Quantitative Evaluation of the Synergistic Sequential Inactivation of

subtilis spores, to make a quantitative evaluation and to improve the mechanistic .... In general, 5-10 samples were taken every 5-8 min to measure th...
0 downloads 0 Views 121KB Size
Download PDF
Environ. Sci. Technol. 2003, 37, 2134-2138

Quantitative Evaluation of the Synergistic Sequential Inactivation of Bacillus subtilis Spores with Ozone Followed by Chlorine MIN CHO,† HYENMI CHUNG,‡ AND J E Y O N G Y O O N * ,† School of Chemical Engineering, College of Engineering, Seoul National University, San 56-1, Sillim-dong, Gwanak-gu, Seoul, 151-742, Korea, and Water Microbiology Division, National Institute of Environmental Research, Kyungseo-dong, Seo-gu, Inchon, 404-170, Korea

This investigation of sequential disinfection, with ozone followed by free chlorine, was carried out using Bacillus subtilis spores, to make a quantitative evaluation and to improve the mechanistic understanding of their synergistic effect. This study shows that the extent of the synergistic effect in the inactivation of Bacillus subtilis spores appears to be dependent upon the level of preozonation. However, when the ozone pretreatment level exceeded the lag phase of the ozone inactivation curve, the chlorine inactivation curves were almost identical regardless of the level of preozonation. When this sequential disinfection was performed in the reverse order, no enhanced disinfection was observed. This difference, depending on the order of disinfectant application in sequential disinfection, was explained in terms of the enhanced disinfection being the result of the greater intracellular diffusion of free chlorine, caused by the cell surface disruption induced by ozone. The practical implications of this synergistic sequential inactivation with ozone followed by free chlorine were discussed, along with the issue of selecting the amount of each oxidant to use in water treatment plants, to achieve a specific level of microorganism inactivation.

Introduction With the appearance of Cryptosporidium parvum and Giardia lamblia, which were resistant to chlorine, many drinking water systems employing free chlorine as a primary disinfectant reportedly encountered difficulties in achieving the required microbial inactivation levels, without excessive DBPs (disinfection byproducts) formation (1). Among the alternatives to chlorine, ozone is widely used due to its strong biocidal oxidizing property and its ability to diffuse through the cell membrane (2-7). Finch et al. (4) showed that ozone could disinfect 99.5% (2.3 log) of Giardia lamblia within 2 min of contact time and 0.79 mg L-1 of initial ozone residual. Marin ˜as group (5, 7) also found that ozone inactivated E. coli and Cryptosporidium parvum effectively by altering the ability of diffusion through the cell membrane. Practically speaking, ozone needs to be employed with chlorine (or combined with free chlorine) to maintain residual * Corresponding author phone: +82-2-880-8927; fax: +82-2-8768911; e-mail: [email protected]. † Seoul National University. ‡ National Institute of Environmental Research. 2134

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 10, 2003

disinfectant in the distribution system, which is required to prevent microbial contamination and biofilm growth. Recently, significant attention was given to the sequential disinfection of Cryptosporidium parvum with ozone followed by free chlorine, which was able to achieve an enhanced level of disinfection when compared with the use of either disinfectant used alone (7-10). However, until now, no detailed quantitative evaluation or mechanistic explanation of the synergistic effect behind this microbial inactivation has been provided. In certain cases, the conditions affecting synergistic inactivation are controversial: Lewin et al. (8) reported no observation of any synergistic effect at pH 6 in the sequential inactivation of Cryptosporidium using ozone followed by free chlorine, whereas Driedger et al. (7) observed a synergistic effect which increased with decreasing pH. These inconsistent results can be partly attributed to the nature of the detection methods used for Cryptosporidium as well as the complicated nature of aqueous ozone chemistry which make exact kinetic measurements difficult. The inactivation of Cryptosporidium parvum is difficult to measure and gives rise to rather variable results, depending on the type of assays employed, which include such methods as infectivity assays of mouse and cell culture assays as well as the viability analysis of excystation and vital dyes (11, 12). As an alternative approach, the use of indicator microorganisms such as Bacillus spores has been suggested (2, 13). These microorganisms acted as good surrogates in inactivation studies for inert protozoan cysts, especially for the comparison and optimization of reaction conditions. Since a better understanding of the synergistic effect could lead to a reduction in the retention time of the disinfection chamber or in the disinfectant dosage, resulting in less DBPs being formed and reducing chemical costs, a quantitative evaluation, supported by a mechanistic understanding of the synergistic effect, is essential for the rational design and practice of the disinfection process in water treatment plants. This investigation of the process of sequential disinfection with ozone followed by free chlorine, using Bacillus subtilis spores as a surrogate for the protozoan cysts was carried out with two specific objectives. First, a quantitative evaluation of the synergistic effect was made as a function of preozonation. The magnitude of the additional disinfection, achieved as a result of synergistic inactivation, was also evaluated. Second, a mechanistic understanding of the synergistic effect was also pursued. In the process of sequential disinfection, ozone followed by free chlorine was compared with free chlorine followed by ozone, to isolate any specific effects of preozonation on the microorganism.

