PureLine continues its series on surveying chlorine dioxide and its effects on reverse osmosis (RO) membranes by looking at case studies of a seawater desalination facility, and chlorine dioxide’s effectiveness in cleaning fouled reverse osmosis membranes along with preventing biofouling in RO membranes.
A large seawater desalination facility (DESAL) was built to produce about 25 million gallons of drinking water per day from approximately 44 million gallons of seawater (Figure 1).
A number of interrelated problems plagued the facility. These areas of concern are described below:
Naturally occurring organics in seawater such as humic and fulvic acids react with chlorine to produce trihalomethanes, which include chloroform, chloro-dibromo methane, dichloro-bromo methane, and bromoform. The national limit for total trihalomethanes (TTHMs) in potable water was 80 μg/L (ppb). The brominated forms are thought to have a greater health risk than their chlorinated counterparts. If bromide ion is present, chlorine will oxidize the bromine to hypobromite ion, with the result that by far the greater part of the TTHMs will be the more highly brominated forms. Given the need for oxidant to control micro and macro biological growth in the influent line, TTHMs has been unacceptably high.
RO Concentrate Toxicity
Disposal of the RO concentrate was to be to the effluent of the power plant. The RO concentrate was diluted by at most a 20:1 ratio (sometimes up to 70:1) with power plant effluent.
Macrofouling of First Stage Filters
A variety of organisms, and most specifically the Asian green mussel, grew in the influent line. This non-native invasive mollusk caused substantial plugging problems in the first stage filters and other water lines. The problems were exacerbated by the lack of filtration of influent water to DESAL, particularly during the warmer months, when cooler seawater from the power plant influent reservoir was blended with the warmer power plant effluent to provide a water temperature suitable for RO membranes. The pump used to provide the cooler water lay outside any strainers that power plant uses. As a result, unfiltered seawater laden with veligers of these mussels can entered the DESAL influent water pipe, attach to walls of piping, and grow to a size where, once they detach from pipe walls, they may be of sufficient size and density that they cannot be removed from the first stage filters by the normal self-cleaning operation of the filters. Cleaning of these filters must be accomplished manually, resulting in unplanned-for increases in maintenance costs.
Biofouling of Dynasand® Filters
Dynasand® is a patented, two-stage, continuously self-cleaning sand filter. Biofilm growth resulted in agglomeration of sand particles and consequently the performance of the filter was compromised. Chlorine was unable to remove biofilm which caused clumps of sand to form which was not cleaned by the self-cleaning process.
Biofouling of 5 μ Cartridge Filters
The 5 μ cartridge filters protected the RO membranes. Replacement of these filters had been a major challenge. Any material that did not get caught on the Dynasand® filters would foul the 5 μ cartridge filters. The differential pressure would then increase, requiring frequent replacement of these filters. This had been a major expense historically in terms of actual cartridge filter costs as well as time and effort required to replace these filters. Filters, replaced based on differential pressure readings, were costing the plant in excess of $1 MM, on an annualized basis.
Chlorine had been used as the primary micro- and macro- fouling control chemical, and although chlorine was an adequate disinfectant, its use may have caused more problems than it solved. Specifically, use of chlorine had resulted in extremely high TTHMs, it use had not prevented macrofouling of the first stage filters, it could not keep the Dynasand® filters from biofouling, and its use required very frequent replacement of the 5 μ cartridge filters. Dechlorination had also been required to ensure that no free chlorine contacted the RO membranes. The addition of another chemical to remove chlorine added to the complexity and cost of the overall operation.
Chlorine dioxide was proposed as a means of addressing some of these problems. However, the use of chlorine dioxide introduced additional concerns, specifically, the introduction of the chlorite ion, which might prove toxic to aquatic life in the RO concentrate disposal and the potential adverse impact chlorine dioxide might have on the RO membrane. Therefore, a trial was proposed to investigate the use of chlorine dioxide, and to compare the expected benefits with the possible drawbacks. The chlorine dioxide would be applied on the pilot plant unit, so that a short term, intensive investigation could be done, and these things could be investigated.
Chlorine Dioxide Field Trial
Due to the small amount of chlorine dioxide (< ~ 2—3 lb/day) required for testing on the pilot unit, an electrolytic generator which could be operated to produce chlorine-free chlorine dioxide was selected. The protocol was to apply chlorine dioxide at a dosage which would clean up any biological growth in the sand filters, so that a measurable residual could be achieved on the effluent of the sand filter. Then, the generator output would be adjusted so that only a trace of chlorine dioxide was found in the sand filter effluent, to minimize the chlorine dioxide residual which contacted downstream equipment. During the field trial, a number of measurements were made both manually and continuously. Parameters tested manually included chlorine dioxide and its primary disinfection by-product, chlorite ion at every point. Parameters measured continuously included turbidities on the effluent of the Dynasand® filters, differential pressures on the cartridge filter and RO unit, and RO permeate conductivity.
