The State of Liquid Redox
As early as the beginning of the century, work has been ongoing to develop a liquid phase, regenerative process for converting hydrogen sulfide (H2S) into pure, elemental sulfur. This work has lead to the introduction of over 25 different processes, most of which with very little commercial success. However, in the late 1940’s, the North Western Gas Board and the Clayton Aniline Company developed the Stretford Process, which utilized an aqueous solution of vanadium and anthraquinone (ADA). Although the Stretford process had some serious process, operational and environmental problems, the process filled a much-needed niche and became fairly popular throughout the 50’s, 60’s and 70’s.
In the late 1960’s the CIP process, which employed an aqueous solution of chelated iron, was introduced in the United Kingdom. The process failed miserably; however, its failure did lead to the successful introduction of the LO-CAT® Process in the late 1970’s, which solved many of the problems encountered with the CIP Process and the Stretford Process. In the late 1980’s, another chelated iron process, the Sulferox Process, was introduced; however, the developers of the LO-CAT and Sulferox processes have recently combined efforts to improve liquid redox processing even further.
Throughout the last half of the century, liquid phase oxidation has played an important role in the recovery of sulfur from various sources of hydrogen sulfide. Of all the processes available for converting H2S to sulfur, current liquid phase oxidation systems are the most versatile. They are able to treat any type of gas stream containing H2S, at a wide variety of operating conditions and all at removal efficiencies exceeding 99%. This paper will discuss the major liquid oxidation processes, describing in detail their advantages and disadvantages. In addition, the current R&D efforts in the field of liquid phase oxidation will be discussed with a glance of what future developments may occur.
Keywords: Desulfurization, sulfur recovery, liquid redox, chelated iron, LO-CAT
Liquid phase oxidation processes for the removal of hydrogen sulfide (H2S) from gas streams were initially developed to correct certain problems associated with dry oxidation processes such as iron sponge. The problems being mainly large plot requirements, replacement of the oxidation media on a frequent basis, and safety problems. This development work led to the utilization of various oxygen carriers dissolved or suspended in a liquid phase, which could be regenerated continuously at ambient temperatures.
Most early development work was done for systems processing coal gas or town gas with the objective of removing both hydrogen sulfide and ammonia by the formation of ammonium sulfate and elemental sulfur. Some of these early processes included the Feld Process, the Gluud Process, and the Koppers C.A.S. Process all of which employed polythionate solutions. Because of the complex chemistry, none of these processes achieved any commercial success.
After the failure of the polythionate processes, development shifted towards utilizing suspensions of iron oxide in aqueous solutions, which in essence was an attempt at a continuous iron sponge process. The hydrogen sulfide reacts with an alkaline compound to form hydrosulfide, which reacts with iron oxide to form iron sulfide, which in turn reacts with oxygen to form iron oxide and sulfur. Development work in this area lead to the introduction of the Burkheiser, Ferrox and Manchester processes.
During the 1920’s the Thylox and Giammarco-Vetrocoke processes met with some commercial success. However, both of these processes employed thioarsenate solutions, which resulted in toxicity problems caused by the arsenic.
Another group of processes which showed technical promise but were limited by toxicity problems were those employing iron cyanide solutions. The Fischer Process and the Autopurification process are examples. These processes employed ferric and ferrocyanide complexes as oxidizing agents.
A summary of the various liquid phase oxidation processes, which have been developed throughout the years, is contained in Table I.
The Stretford Process
The first liquid phase, oxidation process, which gained widespread commercial acceptance, was the Stretford process. The process was developed by the North Western Gas Board and the Clayton Aniline Company in England to remove H2S from town gas. The original process utilized an aqueous solution of carbonate/bicarbonate and anthraquinone disulfonic acid (ADA). However, the process suffered from several inherent problems. The solution had a very low capacity for dissolved sulfides resulting in large liquid circulation rates and hence, high power consumption. In addition, the sulfur formation reaction was very slow requiring large liquid inventories and resulting in high byproduct formation (thiosulfates).
These problems were corrected to a certain extent by the addition of alkali vanadates to the solution, which, in essence, replaced dissolved oxygen as the oxidant in the conversion of hydrosulfide ions (HS-) to elemental sulfur. In the reaction the vanadium ions undergo a valance change from +5 to +4. To reoxidize the Va+4 ions back to the +5 State, ADA is added as an oxygen carrier, and the ADA is subsequently regenerated with air. The process chemistry can be summarized as follows;
|CO3-2 + H2S||→||HS-+HCO3-|
A typical flow diagram of a Stretford unit is illustrated in Fig. 1.
