Flexibility of Liquid Redox Processing in Refinery Sulfur Management

Gary J. Nagl


With over 25 years of experience, sulfur recovery via iron-based, liquid redox processing is a commercially proven approach to refinery sulfur management. The inherent characteristics of liquid redox processing, such as 100% turndown in respect to H2S concentration, flowrate and sulfur loading and single stage removal efficiencies in excess of 99.9%, makes the process attractive as a stand alone device for applications of less than 15 LTPD of sulfur or in conjunction with a Claus unit at higher capacities resulting in an overall system having a removal efficiency of 99.9+% and 100% turndown capability.

Another attractive inherent feature of iron-based, liquid redox processing is its ability to process any type of gas stream such as fuel gas, amine acid gas, and sour water stripper gas. Consequently, for refineries with Claus plants, the liquid redox process can be employed as a tail gas treatment unit, with or without a hydrogenation/hydrolysis unit, while also directly processing the sour water stripper gas relieving the Claus unit of this burden. All of this can be accomplished without recycling gas back to the Claus unit, thus increasing the capacity of the Claus unit.


With the installation of more and more desulfurization processes to meet ultra low sulfur specifications for gasoline and diesel in conjunction with more and more pressure from environmental agencies to reduce sulfur emissions, refiners are faced with the challenge of what to do with the additional H2S being generated. In some cases, there is sufficient spare capacity in the existing amine and Claus units to process the additional load. And in other cases, the existing Claus unit(s) can be updated with oxygen enrichment technology to handle the additional load or in other cases, new sulfur recovery systems will be required to handle the incremental load. Another approach to this challenge, which is often overlooked, is the addition of liquid redox processing. Due to its excellent versatility, sulfur recovery via liquid redox processing can be employed to treat any type of sour gas streams. In a refinery this means that liquid redox technology can be use to treat sour, gaseous hydrocarbon-based streams directly or acid gas streams from amine units, or sour water stripper gas or even Claus tail gas streams. By integrating liquid redox processing into the overall refinery sulfur recovery program, refiners may be able to easily accommodate the additional sulfur recovery burden imposed by the production of ultra low sulfur fuels.

Sulfur Recovery via Liquid Redox Processing

Liquid redox processes employ aqueous-based solutions containing metal ions, usually iron, which are capable of transferring electrons in reduction-oxidation (redox) reactions. Currently, the redox process of choice is the LO-CAT® process, which is licensed by Gas Technology Products LLC. In this process, a non-toxic, chelated iron catalyst is employed to accelerate the reaction between H2S and oxygen to form elemental sulfur.

H2S + O2 S0 + H2O (1)

As implied by its generic name, liquid redox, all of the reactions in the LO-CAT® process occur in the liquid phase in spite of the fact that Equation (1) is a vapor phase reaction. In the process, the sour gas is contacted in an absorber with the aqueous, chelated iron solution where the H2S is absorbed and ionizes into sulfide and hydrogen ions as follows.


H2S + H2O 2H+ + S–– (2)

This reaction is mass transfer limited.

The dissolved sulfide ions then react with chelated, ferric ions to form elemental sulfur as follows.

S–– + 2Fe+++ S0 + 2Fe++ (3)

Again, this reaction is mass transfer limited.

Adding Equations 2, 3, and 4 yields Equation 1.

Treating Acid Gas Streams

As illustrated in Figure 1, Autocirculation type LO-CAT units are used when treating acid gas streams and streams, which can be mixed with air without creating a safety problem. In this type of unit, the absorber and oxidizer are contained in one vessel and separated by baffles. Due to the large differences in aerated densities between the liquids in the absorber and the oxidizer large circulation rates are achieved between the various compartments of the vessel without having to employ pumps. The acid gas enters the absorber section of the vessel (centerwell) where it is contacted with oxidized LO-CAT solution and where the H2S is converted to elemental sulfur. The partially reduced solution then circulates to the oxidizer section where it is contacted with air, which reoxidizes the iron. The exhaust air from the oxidizer and the sweet acid gas from the absorber are combined and exhausted to the atmosphere.

