H2S REMOVAL SYSTEM SHOWS PROMISE OVER IRON SPONGE

Alvin Samuels
Gas Sweetener Associates Inc.
Metairie, La.

Use of a new material as a replacement product for iron sponge (wood chips) for H2S removal has been introduced at two Shell Western E&P central production facilities in Michigan.

At the Chester 21 facility, operating costs between March and August 1989 were reduced by 13% over the same period for 1988. Winter savings are projected at 21%.

The new trademarked material is named SulfaTreat. It is a dry, free-flowing material being used in patented and patent-pending processes for selective removal of up to 2,800 ppm of H2S. The material is non-toxic, nonpyrophoric, and environmentally safe in both its unreacted and ready-for-disposal form.

Control of pH is not required and the process generates no toxic or corrosive gases.

Bed life is predictable with SulfaTreat which allows for the timely scheduling of changeouts and reduces undesirable shutdown from high H2S outlet.

INCREASED COSTS

Shell has been using wood chips at the Chester 21 central production facility in Michigan and has experienced increased operating costs during the winter months. Shell decided to test SulfaTreat at Chester 21 beginning in March 1989.

The material requires no water to be reactive. Therefore, there is no risk of the process failure that occurs when the iron hydrate in iron sponge dries out. Nonetheless, some moisture in the gas stream is desirable to prevent reduction in the rate of reaction.

Because SulfaTreat can operate below the minimum moisture content of wood chips, hydrate formation within the SulfaTreat bed is reduced, decreasing possible channeling.

Because the pressure drop across the SulfaTreat bed is very low, there is little possibility of methane hydrates forming in expanded and cooled gas at the low-pressure side.

The material can work effectively in acidic conditions, eliminating the need for soda ash to buffer the pH, and thereby reducing possible iron carbonate and calcium carbonate scaling.

Another operating problem experienced with wood chip material at Chester 21 is the condensing or carryover of liquid hydrocarbons in the inlet piping or the vessel. Such contaminants can coat the iron-sponge media and prevent complete reaction of the iron oxide.

With SulfaTreat, this contamination appears to have little or no effect.

Finally, odor problems are less troublesome because the material used has less of a tendency to absorb hydrocarbons than do the wood chips.

The Chester 21 vessels consist of two towers piped lead-lag. Each tower is a 45 in. ID by 9-ft high vessel

VARYING H2S LEVELS

The operating conditions for the SulfaTreat runs are summarized in Table 1. Each run had varying H2S levels and flow rates, depending upon which combination of three input wells was being treated.

For Run 1, the pressure was increased from 350 to 750 psi on Day 9. All other runs were at 750 psi.

For Runs 1 and 2, the inlet vessel was changed out when the outlet H2S level in the polishing tower reached 10 ppm.

For all other runs except Run 9, the inlet vessel was changed out before the polishing tower showed any H2S. This may have resulted in lower efficiencies than would have been achieved otherwise.

For optimum efficiency, the second tower outlet's H2S level should be allowed to approach contract levels before changeout.

Run 8 results were lower than typical because this run changed prematurely and was insufficiently loaded with SulfaTreat. Results have demonstrated bed life to be same for the colder months of March and April and the warmer months of May, June, July, and August.

Bed life averages 2.5 times longer and results in 781 lb of H2S removed per changeout.

This compares with an average of 310 lb of H2S removed per changeout with iron sponge.

Field measurements of H2S content made on the inlet and outlet gas from the inlet tower were made with a Sensidyne detector. One comparison involving analysis by gas chromatograph indicated an H2S inlet of 1,100 ppm vs. 620 ppm with Sensidyne.

Laboratory analysis of reacted SulfaTreat with Leco analysis showed that the spent oxide contained 25% sulfur.

Wet washing showed iron content of 18.8%.

This much sulfur in the reacted material would have required inlet H2S levels substantially in agreement with the chromatograph. Leco readings on Run 8 showed only 23.6% sulfur and 22.1% iron, indicating that reactive capacity had not been fully utilized.

The large discrepancy between these readings and the field measurements is subject to further analysis. Nonetheless, all calculations contained herein are based on the field Sensidyne readings. This would suggest that the reported results are conservative for both SulfaTreat and wood chips.

ECONOMICS

Table 2 shows that the treating costs for the Chester 21 have been reduced by 13% in the months of March through August 1989, over the same period for the previous year. The costs include labor, material, and disposal costs, which average $6.16/lb of H2S removed for iron sponge and $5.37/lb of H2S removed for SulfaTreat.

The cost reduction attributable to material is expected to increase to 21% for the months of September through February because iron sponge costs are historically higher during those colder months.

Costs are not expected to increase in the winter because the process should not be affected by methanol injection and hydrate formation which adversely affect iron sponge.

The labor and disposal costs are approximately the same for SulfaTreat ($1,661/changeout) and iron sponge ($1,581/changeout) in spite of the reduced time and equipment required by the SulfaTreat process.

It is expected that further savings will be achieved as competition and experience with the process increase. The material costs ($2,433 vs. $445) and the overall cost per changeout ($4,094 vs. $2,026) are higher with SulfaTreat.

FIELD OPERATIONS

The accompanying box summarizes the changeout procedure.

The time required to remove the material from a contactor varies from 30 min to 1 hr and compares with a minimum of 1 hr and up to 6 hr for iron sponge.

