants are under license, and 13 Superclaus plants are operating.

The process has gained popularity because of its proven reliability in operation, low incremental capital cost, low incremental operating cost, and ease of operation.

The Superclaus 99 process (Fig. 4) typically uses two or three Claus catalytic stages followed by a final Superclaus stage for improving overall sulfur recovery. Superclaus achieves higher recovery than a Claus stage by producing sulfur by direct oxidation of H2S to elemental sulfur.

The key to the process is the specially developed Superclaus catalyst which promotes only the direct oxidation of H2S to sulfur. This direct oxidation reaction is not limited thermodynamically, as in the case for the Claus reaction.

Superclaus is unaffected by the presence of other process gas components such as H2O, CO2, N2, H2, and other substances.

The control scheme used for Superclaus resembles that used for a Claus unit. The air to the main burner (on the reaction furnace), however, is reduced somewhat to produce an H2S/SO2 ratio in the Claus tail gas much greater than 2:1.

This air deficient operation tends to suppress the SO2 content of the process gas. The H2S content in the feed to the Superclaus stage is maintained at 0.5 1.5 mol % H2S, with a corresponding SO2 content typically of 0.05 mol % or less.

As indicated in Fig. 4, oxidation air is introduced upstream of the Superclaus reactor for the oxidation reaction.

This continuous, straight through process is gaining popularity worldwide. Examples of the Claus plus Superclaus process are shown in Table 1.

PLUS SCOT

In many locations, the sulfur recovery efficiency of a three-stage Claus SRU is far too low, and a tail gas cleanup process such as Shell Oil Co.'s SCOT must be used.

In some cases, a two stage SRU plus SCOT is used. In other cases, a three-stage SRU plus SCOT is selected. In either case, an overall sulfur recovery of 99.8 99.9% or greater is normally achieved.

The SCOT process (Fig. 5) occurs in essentially three stages:

Heating and reduction
Cooling and quenching
H2S absorption, stripping, and recycle.

The Claus SRU tail gas enters a reducing gas generator where the gas is heated to the proper temperature, and reducing gas (H2 and CO) is formed by substoichiometric combustion of fuel gas.

The hot stream, containing the reducing gas, flows to a cobalt-molyb-date catalytic hydrogenation reactor, where essentially all sulfur compounds (SO2, Sx, COS, CS2) are hydrogenated or hydrolyzed to H2S.

The effluent stream flows to a waste heat exchanger and a quench tower where the gas is cooled and the water content of the stream reduced from 25 35 mol % H2O to 4 5 mol % H2O. The cooled stream, containing mostly N2, H2S, and CO2 (when present), flows to the SCOT absorber.

In this tower, a selective solvent (usually MDEA) absorbs the H2S and rejects the majority of the CO2 (when present). The effluent vent stream, containing typically 100 200 ppm (v) of H2S, flows to a thermal incinerator.

The rich solvent is regenerated in the conventional manner; and the H2S rich regenerator overhead stream is recycled back to the Claus unit.

As long as the upstream Claus SRU is operated properly, the SCOT unit is reliable and performs well. If the SRU becomes upset and the H2S/SO2 ratio in the tail gas becomes grossly "off ratio," the SCOT unit will experience operating problems.

Excessive SO2 content of the Claus tail gas can cause major problems in the SCOT unit. In general, the SCOT process is considered to be reliable, flexible, and widely accepted worldwide.

The process equipment used in the SCOT process is familiar to many gas plant operating personnel. This is certainly not the case with some older, competing TGCU technologies.

Examples of Claus SRUs plus SCOT units appear in Table 1.

COLDBED SUB DEW POINT

The three cold bed sub dew point Claus processes currently in commercial operation are Sulfreen, CBA (Cold Bed Absorption), and MCRC. These are all variations of the same basic concept, differing in the regeneration technique.

Sulfur adsorption and desorption from catalytic reactors results in these processes being cyclical and semicontinuous.

Introduced in 1970, Sulfreen was the first sub dew point process; it is licensed by Societe Nationale Elf Aquitaine, France.