In the conical portion of the vessel, the sulfur will settle into a slurry of approximately 15 wt%. A small stream is withdrawn from the cone and sent to a vacuum belt filter where the sulfur is further concentrated to approximately 65 wt% sulfur. Some units stop at this stage and sell the sulfur cake as a fertilizer. Drier sulfur cake can be formed by employing pressure filters. If molten sulfur is required, the cake is reslurried and melted as shown in Figure 6.

Removal efficiencies of greater than 99.99 % and turndowns of 100% can easily be achieved with an Autocirculation unit.

Treating Sour Water Stripper Gas Streams

Contrary to Claus units in which sour water stripper gas (SWS) can create operational problems, processing SWS gas in a LO-CAT unit is actually beneficial. As previous described, all reactions in a liquid redox process occur in the liquid phase and are very rapid; however, the mass transfer of the H2S from the gas phase to the liquid phase is relatively slow, and is thus the rate limiting step in the process. One method of increasing the mass transfer rate is to increase the alkalinity or pH of the circulating liquid solution. This is usually accomplished by adding caustic in the form of NaOH or KOH to the system, which creates an operating expense; however, due to the ammonia, which is usually contained in the SWS gas, no caustic addition is required when processing SWS gas, thus the operating cost is reduced.

Gas from sour water strippers, which contain reflux overhead systems, can be processed in a LO-CAT unit as illustrated in Figure 2; however, for non-reflux sour water strippers, the SWS gas will require cooling to reduce the temperature to approximately 140–150°F (60–65°C), otherwise a water balance problem will be created in the LO-CAT system. The condensed water from this cooling step would be recycled back to the sour water stripper.

As shown in Figure 2, the SWS gas is contacted with LO-CAT solution in a venturi absorber, which serves two purposes. First, the venturi is an excellent, non-plugging absorber for this service and second, the venturi will produce a draft, which reduces the pressure requirement of the sour water stripper. The oxidizer and sulfur removal is the same as previously described.

The effluent gas from the Oxidizer will contain ammonia and water vapor and a good deal of dilution air. Depending on the amount of ammonia present and local regulations, this stream may need to be process through a catalytic reduction unit to remove the ammonia or combusted in a Claus incinerator if one is available. Again removal efficiencies of 99.99% and turndowns of 100% can easily be achieved with this processing scheme.

Removing the SWS gas from the Claus unit and processing it in a liquid redox system will not only free up capacity in the Claus unit but it will also eliminate the operational problems associated with processing this gas stream.

Treating Refinery Fuel Gas

When treating combustible gas streams such as fuel gas streams or recycle gas streams in which the gas stream cannot be mixed with air, a conventional processing scheme as illustrated in Figure 3 is employed. This processing scheme is similar to that of the sour water stripper gas processing scheme previously described with the exception that the venturi absorber is replaced with a non-plugging, liquid full absorber.

Fuel gas, generally at 60–70 psig (4–4.8 barg) is passed through the liquid full absorber. If so desired, the H2S concentration of the fuel gas can be decreased to less than 4 ppm (V/V) in this type of absorber. The spent solution from the absorber is directed to the oxidizer, which is simply a liquid full, atmospheric tank with a cone bottom through which air is sparged. The flue gas from the oxidizer is moist air that is slightly depleted of oxygen and contains no H2S. The flue gas is normally discharged to the atmosphere; however, if the fuel gas contains BTX components, it will need to be incinerated. Sulfur is separated and removed from the system in the same manner as with the previously described Autocirculation processing scheme.

High pressure gas streams up to several hundred pounds of pressure can be treated in this type of processing scheme; however, a flash drum will be required between the absorber and the oxidizer. For very high pressure gas streams, it is generally more economical to treat the gas in an appropriate amine unit and then treat the acid gas in an Autocirculation LO-CAT unit.

