New filtration process cuts contaminants from offshore produced water

Syed A. Ali, Larry R. Henry
Chevron U.S.A. Production Co.
New Orleans
Jerald W. Darlington
Cetco
Chicago
John Occapinti
Cetco
Lafayette, La.

The 6 ft x 8 ft skid quick-change, produced-water polishing units take up a relatively small space on the platform (Fig. 2).
Laboratory and field tests demonstrate that a two-part filtration process can effectively reduce oil and grease content in offshore-discharged produced water.

The first part of the process is the patented Crudesorb technology and the second part is a proprietary patented polymeric resin. Crudesorb removes dispersed oil and grease droplets from the produced water, and thereby protects the resin from being coated with the oil droplets.

The resin removes dissolved hydrocarbons, aliphatic carboxylic acids, cyclic carboxylic acids, aromatic carboxylic acids, and phenolic compounds.

Field trials have demonstrated that after 6 months of produced water passing through the filtration system, the oil and grease values were 100% in compliance with the federally mandated level of 29 ppm.

Produced waters

During crude oil production, a significant quantity of water is typically coproduced. Water represents the largest waste stream in oil production.

Typically, a new field produces relatively little water, such as 5-15% of the produced stream. But as the field ages, water volumes can increase significantly to the 75-95% range. Waterflooding, for recovering additional oil, compounds the water-handling problem.

Offshore production facilities primarily treat oil/water mixtures by mechanically separating the fluids. After removing as much oil as possible, the produced water is typically discharged below the sea surface.

These discharges are a potential pollution problem because oil and grease could create a toxic effect at the discharge point and also sheen the sea surface.

The U.S. Environmental Protection Agency (EPA), as of August 1997, requires a maximum extractable content of 29 ppm for produced water discharged in the Gulf of Mexico.

Typically, the oil and grease content of produced water is classified as either dispersed or soluble oil. This article will further differentiate the soluble oil category into two subgroups: partially soluble and very soluble.

This differentiation is necessary because of the distinct solubility profiles, the difference in component concentration in the produced water, and most importantly, the contribution to the measured oil and grease content.

Dispersed oil

Dispersed oil consists of small individual oil droplets that are suspended in the aqueous phase. In general, oil drops are sheared into smaller droplets by the differential pressure and mechanical forces of flow restrictions such as pipes, valves, pumps, and other mechanical devices that constrain flow or induce turbulent flow.

These droplets become dispersed in the aqueous phase because of the low interfacial tension between the oil and water layers.

A small concentration of dispersed droplets remains in the produced water because of soluble organics and treatment chemicals that further lower interfacial tension between the oil and water layers.

Several factors affect the dispersed oil concentration. These factors include:

Oil density
Shear encountered by the oil/water mixture
Interfacial tension between the oil and water phases
Chemical treatments
Separation equipment.

Commonly used chemical treatments and equipment attempt to force these individual droplets to combine until the agglomerated oil drop size is sufficient for recovery with coalescence and separation, prior to the water being discharged to the sea.

Once discharged, the dispersed water can have several effects on the surrounding area. The effects depend on the distance between the discharge point and the sea bottom and surface, as well as sea motion at the location.

If the produced water discharge contacts the sea bottom, oil can contaminate and accumulate in the sea sediments. Motion caused by tides, currents, wave action, and vertical mixing can effect this accumulation cycle in an unpredictable fashion.

Dispersed oil droplets not accumulated on the bottom eventually rise to the surface and spread. Depending on surface conditions, an oil sheen may form. Sheen is the most visible evidence of pollution that causes concern.

Dispersed oil concentration can be monitored with visual and analytical techniques. Visually, when one holds a sample up to a light, the dispersed droplets can be seen in the aqueous phase or sheen is seen on the sample surface. In addition, the sample will often be turbid and cloudy.

Because dispersed droplets are very soluble in organic solvents, their concentration can be monitored by extracting the sample with a water-immiscible organic solvent (freon or hexane) and the oil and grease extracted then can be measured by infrared or gravimetric analysis.