Although the addition of vanadium to the Stretford Process increases the reaction rate of hydrosulfide ions to sulfur sufficiently to make the process commercial, it still produces a significant amount of byproduct thiosulfate. The reaction is still slow enough that air streams cannot be treated due to the high rate of thiosulfate formation.
Hundreds of Stretford units have been installed throughout the world; however, their popularity vanished in the mid 1980’s after the introduction of the chelated iron processes which addressed many of the deficiencies of the Stretford process
Chelated Iron-Based, Liquid Redox Processes
Iron is an excellent oxidizing agent for the conversion of H2S to elemental sulfur; however, due to the very low solubility of iron in aqueous solutions, the iron had to be present in the dry state (iron sponge) or in suspensions (the Ferrox process) or compounded with toxic materials such as cyanides. In the 1960’s development work was begun in England to increase the solubility of elemental iron in aqueous solutions. This work led to the introduction of the CIP process, CIP being an acronym for "Chelated Iron Process".
In this process, iron, in its’ ferric state (+3), is held in solution by a chelating agent, namely ethylenediaminetetraacetic acid (EDTA). The intent of the process was to oxidize sulfide (S=) and hydrosulfide (HS-) ions to elemental sulfur by the reduction of the ferric (Fe+3) iron to ferrous (Fe+2) iron, and the subsequent reoxidation of the ferrous ions to ferric ions by contact with air. The chemistry of all chelated iron processes is summarized as follows with (l) and (v) representing the liquid and vapor states, respectively;
|H2S(v) + H2O(1)||→||H2S(1)||(1)|
Equations 1 and 2 represents the absorption of H2S into the aqueous, chelated iron solution and its subsequent ionization, while equation 3 represents the oxidation of sulfide ions to elemental sulfur and the accompanying reduction of the ferric iron to the ferrous state. Equations 4 and 5 represents the absorption of oxygen into the aqueous solution followed by oxidation of the ferrous iron back to the ferric state.
Equations 3 and 5 are very rapid. This is in contrast to the oxidation reactions in the Stretford process when using vanadium. Consequently, iron-based systems generally produce relatively small amounts of byproduct thiosulfate ions, and in properly designed units, air streams can actually be processed. However, as in the Stretford process, equations 1 and 4 are relatively slow and are the rate controlling steps in all chelated iron processes.
It is interesting to note that the chelating agents do not appear in the process chemistry, and in the overall chemical reaction, the iron cancels out. So the obvious question is why is chelated iron required at all, if it doesn’t part take in the overall reaction. The iron serves two purposes in the process chemistry. First, it serves as an electron donor and acceptor or in other words, a reagent. Secondly, it serves as a catalyst in accelerating the overall reaction. Because of this dual purpose, the iron is often called a "catalytic reagent". The chelating agent(s) do not part take at all in the process chemistry. All the chelating agents do is to increase the solubility of iron in water, thus reducing the circulation rates required to furnish the two moles of iron required in equation 3.
Although it appears that chelated iron would solve many problems associated with previous liquid oxidation processes, the CIP Process failed miserably. In very short order after starting up, all of the iron precipitated out of solution as iron sulfide, FeS. The problem was that the chelation strength of many chelating agents varies considerably with solution pH, and unfortunately, the pH’s experienced in the CIP process were outside the range of EDTA.
In the early 1970’s, a small company in the Chicago area (ARI Technologies) started development work on a process, which employed multiple chelating agents. The idea being that by employing chelating agents with overlapping pH ranges where the chelation strengths were high, the iron would stay in solution at all times. This development work led to the introduction of the LO-CAT® process in the late 1970’s. The process worked well for units processing small quantities of H2S; however, in the first unit processing tons per day of sulfur, the process, in essence, failed. It was found that the chelating agents were disappearing very rapidly requiring extremely high chemical makeup rates.
Research found that the chelating agents were being oxidized to useless byproducts by a free radical mechanism. After a few years of experimentation, the chelate oxidation rate was reduced to an acceptable level by the introduction of free radical scavengers and by switching to chelating agents, which were much more resistant to oxidation.
Iron-based, liquid oxidation has developed into a very versatile processing scheme for treating gas streams containing moderate amounts of H2S. Advantages of these systems include the ability to treat both aerobic and non-aerobic gas streams, removal efficiencies in excess of 99.9%, essentially 100% turndown on H2S concentration and quantity, and the production of innocuous products and byproducts.