Treating Claus Tail Gas

There are two approaches to coupling a liquid redox process to a Claus unit(1). The first method involves processing the Claus tail gas through a cooler and then directly into a liquid redox unit. The second method involves processing the Claus tail gas through a hydrogenation/hydrolysis reactor, which will convert all of the SO2, CS2 and COS to H2S followed by a cooler and a liquid redox unit. Both methods will result in removal efficiencies exceeding 99.9%. The only difference will be the operating cost of the first approach will be greater than that of the second. Both approaches will also result in turndown capabilities of 100%. This is accomplished by by-passing the Claus unit when it has reached its turndown capability and directing the acid gas to the redox unit.

When considering liquid redox to treat Claus tail gas without the inclusion of a Hydrogenation/Hydrolysis reactor, the amount of SO2 in the tail gas is an important operating parameter. Since liquid redox units operate at alkaline pH’s in the range of 8 to 9, any SO2 in the tail gas will be easily absorbed, and form sulfates and sulfides in accordance with Reaction 1 and 2.

SO2 + 2NaOH + 1/2 O2 Na2SO4 +H2O (1)

It is important to note that SO2 does not interfere with the liquid redox chemistry and consequently does not affect the H2S removal efficiency of the process. However, Reactions 1 and 2 do affect the operating cost of the process in two ways. First, caustic is consumed for each mole of SO2 absorbed, which increases the operating cost of the unit. Secondly, the resultant sulfate/sulfide product will accumulate in the liquid redox solution, and eventually a continuous blowdown will be required resulting in loss of valuable catalyst solution, which must be replaced, again, increasing operating cost even further. Consequently, if this process configuration is to be employed, it is advantageous to minimize the formation of SO2 in the Claus unit by operating the Claus unit with sub-stoichiometric quantities of oxygen, thus increasing the H2S:SO2 ratio in the unit.

A flow diagram of a typical LO-CAT liquid redox unit for treating Claus tail gas directly is shown in Figure 4. Since the liquid redox system is aqueous-based, elevated temperatures will cause water balance problems; consequently, the tail gas is first passed through a quench tower where the gas temperature is reduced from approximately 135°C to 50°C.

For direct treatment of Claus tail gas, the LO-CAT process would employ a venturi absorber followed by a proprietary Mobile Bed Absorber (MBA). The venturi not only supplies much needed draft to the system, but it also provides a fair amount of H2S removal. The MBA employs hollow, ping-pong-like spheres as contacting media which, when fluidized, are self-cleaning

In the event that the Claus unit is unable to operate due to turndown requirements beyond its capabilities, the system can be designed to bypass the Claus unit entirely and route the acid gas directly into the LO-CAT unit. In this mode of operation, the MBA will be bypassed, and the LO-CAT unit will operate as an Autocirculation unit as illustrated in Figure 1. To accommodate this operating scheme, the operating capacity of the Autocirculation vessel will need to be equivalent to that of the turndown capacity of the Claus unit.

For direct treatment of Claus tail gas, the LO-CAT process would employ a venturi absorber followed by a proprietary Mobile Bed Absorber (MBA). The venturi not only supplies much needed draft to the system, but it also provides a fair amount of H<sub<>2S removal. The MBA employs hollow, ping-pong-like spheres as contacting media which, when fluidized, are self-cleaning

In the event that the Claus unit is unable to operate due to turndown requirements beyond its capabilities, the system can be designed to bypass the Claus unit entirely and route the acid gas directly into the LO-CAT unit. In this mode of operation, the MBA will be bypassed, and the LO-CAT unit will operate as an Autocirculation unit as illustrated in Figure 1. To accommodate this operating scheme, the operating capacity of the Autocirculation vessel will need to be equivalent to that of the turndown capacity of the Claus unit.

This mode of operation (Figure 4) will still yield H2S removal efficiencies of 99.99+%. In addition, the effluent gases from the liquid redox unit will not require incineration since the tail gas will contain only a very small amount of H2S and no SO2, and it will be diluted with the effluent air from the Oxidizer. It is also important to note that there is no recycle stream back to the Claus unit; hence the capacity of the Claus unit will not be reduced by the addition of a liquid redox system to treat the tail gas. In addition, by coupling the treatment of the tail gas and the SWS gas into a liquid redox system, addition Claus capacity is created.