Partially soluble components

Partially soluble components are medium to higher molecular weight (C ₆ to C ₁₅) aliphatic and aromatic carboxylic acids, phenols, and aliphatic and aromatic hydrocarbons. These components are soluble in water at low concentrations, but not infinitely soluble as are their lower molecular weight analogs.

These components are difficult to remove from produced water and for the most part, these components are being discharged directly into the sea.

Partially soluble compounds contribute to sheen, but the primary concern is the potential for toxicity. In a normal produced-water stream with a pH of about 6-8, the carboxylic compounds are primarily present in the sodium salt form.

One widely practiced technique for removing these partially soluble components is to acidize the produced water. This converts the sodium carboxylate compound to a carboxylic acid compound because the carboxylic acid compound has less solubility in the aqueous phase than the sodium carboxylate compound.

Acidizing does have problems such as handling large acid volumes, having a relatively high cost, dealing with corrosion and subsequent replacement of equipment, and ensuring a consistently low oil and grease measurement.

Because carboxylic acid and phenolic compounds limit solubility in water, the organic solvents (freon or hexane) used in the EPA oil and grease methods extract at least a small percentage of the compound present in the produced water.

Carboxylic acids and phenolic compounds are not volatilized during the solvent evaporation step of the gravimetric oil and grease analysis because of their molecular weight and vapor pressure profiles. Therefore, these compounds contribute to the oil and grease results obtained by infrared and gravimetric analysis.

Very soluble components

Very soluble components are low molecular weight (C ₂-C ₅) carboxylic acids, ketones, and alcohols that are very soluble in produced water. Components in this category include acetic and propionic acid, acetone, and methanol. The concentration of these components in some produced water has been measured in excess of 5,000 ppm.

Because the components are highly soluble in the aqueous phase, typical separation equipment on production platforms does not remove them and acidizing the produced water stream is ineffective at promoting the separation.

Also, because of their high solubility in the aqueous phase, the organic solvent in the oil and grease analysis does not extract virtually any of these components. Therefore, these components do not significantly contribute to the oil and grease measured in spite of their high concentration in the produced water.

These components do present an additional problem in that their high concentration can consume a large portion of a nonselective treatment system.

Attempts have been made to remove these oil and grease contributors from the stream before discharge. Zaidi, et al., evaluated the use of membranes for removing oil and suspended solids.¹ Simmons has investigated reverse osmosis membranes.² Tellez and Khandan investigated biological treatment.³

The patent literature also focuses on ion-exchange resins. Means, et al.,⁴⁵ was granted two patents for removing oil with a styrene/divinylbenzene polymer.

The authors believe all these attempts to remove the oil and grease contributors have had limited success. References ⁶ and ⁷ contain excellent reviews about produced water.

Analytical procedures

The laboratory procedure for determining the removal efficiency for each component consisted of pumping aqueous solution containing oil and/or carboxylic acids and phenolic compounds through multiple beds of media.

The evaluation used a clean 250-ml burette with about 2 in. of glass wool at the bottom to prevent the media from escaping the burette. The burette was filled with water to purge air from the system. In a 500-ml beaker, about 100 g of media was stirred in deionized water to deaerate the media pores. After mixing for 1 hr, the media was slurried into the burette in a manner that prevented introduction of air into the media pores.

The media slurry was added, and the media was fluidized by backflushing with deionized water at a sufficient flow rate to agitate the bed. When all of the media was fluidized, the water flow was terminated and the bed allowed to settle. Two inches of additional glass wool was lightly placed on top of the media to prevent the aqueous solution from disturbing the media bed.

A synthetic dispersed-oil solution was prepared by adding 300 ppm of Aldrich vacuum pump oil to an aqueous deionized water solution containing 4.0% sodium chloride and 0.5% calcium chloride. The oil/salt water solution was emulsified with an overhead rotary stirrer and pumped at a rate of 10 bed volumes/hr into the 250-ml burette containing the media.