Experimental Section Two types of sequential disinfection experiments, with ozone followed by free chlorine and with chlorine followed by ozone, were each carried out at two pH levels, 5.6 and 8.2, and at a constant temperature of 20 °C. Each sequential disinfection experimental result was compared with that from the experiment using a single disinfectant, ozone or chlorine, to quantitatively estimate the enhancement resulting from the sequential disinfection. Bacillus subtilis spores (ATCC 6633) were selected for this study, because they are sufficiently resistant to ozone to be able to study the inactivation kinetics within a reasonable time scale, and their inactivation mechanisms of ozone and free chlorine were reported (14). The laboratory procedures used in this study, for the preparation of the microorganisms and for the ozone disinfection, were similar to those used by Cho et al. (2). All water was prepared using deionized/distilled water, treated 10.1021/es026135h CCC: $25.00

 2003 American Chemical Society Published on Web 04/10/2003

with a Barnstead NANO pure system (Barnstead, Iowa, U.S.A.), and analytical reagent grade chemicals were used (Fisher Scientific, Springfield, U.S.A.). The pH was maintained at pH 5.6 or 8.2 using 20 mM phosphate buffer solution. The resulting buffer solution, which did not have any measurable chlorine demand, was preozonated for 60 min to satisfy any ozone demand and left standing for 24 h. All glassware used in these experiments was washed with distilled water and then autoclaved at 121 °C for 15 min. Spore suspensions of Bacillus subtilis (ATCC 6633) were prepared following the procedure described by Nakayama et al. (15), except for some minor modifications such as the use of 1/10 nutrient agar with an extended incubation time of 5-7 days. The stocks were diluted and treated at 80 °C for 20 min just before each experiment. The numbers of viable spores were measured by the spread plate method with nutrient agar grown at 37 °C for 24 h. One milliliter of solution was withdrawn at each sampling times from the experimental reactor and was 10-fold diluted up to 1/1000 or up to 1/100 000, depending on the initial viable number of Bacillus subtilis spores, which ranged approximately from 105 to107 cfu mL-1. Then, 0.1 mL of the undiluted and diluted solutions was inoculated onto the three replicate plates for each dilution, respectively. Selection of the statistically relevant plate counts and calculation was carried out according to the standard method (16). The ozone disinfection experiments were conducted in a piston type batch reactor (50 mL, Pyrex), which did not have free headspace. Several appropriate amounts of concentrated ozone solution (>40 mg L-1) made with an ozone generator (CFS-1, Ozonia Co., Du ¨ bendorf, Switzerland) were mixed immediately with a solution containing microorganisms in the piston type batch reactor so as to attain an approximate initial concentration of 1.50-2.00 mg L-1. Ozone measurement was made using indigo method by means of a UV-vis spectrophotometric detector (Gilson 151 UV/VIS, Middleton, U.S.A.) which is based on the quantitative decolorization of indigo trisulfonate at 600 nm as result of its reaction with ozone (17). To measure the continuous ozone concentration with time, a small fixed quantity of the ozonated sample was taken continuously from a piston type batch reactor, and mixed rapidly with an indigo trisulfonate solution using Flow injection analysis (FIA) system. This is a specially designed instrument to continually and accurately measure the residual ozone concentration as a function of time (2). The detection limit of residual ozone by the flow injection analysis technique was about 0.01 mg L-1. In general, 5-10 samples were taken every 5-8 min to measure the number of the viable microorganisms. Na2S2O3 was used for stopping the inactivation reaction by residual ozone. All materials used in this experimental apparatus were carefully selected to not have any undesirable reactivity with ozone. The chlorine disinfection experiments were performed using a 50 mL batch reactor. A chlorine stock solution (300 mg L-1) was prepared by dilution with sodium hypochlorite solution (5% Junsei Co., Tokyo, Japan). The chlorine dose ranged from 2.00 mg L-1 to 4.00 mg L-1 as Cl2 at pH 8.2 and pH 5.6. Residual free chlorine levels were assayed with DPD (N,N-dimethyl-p-phenylenediamine) reagent at 530 nm (DR/ 2010, HACH Co., Loveland, U.S.A.). In general, 5-10 samples were taken during the disinfection experiments (120-300 min) to measure the microorganisms. The experiments designed to measure the sequential inactivation of Bacillus subtilis spores with ozone/free chlorine and free chlorine/ozone were carried out following the procedures described in the previous studies (7, 18) with several modifications. In the experiment involving the sequential inactivation with ozone/free chlorine, concen-