The process (Fig. 6) converts H2S and SO2 contained in the Claus tail gas to sulfur at sub dew point temperatures in the range of 260 290 F. Activated alumina is used as the catalyst and adsorbent in the reactors.

The cycle of switching gas flow from one reactor to another is set by the sulfur holding capacity of the catalyst in each bed.

In the regeneration step, inert heating gas is circulated by a blower. The hot gas vaporizes the sulfur and is cooled before going back into the process.

A portion of the cool inert gas is taken from the blower discharge and flowed through the bed on the cooling cycle. Sulfur recovery of up to 99.0% can be achieved for rich acid gas feeds.

The major problems encountered with the Sulfreen process have been mechanical and maintenance ones associated with the gas-switching valves and the regeneration gas blower and heater.

These problems can be lessened by selection of good equipment and proper Operational and maintenance procedures. The cyclical nature of the process makes it a high maintenance scheme. ]t is not very energy efficient, as well.

AMOCO Production Co., Chicago, commercially introduced the "Cold Bed Adsorption" (CBA; Fig. 7) process in 1976 and licenses it. CBA is similar to the Sulfreen process, except that CBA uses an internal stream within the process to regenerate the sulfur-loaded alumina catalyst beds in the reactors.

Sulfur recovery of up to 99.0% can be achieved for rich acid-gas feeds.

As with Sulfreen, major problems encountered with the CBA process have been mechanical and maintenance ones associated with the gas-switching valves and the regeneration gas blower.

The problems can be reduced by selection of good quality equipment and proper operational and maintenance procedures.

The Mineral & Chemical Resource Co. (MCRC) process (Fig. 8) was introduced commercially in 1980; it is licensed by Delta Projects Ltd., Alberta.

Regeneration is achieved by cycling the three MCRC converters in turn to receive high temperature process gas from the gas gas exchanger located downstream of the first Claus converter.

This means that the outlet from the first Claus converter must be maintained it about 650 F. at all times.

This hot gas desorbs the sulfur which was deposited on the active bed during its sub dew point operation.

The MCRC process uses conventional alumina catalyst and achieves a sulfur recovery of up to 99.0% for rich acid gas feeds.

Problems encountered with the MCRC process have been mechanical and maintenance ones associated with the gas switching valves and severe corrosion in "dead legs" of piping.

The valve problems can be overcome with selection of good equipment and proper operational and maintenance procedures. The corrosion problems can be reduced by, proper heat tracing and steam jacketing of process piping.

Also, problems have occurred with H2S spikes during the regeneration procedure, which have resulted in occasional environmental violations.

Examples of the sub dew point Claus processes are shown in Table 1.

RECYCLE SELECTOX

If the acid gas feed stream contains less than 40 mol % H2S, more typically 8 25 mol % H2S or less, the Recycle Selectox process (Fig. 9), licensed by Unocal Corp., Brea, Calif., should be considered.

Typically, this process has been applied to relatively small plants (30 ltd or smaller), but it can be applied to larger ones. Recycle Selectox competes with the split flow Claus process.

It is an all catalytic process which eliminates the need to burn one third of the H2S in a furnace. No fires burn in this process.

The incoming acid gas feed is preheated, mixed with recycle gas, and then mixed with process air in the exact same stoichiometric amount as used in the conventional Claus process.

The hot, mixed gases enter the Selectox catalyst bed where a selective oxidation of H2S to SO2 occurs in the upper few inches of the bed. In the remainder of the bed occurs the Claus reaction (to about 80% completion).

The hot gases leave the Selectox converter and enter a sulfur condenser where they are cooled, and about 70 75% of the sulfur present in the feed is condensed and recovered.

Effluent gases from the condenser are split, and a portion of the stream is returned to the Selectox bed inlet as recycle gas. This recycle of mainly inert gas senses to dilute the feed and keep the bed's outlet temperature from getting too hot for carbon steel materials of construction.

The balance of the stream (net forward flow) is reheated and passed to one or more Claus stages with the final tail gas being incinerated and vented to the atmosphere.