In indirect processing scheme all sulfur compounds in a Claus tail gas are converted to H2S by passing the tail gas through a hydrogenation/hydrolysis, catalytic reactor at elevated temperatures. Reactions 3 and 4 (hydrogenation) and Reactions 5 and 6 (hydrolysis) represent the major reactions, which occur in the reactor.

SO2 + 3H2 H2S + 2H2O (3)
S2 + 2H2 2H2S (4)
CS2 + 2H2O CO2 + 2H2S (5)
COS + H2O CO2 + H2S (6)

In this processing scheme (Figure 5), a fuel gas is subjected to partial oxidation, which not only generates sufficient heat to raise the tail gas to reaction temperatures but also generates sufficient hydrogen to satisfy the requirement of Reactions 3 and 4.

After passing through the reactor, the effluent gas must be cooled to approximately 50°C, which will generate sour condensate. Again, a portion of the sour condensate may be used as makeup water for the liquid redox unit; however, some of it will need to be sent to a sour water stripper with the vapor being routed back to the liquid redox unit. The operating scheme of the LO-CAT unit will be identical to those described for the indirect treat treating case.

In summary, employing a liquid redox process as a tail gas cleanup yields the following benefits:

  • 99.99+ % overall efficiency without having to recycle gas back to the Claus unit.
  • Insensitivity to SO2 break through.
  • By incorporation of sufficient design capacity, 100% turndown for the sulfur recovery unit can be achieved.

Sulfur Disposal

Because of its poor quality, sulfur produced from liquid redox processes has a bad reputation, which in some cases is well earned. However, due to the relatively small quantities of sulfur produced in liquid redox installations, most liquid redox sulfur has been either landfilled or disposed of as solid, agricultural sulfur; hence, not a lot of effort has been exerted to improve its quality. However, great progress has been made in improving the quality of sulfur produced in LO-CAT units(2).

Sulfur is produced as a solid in a liquid redox unit. Since the reactions are not gas phase, there is no dissolved H2S in liquid redox sulfur, thus sulfur degassing is never required. The sulfur is normally filtered and washed to produce a filter cake, which is 65% to 85% sulfur depending on the type of filter used, with the remainder being water and dissolved salts, albeit the salts are also good fertilizers and soil nutrients. It is not possible to simply dump this cake into the Claus sulfur pit, since there is insufficient heat in the pit to evaporate the water and to melt the sulfur. And even if the moisture is removed prior to dumping the sulfur in the pit, the solid particles have a tendency to float on top of the molten sulfur thus making heat transfer and consequently melting very difficult. Due to these problems, the sulfur from the liquid redox system must be disposed of as a solid or melted prior to being introduced into the sulfur pit. A typical melter system for a LO-CAT system is shown in Figure 6. It is worth noting that solid LO-CAT sulfur is in demand for use as a fertilizer and consequently, is demanding prices of around $35/LT.

Chemical Operating Costs

Chemical operating costs for a LO-CAT unit are generally between $100 and $200/LT of sulfur processed in the LO-CAT unit. If a Claus unit is operating at 96% conversion with the remaining 4% being processed in the LO-CAT unit, the chemical operating cost of the LO-CAT unit will add approximately $6/LT of total sulfur processed in the overall sulfur recovery unit.


Liquid redox systems are extremely versatile and capable of achieving very high removal efficiency (99.9+ %) while yielding 100% turndown. These capabilities result in a proven technology that is capable of meeting today’s sulfur management requirements in refineries. By incorporating liquid redox processing in refineries, the capacities of existing Claus units can be increased by the removing sour water stripper gas and the tail gas treating unit recycle gas from the Claus unit, thus creating surplus Claus capacity.


  • Nagl, G. “Employing Liquid Redox as a Tail Gas Cleanup Unit”, Sulphur 2001 Conference, Marrakech Morocco, October 2001.
  • Nagl, G. “Emerging Markets for Liquid Redox Sulfur”, Sulphur 1997 Conference, Vienna Austria, November 1997.