In general, the pilot plant was arranged to simulate as nearly as possible, the residence times, relative flows, and feed rates that would be required by the plant (Figure 2). Sample points and descriptions are given in Table 1.
The unit was installed and started initially to insure it was operating properly. Then the unit was shut down until the Dynasand® filters could be placed into service. Once the Dynasand® filters were placed into service the generator was set to provide a fairly high level of chlorine dioxide to clean up the filters prior to the trial. As the filters cleaned up, the residual out of the filters increased. At that point, the generator production rate was then reduced to provide a low residual out of the Dynasand® filters.
Once the RO was put into service, a complete data set was taken twice per day during the test period, although some parameters such as chlorine dioxide residuals were taken more frequently.
There were several goals in this trial. The goals, in no particular order of importance, were as follows:
- To reduce TTHMs
- To minimize biofouling of the sand filters
- To minimize or eliminate macrofouling of the primary stage filters
- To extend the life of the cartridge filters
- To provide a non-toxic effluent, as measured by third party toxicity testing
- To provide water that has no adverse impact on the RO membranes
In pre-trial work, a chlorine dioxide demand study was done to identify the dosage required. The demand varied between about 0.4 and about 2 mg/L.
TTHMs. TTHMs were measured in grab samples of the seawater treated with chlorine and chlorine dioxide to be 845 and 1 μg/L TTHM, respectively. This kind of spread is somewhat unusual, although spreads of this magnitude have been observed by others.
Biofouling of Sand Filters. A plot of the turbidity along with the chlorine dioxide residual out of the sand filters demonstrated that toward the end of the field trial, the chlorine dioxide residual increased, although the applied dosage did not. This suggested that the demand for chlorine dioxide in the filter had been reduced, i.e., the biofilm/gumbo was being removed.
This was supported by observations of the first stage reject water by the pilot plant operator. revealed some interesting anecdotal information. The highly experienced pilot plant operator noted that the first pass reject water out of the pilot unit was dirtier than he had ever seen it before. It appeared that the self-cleaning process of the filters was enhanced by the chlorine dioxide.
Macrofouling of Sand Filters. While the field trial was not designed to demonstrate the efficacy of chlorine dioxide towards Asian green mussels, the efficacy of chlorine dioxide towards Asian clams, a benthic organism, the zebra mussel, and other mollusks is well established.
A periodic, i.e., once per week or so treatment of the influent line for 2 hours at ~ 1.0 mg/L above demand was expected to encourage the relocation of any mollusks which may have attached and prevent any organisms from achieving a size which will prevent them from being backwashed during the routine operation of the Dynasand® filters. The approach of using low levels of chlorine dioxide to prevent settlement of mollusks has been utilized in a number of applications since the mid 1990s and is described elsewhere.
Extending the Life of 5 μ Cartridge Filters. The chlorine dioxide and chlorite residuals were measured after the 5 μ cartridge filters but before the RO unit. The residuals were generally < ~ 0.3 mg/L.
Figure 4 shows the differential pressure of the 5 μ cartridge filters along with the chlorine dioxide residual out of the cartridge filter. The differential pressure showed a slow steady trend upwards, indicating that the filter was slowly fouling. This may be biofilm/gumbo being removed from the sand filters. Then differential pressure then reached what appeared to be a plateau. Note that the differential pressure of the cartridge filter never reached a point which required replacement of the filter over the ~ 2-week trial period. At the same time, filter replacements in the facility were being required every ~ 3 days. This indicates that use of chlorine dioxide can significantly extend the life of the cartridge filter, which in this case was a major unplanned-for expense (approaching $1MM annually).
Toxicity Testing of the Reject Water. The undiluted RO concentrate had a chlorite residual of about 3.5—5 mg/L. The RO concentrate was discharged to the power plant effluent basin where it was diluted from 20:1 to 70:1 prior to discharge. Therefore a 20:1 dilution was allowed for toxicity testing. The diluted effluent was not found to be toxic to the organisms tested (Table 2).
Impact of Chlorine Dioxide on RO Membranes. Testing of the RO permeate was done for chlorine dioxide and chlorite ion. Results are shown in Figure 5.
Chlorine dioxide concentrations were measured in the permeate that were on the same order as those found in the concentrate. This indicates that the chlorine dioxide readily penetrated the RO membrane.
With a single exception, the chlorite concentrations were very low. Even then, the concentration was well below the potable water limit for chlorite ion.
Figure 6 shows the conductivity of the permeate. The conductivity increased slightly during the first few days of testing, and then a plateau was reached.