The three most common processing schemes encountered in iron-based, liquid oxidation systems are illustrated in Fig. 2 through 4. Fig. 2 shows a "Conventional" unit, which is employed for processing gas streams, which are either combustible or cannot be contaminated with air such as carbon dioxide, which is being treated for beverage purposes. In this scheme, equations 1 through 3 are performed in the Absorber while equations 4 and 5 are performed in the oxidizer. Fig. 3 illustrates an "Autocirculation" unit, which is used for processing acid gas (CO2 and H2S) streams or for other non-combustible streams, which can be contaminated with air. In this scheme, equations 1 through 3 are performed in the "Centerwell" which is nothing more than a piece of pipe open on each end. The purpose of the centerwell is to separate the sulfide ions. from the air to minimize byproduct formation. The volume within the centerwell is essentially the same as the absorber in a conventional unit. The other unique feature of the Autocirculation scheme is that no pumps are required to circulate solution between the centerwell (absorber) and the oxidizer. In these units there is a larger volume of air than acid gas; consequently, the aerated density on the outside of the centerwell is less than on the inside resulting in a natural circulation from the oxidizer into the centerwell.
The last type of processing scheme (Fig. 4) is the aerobic unit (air contaminated with H2S) in which equations 1 through 5 all occur within the same vessel, at the same time and without separation of the absorber and the oxidizer. These are generally less expensive units than the other two schemes; however, because there is always oxygen in the presence of sulfide ions, consequently, these units produce the most byproducts.
Although references are continuously made that iron-based, liquid redox systems are plagued with plugging and foaming problems and that the process cannot be operated at high pressure due to pump and foaming problems, these problems for the most part have been solved for some time.
Foaming occurrences are either a start-up phenomenon or the result of large amounts of liquid hydrocarbons entering the unit. During the initial start-up of a unit, the surface tension properties of the fresh solution are such that the foaming may occur during the first few days of operation. However, by following proper start-up procedures, this foaming is easily avoided. In addition, this foaming tendency is only experienced when the entire unit is filled with fresh solution, which only happens during the initial start-up of the unit. Foaming does not occur during subsequent start-ups.
Continuous incursions of small amounts of liquid hydrocarbons are frequently experienced with no adverse effect on the operation of a unit; however, the introduction of large amounts of liquid hydrocarbons can present foaming and plugging problems. This would also be true of Claus units, selective oxidation processes and hydrocarbon-based, redox systems. However, for aqueous-based redox systems, "Designer" surfactants1 have been developed, which in essence totally alleviates the problems caused by the introduction of large amounts of liquid hydrocarbons.
During the early days of liquid redox, sulfur plugging was a severe operating problem. Packing plugged, static mixers plugged, pipes plugged, heat exchangers plugged and distributors plugged. For the most part, all of these plugging problems have been eliminated. Vessels with random packing are no longer used, on-line cleaning procedures have been developed for static mixers1, which require very little operator attention, proper pipe design has eliminated pipe plugging, proprietary heat exchanger designs and proper operating procedures have minimized heat exchanger plugging, newly designed absorber spargers are being installed, which have greatly extended the life of sour gas spargers and improved quality control of oxidizer sparger materials and proper operation of the process has minimized oxidizer sparger plugging.
Operation of aqueous-based liquid redox systems at high pressure has been a problem due to difficulties with keeping the liquid circulation pumps running. Circulation pumps were always specified as ANSI, open-impeller centrifugal pumps. The logic being that closed-impeller pumps would plug with sulfur particles or possibly erode. Consequently, for high head applications in which open-impeller pumps would not apply, plunger type pumps were chosen. The plunger pumps had no difficulty supplying the required head; however, seal rings had extremely short lives. To solve this problem, a multi-staged, closed-impeller, centrifugal pump was installed in one high pressure application with excellent results. The pump was in continuos operation for approximately1_ years without any signs of plugging or erosion. For all future high-pressure applications, closed-impeller single or multi-stage centrifugal pumps will be specified. Obviously, the original concern about plugging had no basis.
Although iron-based, liquid redox processes have gained acceptance as evidenced by over 150 units being licensed worldwide, there are still areas in the process, which need to be improved upon. Current areas of R&D efforts are reduction in operating costs, reduction in equipment size and improvement in molten sulfur color.
Operating costs for aqueous, iron-based redox systems are composed of replacing chemicals which are either oxidized in the unit or which are physically lost from the unit and of electrical power required for circulating solution and injecting air. For any iron-based system there is an economic tradeoff between iron concentration and the solution circulation rate. Since 2 moles of iron are required for every mole of hydrogen sulfide (equation 3), the amount of circulating solution required is dependent on the iron concentration in the solution and the amount of H2S in the sour gas stream — the higher the iron concentration, the lower the circulation rate and hence, the lower the power consumption. Conversely, the higher the iron concentration, the higher the catalyst makeup rate required to replace iron from physical losses such as solution lost with sulfur withdrawal. There is an optimum iron concentration based on the incremental cost of power and the amount of solution, which is normally lost from the system. This relationship is shown in Figure 5.