Oil and grease breakthrough (EPA Method 413.1) was monitored qualitatively by visual observation and quantitatively by freon extraction.

The synthetic carboxylic acid and phenol solution was prepared by adding 50 ppm of phenol, benzoic acid, heptanoic acid, and napthalenic acid to deionized water containing 4.0% sodium chloride and 0.5% calcium chloride. The mixture was stirred for 24 hr to dissolve all organic compounds.

The carboxylic acid/salt water solution was stirred with an overhead rotary stirrer and pumped at a rate of 10 bed volumes/hr into the 250-ml burette containing the media.

Breakthrough of the carboxylic acid/phenol compounds was monitored quantitatively with freon extraction (EPA Method 413.1).

Liquid chromatography (Waters high-pressure liquid chromatograph with a C8 column) on the media effluent determined the organic components not absorbed by the media. The mobile phase was 20% acetonitrile/80% water with an isocratic flow rate of 1.5 ml/min. A Waters UV/VIS detector and a wavelength of 210 nanometers was used.

During pilot testing, a rapid qualitative field analysis was done on the production platform.

All samples were acidified to a pH of 2 with hydrochloric acid before the analysis commenced. The field analysis used 100 ml of freon and 1 l. of produced water. The two solvents were vigorously mixed in a 2-l. separation funnel, and the phases were allowed to settle.

To obtain an absorbency measurement, the freon phase was collected and an aliquot was placed into an infrared cell that was inserted into a previously calibrated infrared spectrophotometer. The oil and grease content in the sample was determined by taking the absorbency reading from the sample and comparing it to a previously established chart.

Commercial laboratories performed qualitative field analyses in accordance with EPA Method 413.1.

The samples were acidified to a pH of 2 with hydrochloric acid prior to sending ashore. The typical time from sample collection to analysis was 4 days.

Results

The initial laboratory investigation consisted of analytically determining the components present in a commercial sample of offshore produced water. The primary goal was to ensure that Cetco's analytical results were consistent with published literature.

A sample of produced water was obtained from a Chevron Main Pass production platform. It was extracted with EPA Methods 413.1 and 1664, and the extracted portion was analyzed by GC-MS (gas chromatography and mass spectrometry). The results (Table 1 [31,962 bytes]) show that the Cetco analytical methods were consistent with established composition profiles of offshore produced water.

A synthetic produced water was then prepared based on these data and literature data.

As previously discussed, oil and grease contributors consist of dispersed oil and partially soluble organics that are primarily carboxylic acids and phenols.

Filter media

The first problem was to remove the dispersed oil because it is a main contributor to the measured oil and grease content. More importantly, the dispersed oil droplets will foul any media designed to remove the partially soluble components. Therefore, a specialized adsorption filter media was used for removing emulsified and dispersed oil from the aqueous phase prior to overboard discharge.

The filter media, Crudesorb, is a proprietary patented granular media. It is designed for removing oil and grease from aqueous streams and is formulated to withstand several common solvents used in oil production.

The filter media effectively removes free, emulsified, and dissolved organics that are sparingly water soluble. Crudesorb has a very short mass transfer zone and can adsorb up to two times its weight in hydrocarbons.

Because of its high BTU value, the Crudesorb can be either disposed of as normal oil field waste in the cutting box or incinerated. Based on the documented success of this filter media in produced water, deck drainage, and sump drainage applications, the researchers were highly confident that it would successfully remove dispersed oil from produced water.

This filter media is very important from a practical sense because the dispersed oil would coat most other media rendering them useless. In addition, dispersed oil tends to plug pore openings in activated carbon and other porous media. The plugging significantly diminishes the media's capacity because of reduced surface area and reactive sites. Thus, without Crudesorb, the media has very limited ability to remove partially soluble carboxylic acids and phenols.

Other media

The next phase of the investigation evaluated the carboxylic acid removal efficiencies of several commercially available media. A stream of laboratory-generated synthetic produced water was pumped as previously described through a column of Crudesorb and into a column containing the media being evaluated.