A primary reason for the ease of cleanout is the uniform particle size and shape of the material. This uniformity not only reduces pressure drop, which is so low that it is not measurable, but also minimizes the caking and cementing common to iron sponge.

Bed porosity and permeability are uniform from run to run. The uniform porosity and permeability reduce the possibility of gas channeling which can result in substantial variation in iron sponge bed life.

Hydrocarbon precipitation is substantial in this location with up to 10 b/d condensing in the inlet vessel. No adverse effects have been observed resulting from hydrocarbon contamination.

With iron sponge, this contamination has resulted in short runs as a result of apparent interference of hydrocarbons in the reaction with hydrogen sulfide (Fig. 1).

The difference in process reliability between iron sponge and SulfaTreat is easily seen in the variation in run life as shown in Fig. 1.

The amount of H2S removed per changeout varies from 101 to 732 lb with an average of 310 lb of H2S removed with 1.7 off-spec days/month with iron sponge. This compares with a range of 393-1,102 lb with an average of 781 lb of H2S removed and only 0.2 off-spec days/month using SulfaTreat. (Offspec readings are shown in parentheses.)

Iron sponge run life can be influenced by poor loading of the iron sponge, hydrocarbon contamination, methanol injection, drying of the iron oxide hydrate, pH variation, hydrate formation, and low gas velocity.

None of these variables has been observed to affect the SulfaTreat process.

RETROFITTING VESSELS

Installation of suitable retainer screens and a "filterfoam" sponge-like material is all that is required to convert an iron-sponge vessel to Sulfatreat. This assumes that a support tray is already in place.

The bulk density of SulfaTreat is 70 lb/cu ft and results in an increased load compared with the bulk density of iron sponge of about 27 lb/cu ft.

SulfaTreat is a dry, freeflowing material consisting of an inert substrate with uniform permeability and porosity and containing a proprietary reactive iron compound.

The particle size of the material varies from 4 mesh to 30 mesh, and it has a bulk density of 70 lb/cu ft. The material is packaged in 50-lb bags and in bulk bags of up to 2,500 lb capacity.

STOICHIOMETRY, REACTION RATE

The iron oxide in SulfaTreat is present in two forms, Fe2O3 and Fe3O4. Reaction of these oxides with H2S follows several simultaneous paths as shown in the accompanying box.

Although the extent of each reaction is not known, the net overall efficiency E = lb of H2S reacted/lb of oxide is between 0.55 and 0.716 based on field measurements.

Because it takes a short time for H2S to react with the iron oxide in SulfaTreat, the time rate of change of H2S in the gas and the time rate of change of oxide must be described.

This has been accomplished by suitable differential equations incorporated into a computer program.

With these equations, the outlet level of H2S can be calculated for any set of field operating conditions. The variables in these calculations are inlet H2S (PPM), inlet mercaptan concentration (ppm), gas flow rate (MMscfd), inlet pressure (psig), reactor dimensions, and gas temperature.

Two important properties of the oxide that are constant regardless of gas flow rate, pressure, or H2S level are the rate constant (R) and the maximum efficiency (E) which has already been discussed.

R and E are intrinsic properties that combine to determine the amount of gas that can be processed.

The rate constant has been determined in laboratory experiments on water-saturated gas to depend on temperature: R = 0.0002 (40 F.), R 0.0005 (77 F.), and R 0.0012 (130 F.)

The rate constant increases as temperature increases higher than 40 F. Operation below 40 F. is not recommended.

The rate constant can also decrease if a dry gas is treated.

For example at a relative humidity of 23%, R (at 130 F.) 0.0004; at 100%, 0.0012.

If dry gas is being treated, water injection may be desirable in order to increase the rate constant and allow for more effective removal of H2S.

PRESSURE DROP VELOCITIES

A bed of SulfaTreat has a uniform porosity and permeability and offers small resistance to flow.

Laboratory experiments with a water differential manometer resulted in the following equation relating pressure drop, dP (psi), to gas velocity, y (fpm), and bed height, H (ft): dP = 0.0009 x v x H

For example, if flow velocity is 4 fpm through a 10-ft bed, the pressure drop along the bed will only be 0.036 psi. This was confirmed at the Chester 21 where the pressure drop through two 9-ft beds in series was too low to be measured.

The material has proven effective at treating H2S at gas rising velocities as low as 0.62 fpm (see Run 3 in Table 1).

This is another reflection of the uniform porosity and permeability of the SulfaTreat medium.

A low-velocity limit has not yet been determined for Sulfatreat.

EPA CONSIDERATIONS

Evaluation of both unreacted and reacted SulfaTreat with "EPA Guidelines for the Identification and Listing of Hazardous Waste" and the MSDS data show that SulfaTreat is classified as nontoxic, nonhazardous waste.

In Texas it has been approved for burial on site. In Michigan it has been approved for disposal in a Class 11 landfill suitable for nonhazardous materials.

Evaluation of samples obtained from Chester 21 Runs 8 and 12 show that the dried, reacted material meets the criteria for nonhazardous classification based on reactivity, ignitability, and corrosivity.

Previously the material has been determined to be free of heavy metals and pesticides.

Oral and dermal toxicity tests to evaluate possible personnel handling and health problems also showed the product to be safe and nonhazardous.