Besides replacing iron, which is physically lost from the system with the sulfur and blowdown streams, chelating agents are chemically oxidized into useless, non-toxic byproducts within the system and must be replaced. As stated previously, different chelating agents have different resistances to chemical oxidation. In addition, chemicals may be added to the system or made in the unit, which act as free radical scavengers, thus retarding chelate oxidation. Research continues on the development of oxidation resistant chelates and on economical, free radical scavengers. For example, many compounds from the polyamine family have proven to be excellent free radical scavengers reducing chelate degradation to essentially zero. Unfortunately, the oxidation rates of the polyamines are extremely high and consequently, uneconomical. Also new chelating agents have been developed which have very high resistances to oxidation; however, they currently are uneconomical to manufacture. The search for stable chelating agents and/or economical free radical scavengers continues.
A large portion of the electrical consumption in an iron-based, liquid oxidation system is associated with blowing air through the solution to satisfy the oxygen demand of equation 5. Due to the low solubility of oxygen in water, a large excess of air is generally employed depending on oxidizer design. Although the air is free, the back pressure on the blowers is usually between 0.5 and 1.0 bar (g), thus the blower usually represents a large portion of the unit’s electrical consumption.
A considerable amount of research is currently underway to develop new mass transfer devices which will improve the oxygen utilization in liquid oxidation systems with the aim of reducing the quantity of air required (operating cost) and reducing the size of the oxidizing vessels (capital cost). Currently, there are two types of oxidizers employed -low head and high head oxidizers. In the low head oxidizers air is sparged through approximately 3 meters of solution at superficial air velocities of less than 3.5 M/min by means of distributors equipped with EPDM sleeves which are perforated with very small holes. Solution flow is perpendicular to the airflow. These oxidizers are relatively poor mass transfer devices; however, they do provide much need solution inventory for proper operation of the system.
In high head oxidizers air is sparged through approximately 7 meters of solution at superficial velocities of greater than 10 M/min by means of course bubble pipe distributors. Solution flow is co-current to the airflow. These oxidizers provide mass transfer coefficients which are approximately 4 times better than the low head oxidizers; however, this is at the expense of higher discharge heads on the air blowers.
Neither the low or high head oxidizers are very good mass transfer devices; however, they generally do not plug with sulfur and they do supply solution inventory required for proper operation of the system. A main goal of current research is to develop a mass transfer device, which will reduce the amount of air required to approximately stoichiometric quantities, will reduce the oxidizer volume and will not plug with sulfur. To this end, special synthetic membranes are being tested. Initial results indicate that the first two objectives of the research — stoichiometric air and high mass transfer coefficients — can be obtained. Long term testing is currently being carried out to determine the plugging tendencies of the membranes. If this last phase of testing is successful, liquid oxidation systems will become much smaller and less expensive to operate.
Sulfur produced from liquid redox systems has the same chemical assay as Claus sulfur, and it does have several commercial uses2 in its unmelted form. In fact Lubrizol in France has been recycling its produced sulfur with no adverse effect for quite some time. However, there is always a desire to improve the appearance (color) of redox sulfur, which is degraded due to the presence of iron polysulfides. Development work is being conducted on a variety of filter/wash systems with the goal of removing the chelated iron from the sulfur cake prior to final use. However, due to the extremely low price of sulfur and due to the relatively low quantities of sulfur produced in liquid redox plants, it is difficult justifying much work in this area.
Liquid phase oxidation systems have undergone considerable evolution during the 20th century, and this will continue into the 21st century. Foreseeable developments for the near future will be smaller equipment sizes and lower operating costs which will be achieved by the development of better oxygen mass transfer devices reducing the amount of air required and the size of oxidizers and by the addition of free radical scavengers into the systems.
1. Reicher, Myron, Niemiec, Bill & Katona, Tamas, "The Use of Liquid Redox Desulfurization Technology for the treatment of Sour Associated Gases at Elevated Pressure", presented at the GPA European Chapter Continental Meeting, 1999
2. Nagl, Gary, " Emerging Markets for Liquid Redox Sulfur", presented at Sulphur97, Vienna, Austria
Table I. Liquid Phase Oxidation Processes
|Polythionate Solutions||Feld Process
Koppers C.A.S. Process
|Iron Oxide Suspensions||Burkheiser Process
|Thioarsenate Solutions||Thylox Process
|Iron Cyanide Solutions||Fischer Process
|ADA Solutions||Stretford Process (4)
Takahax Process (4)
|Iron Complex Solutions||Cataban Process
Sulferox Process (4)
LO-CAT Process (4)
LO-CAT II Process (4)
|(4) Indicates commercial success|