After passing through the adsorbing media bed, the effluent was analyzed for remaining freon and hexane-extractable material, and removal efficiencies were determined from these data.

Odor and color were also monitored because these simple measurements gave a quick indication of the resin bed capability and capacity. Table 2 [108,689 bytes] shows the results.

The data indicate that the adsorption capacity of several common adsorbents was quickly consumed when the adsorbent was exposed to the synthetic produced water.

Several types of activated carbon were evaluated because activated carbon is well known for its organic removal capability. But the results show that activated carbon was not successful at long-term removal of carboxylic acids, and its on-site regeneration would be problematic.

Several weak and strong base ion-exchange resins were evaluated because these resins should be able to interact with carboxylic acids; however, the results indicate that the basic ion-exchange capacity was quickly exhausted. The rapid consumption of ion-exchange capacity was believed to have been caused by the produced water containing about 5% salt.

Table 2 shows that the proprietary Resin A reduced the hexane-extractable material, odor, and color of the produced water significantly more than the other treatment agents evaluated. This resin removed at least twice as much as most of the other resins. Further investigation revealed that the resin had specific interaction potential with carboxylic acids and phenols and could be regenerated with methanol, acetone, hydrochloric acid, sulfuric acid, and sodium hydroxide.

An added benefit is that this resin had only limited attraction for acetic acid, propionic acid, and positively charged metallic cations.

The investigation also showed that Crudesorb was required because the resin quickly and irreversibly fouled when contacted by dispersed oil.

Large-scale test

Crudesorb and Resin A were evaluated in a large-scale laboratory test conducted with actual produced water samples from several platforms in the Gulf of Mexico.

The laboratory procedure consisted of pumping produced water through a 1-in. diameter column filled with 150 ml of Crudesorb followed by a second 1-in. column filled with 150 ml of Resin A. The produced water influent and effluent from each column was monitored by the gravimetric EPA method 413.1.

The results clearly showed that the combination of media lowered the freon-extractable material to below the federally mandated level of 29 ppm.

Sufficient material was obtained to run 200 bed volumes into the system.

The results for Main Pass Field A (Table 3 [78,330 bytes]) show very high oil and grease content. The reason for this is that it was not possible to obtain a 5-gal sample from the normal sample point that accommodates the 1 l. sample container. To obtain the sample, the platform technician dipped the bucket into the last port of the flotation cell; thus, the sample contained material that would not normally be discharged as oil and grease.

In spite of this large abnormality, the results indicated that the combined media remove enough contaminants to meet regulatory targets.

Field test

To evaluate the system under actual field conditions, Cetco conducted a 6-month field trial at Chevron's Main Pass Field A platform. At that time, the platform discharged about 180 gpm of produced water.

For the test, Chevron diverted a small, 30-gpm, produced-water stream from the platform's Wemco dissolved-gas flotation separator. This 30-gpm stream passed initially through a bag filter to remove suspended particles. The produced water then was pumped through two vessels of Crudesorb in series and subsequently through a vessel containing Resin A (Fig. 1 [808,563 bytes] and Fig. 2).

The 30-gpm stream was then discharged in the same manner as the original 180 gpm.

Fig. 3 [339,204 bytes] shows the analytical results as determined by a freon extraction followed by a gravimetric determination of the oil and grease. The data indicated that the produced-water composition after the Wemco is highly variable and ranges from about 10 to 47 ppm. The Wemco produced water had a significant odor and was greenish in color.

After passing through the Crude sorb columns, the oil and grease content were reduced to the 20-25 ppm range. The treated water was clear in color but still had significant odor. But after passing through Resin A, the odor was removed from the treated produced water and the oil and grease were reduced to 5-10 ppm for the first 600 bed volumes.

During the 600-900 bed volumes, breakthrough occurred on the resin, and the resin capacity was exhausted after about 1,000 bed volumes.