Figure 7 shows the differential pressure of the RO compared with chlorine dioxide residual. The differential pressure did not change significantly over the trial period. In addition, the chlorine dioxide residual which was exposed to the RO membrane varied from about 0.1—0.22 mg/L. Only an extended run of a number of months would show with any degree of confidence whether chlorine-free chlorine dioxide had any long-term impact on the membrane.
Neither plant THM data nor RO membrane autopsy data were made available to us, because of pending litigation between several of the groups involved in the design, construction and operation of the plant.
A longer-term trial would be required to be able to say with any degree of certainty that chlorine-free chlorine dioxide, applied at very low dosages, would have any negative impact on the RO membrane integrity. However, there is an increasing body of evidence from both research and field work that suggests such is the case.
Fairly detailed case histories have been presented, one using chlorine dioxide for cleaning of fouled RO membranes. The second was for low level application to prevent biofouling in RO membranes.1
Cleaning Fouled RO Membranes with Chlorine Dioxide: The principal thing learned from this case history is that badly fouled RO membranes already have tears or holes created by bacteria. The biofilm is known to act as a secondary membrane, and in such systems, when the biofilm is removed, leakage occurs. It is better to prevent biofilm formation on the membrane than to try and cleanup a badly fouled membrane, although this can be done if the membrane is not too badly fouled.
Biofilm Prevention in RO Membranes with Chlorine Dioxide: Although chlorine dioxide can be used to clean up moderately fouled RO membranes, it can be used in a much better, less costly, and more effective preventative tool as a routine part of the biofouling prevention program.
 Aieta, E. and Berg, J., “A Review of Chlorine Dioxide in Drinking Water Treatment,” Journal of the American Water Works Association, 78(6), 62(1986).
 Cooper, W., Zika, R., and Steinhauer, M., “Bromide-Oxidant Interactions and THM Formation: A Literature Review,” Journal of the American Water Works Association, Management and Operations, 116(April 1985).
Robeck, G., “Chlorine, Is There a Better Alternative?,” Science of the Total Environment, 18, 235(1981).
 Suozzo, J. and Suozzo, K., “Wastewater Management System,” U.S. Patent 5,843,308, December 1, 1998.
 Pureline Treatment Systems, Bensenville, IL.
 Simpson, G., “Analytical,” Chapter 8, Practical Chlorine Dioxide: Volume I – Foundations,” Greg D. Simpson & Associates, Publishers, 2005.
 Lykins, B. and Griese, M., “Using Chlorine Dioxide for Trihalomethane Control,” Journal of the American Water Works Association, 78(6), 88(June, 1986).
 Cameron, G., Symons, J., Spencer, S., and Ma, J., “Comparison of Various Disinfectants for Control of the Asiatic Clam (Corbicula fluminea),” Proceedings, American Water Works Association Annual Conference, 1987.
Cameron, G., Symons, J., Spencer, S., and Ma, J., “Minimizing THM Formation During Control of the Asiatic Clam: A Comparison of Biocides,” Journal of the American Water Works Association, 81(10), 53(1989).
Cameron, G., Symons, J., Bushek, D., and Kilkarni, R., “Effect of Temperature and pH on the Toxicity of Monochloramine to the Asiatic Clam,” Journal of the American Water Works Association, 63(October 1989).
Goss, L., Jackson, J., Flora, H., Isom, B., Gooch, C., Murray, S. Burton, C. and Bain, W., “Control Studies on Corbicula for Steam-Electric Generating Plants,” Proceedings of the First International Corbicula Symposium, 139(1977).
 Pitochelli, T. and Mason, J., “Large-Scale Field Applications of Chlorine Dioxide,” Intertech Oxidative Treatment of Pollutants in Wastewater Conferences, Philadelphia, Pa., March 21 – 24, 1994.
Rybarik, D., Byron, J., and Germer, M., “Chlorine Dioxide Treatment for Zebra Mussel Control,” Proceedings Fifth International Zebra Mussel and Other Aquatic Nuisance Organisms Conference, Toronto, Canada, February 1995.
 Kemp, P., and Okimoto, N., “The Use of Subtoxic Levels of Chlorine Dioxide to Prevent Surface Condenser Fouling in Once-Through Seawater Cooling Exchangers,” Proceedings 43rd International Water Conference, IWC 82-8, October 25 – 27, 1982.
 Pickrell, W., Germer, M., Miller, J., and Simpson, G., “Plants Increase Production Through Use of Chlorine Dioxide,” Practical Industrial Water Treatment Technology for the New Millennium, Houston, TX, January 27 – 29, 1999.
 Averett, W., Simpson, G., and Miller, J., “Cleaning RO Membranes with Chlorine Dioxide,” Southwest Chemistry Conference, sponsored by TXU Energy, Dallas, Texas, July 28 – August 1, 2003.