If the acid gas feed contains more than approximately 1,000 ppm (v) aromatic hydrocarbons, the catalyst life will be shortened.

Overall sulfur recoveries of 94 97% are common. If a higher recovery is required, a Superclaus stage can be added after the final Claus stage.

The process is ideal for offshore oil and gas platforms because of its simplicity, compact equipment arrangement, and the elimination of any fired equipment. The Recycle Selectox process is well established in industry and thoroughly proven in several commercial installations.

Examples of Recycle Selectox SRUs are shown in Table 1.

THE FUTURE

Several new sulfur recovery technologies are being studied or developed. Following are some of the major ones.

THERMAL, MICROWAVE CRACKING

Thermal cracking of H2S at temperatures of 2,500 3,000 F. are being studied by, among several groups, the Alberta Sulphur Research Laboratory (ASRL), Calgary.

After many years of laboratory study of the process, ASRL has built a semi works unit and installed it at Petro Canada's Wildcat Hills plant (near Cochrane, Alta.). The unit is scheduled to be in full operation later this year.

ASRL plans to use a special ceramic membrane to separate the produced hydrogen from the elemental sulfur. The initial thermal-cracker design is fabricated from ceramic material and will be used to demonstrate the process.

The laboratory is also working on a further development stage of the cracker in which the ceramic material will also serve as a semipermeable membrane to allow removal of hydrogen as it is formed in the cracker.

Another technology, cracking of H2S by microwaves, is still in an early stage of development. It has been studied for several years at Argonne National Laboratories in the U.S. and at the Kurchatov Institute in Russia.

A fairly large unit (plasma type, using microwave energy) has been in operation at Orenburg, Russia, for several years. Early results were promising, but technical difficulties have presented several challenges.

HIGH PRESSURE PROCESSES

Two approaches have been proposed for a high pressure Claus process.

Sometimes called the Richards' process and defined in U.S. Patent No. 4,280,990, the RSRP process involves an all catalytic Claus process operating at relatively high pressures (100 300 psia or higher). No field pilot unit has been reported.

Air Products' high pressure, oxygen enriched Claus process is defined in U.S. Patent No. 4,684,514.

It involves operation of the Claus process in a more conventional manner (thermal stage, followed by a catalytic stage) but using pure oxygen rather than air and operating the entire process at a high pressure, typically 50 160 psia or higher.

No field pilot unit has been reported.

Although the process seems to have merit, many technical problems must be conquered before it will become commercial.

The advantage of a high pressure Claus process, especially when using oxygen rather than air, would be construction of an SRU with much smaller equipment and lower investment cost than a conventional SRU which typically operates at 8 10 psig.

The challenge will be to develop materials of construction and fabrication techniques that the industry will accept for such vessels operating with H2S, SO2, and sulfur at relatively high pressures.

FUTURE OF OXYGEN ENRICHMENT

The current trend among environmental regulatory agencies is to require an operating company to install multiple SRU trains and TGCUs for redundancy. Oxygen enrichment can be used to good advantage to solve the redundancy problem.

An operating company can build two air based SRU trains, each sized for 50 60% of the total sulfur load.

The smaller SRU trains and TGCUs would be much lower in capital investment. During normal operations, both trains would be operated using air only.

If the trains are equipped with a "high level O2 enrichment" process capability, either train can be operated to handle 100% or more of the total sulfur load (using high levels of O2), while the other train is down for repairs or inspection.

The pure oxygen can be supplied by one of several sources: connection to a gaseous 02 pipeline, if one is in the area; use of liquid oxygen for short-term operations such as during turnaround of a train; use of a pressure swing adsorption (PSA; or vacuum swing adsorption VSA) oxygen plant located near the SRUs; or, use of on site cryogenic oxygen.

If the need for pure O2 is frequent or continuous and exceeds 100 150 tons/day of O2, an on site cryogenic air-separation plant should be considered.

Use of oxygen enrichment in Claus plants will continue to increase. The economics of building much smaller SRU/TGCU trains certainly favors oxygen enrichment and should be evaluated for any new SRU project.

Availability and cost of pure oxygen will be a key consideration.