The resin was regenerated by backflushing methanol at a rate of 1 bed volume/hr for 4 hr. The methanol solubilizes the carboxylic acids and phenols and removes them from the resin. During the regeneration cycle, the resin bed swelled about 25% over the original volume; thus, sufficient head space must be available in the container.

Following the methanol regeneration, water was backflushed at a rate of 10 gpm for 5 min to refluidize the resin bed and remove any trapped gases and methanol. When the bed had resettled, the produced water flow resumed. The methanol was used for several cycles without determining the point at which removal of resin contaminants was not achieved.

The methanol is an ideal regeneration solvent because once spent, it can be disposed of by pumping it into the gas or oil sales line. During this trial, the methanol was pumped into the gas line to prevent hydrates.

In this 6-month trial, neither the Crudesorb nor Resin A was replaced. The Crudesorb vessels were backflushed to remove accumulated solids. Resin A was regenerated with methanol.

The 6-month trial term is believed to be sufficient to indicate that the media can successfully withstand all of the production chemicals used on the platform.

The data show that the produced water exiting the Wemco is highly variable and ranges from about 10 to 47 ppm. But after passing through the Crudesorb, it is significant to note that every data point is below the federally mandated 29 ppm oil and grease value.

The data illustrate that the main contributor to the measured oil and grease values is from the partially soluble components instead of the dispersed oil. To consistently obtain values below 20 ppm, Resin A was required.

For fields where a greater percentage of the oil and grease value contributors are from the partially soluble components, both Resin A and Crude sorb are required.

Acknowledgment

The authors thank the management of Chevron U.S.A. Production Co. and Cetco for granting permission to publish this article.

References

Ray, J.P., and Engelhart, F.R., (eds.), Produced Water, Technological/Environmental Issues and Solutions, Environmental Science Research, Vol. 46, Plenum Press, New York, 1992.
Reed, M., and Johnsen, S., (eds.), Produced Water 2, Environmental Issues and Mitigation Technologies, Environmental Science Research, Vol. 52, Plenum Press, New York, 1996.
Means, C.M., and Braden, M.L., "Process for Removing Water-Soluble Organic Compounds from Produced Water," U.S. Patent 5,104,545, Apr. 14, 1992.
Means, C.M., and Braden, M.L., "Process for Removing Water-Soluble Organic Compounds from Produced Water," U.S. Patent 5,135,656, Aug. 4, 1992.
Beall, G.W., "Method of Removing Water-Insoluble Organic Contaminents from an Acidic Aqueous Stream," U.S. Patent 5,567,318, Oct. 22, 1996.
Ali, S.A., Hill, D.G., McConnell, S.B., and Johnson, M.R., "Process, optimized acidizing reduces production facility uspsets," OGJ, Feb. 10, 1997.

The Authors

Syed A. Ali is a technical advisor for Chevron U.S.A. Production Co. in New Orleans. He specializes in sandstone acidizing, formation damage control, rock-fluid interaction, mineralogy, and oil-field chemistry. Ali has an MS from Ohio State University and a PhD from Rensselaer Polytechnic Institute.

Larry R. Henry is a senior environmental advisor for Chevron U.S.A. Production Co.'s Gulf of Mexico deepwater and shelf business units. His primary technical focus is in treatment and in environmental and regulatory impacts of exploration and production wastewater discharges.
Henry is a 1969 graduate of St. Louis University and a member of SPE. He serves on several industry and agency cooperative committees working to improve the overall compliance, communication, and effectiveness of state and federal water discharge regulatory programs.

Jerald W. Darlington is technical director of Cetco a division of Amcol International. His main areas of interest include water treatment, flocculation, and filtration technology for industrial use. Darlington has a BS and MS in chemistry from Baylor University.

John Occapinti is manager of product development with the environmental offshore unit of Cetco. He has held various positions in the environmental wastewater treatment industry. His interests include the filtration and remediation of environmental sensitive fluids. Occapinti has a BS in chemical engineering from the University of Waterloo.