trated ozone (>40 mg L-1) was introduced into a reactor containing Bacillus subtilis spores and mixed immediately with a solution to attain an initial ozone concentration of about 0.05-1.15 mg L-1. Initial ozone dosage was controlled so as not to have any residual ozone after required contact time. Since then, ozone treated sample solution, containing Bacillus subtilis spores, was transferred into a 50 mL flask reactor for the subsequent disinfection experiment with free chlorine. This step contrasts with the method used in a previous study (18), which employed Na2S2O3 for removing residual ozone, followed by filtering and centrifuging in order to transfer spores only of interest for the secondary disinfection step. In the experiment involving the sequential inactivation with free chlorine/ozone, the residual disinfectant in the free chlorinated sampled solution was immediately quenched with Na2S2O3 after running the planned CT values, and then the spore suspension was centrifuged at 1200×g for 12 min. The supernatant was discharged, and the spores were resuspended in a phosphate buffer (10 mM) and stored at 4 °C for 24 h. The recovery yield of centrifugation was over 99%. These spores treated with free chlorine were used for the subsequent ozone disinfection experiment. All of the disinfection experiments were repeated three times. The averaged value for each experimental condition was used for data analysis, as the three repeated measurements fall within 10% deviation. The Delayed Chick-Watson model with a time average ozone concentration was chosen in this study to explain the kinetics of the Bacillus subtilis spores inactivation, as described by eq 1 (2, 7, 12).

(

0

N Log ) N0 -kC h T - kC h Tlag

1 if C hT e C h Tlag ) Log k 1 if C hT g C h Tlag ) Log k

( ) ( ) N N0 N N0

)

(1)

where N0 initial number of viable Bacillus subtilis spore (cfu mL-1), N is the remaining number of viable Bacillus subtilis spore at time T (cfu mL-1), C is the ozone concentration (mg L-1), C h ) ∫0tC dT/T is the time averaged ozone concentration (mg L-1), and k is the inactivation rate constant (min-1). Although Delayed Chick-Watson model is not true rate model, this model was easy to analyze our results of inactivation kinetics especially with shoulder. The suitability of the Delayed Chick-Watson model for all experimental measurements of sequential disinfection was supported by finding high correlation coefficients (all R2 > 0.96), as shown in the previous study (2). The value of C h T was calculated by the integration of residual ozone measured by FIA. Neither the levels of initial microorganisms nor the ozone concentration had any significant impact on the inactivation curves under the experimental conditions used in this study.

Results and Discussion Inactivation of Bacillus subtilis Spores with a Single Disinfectant; Ozone and Free Chlorine. The results of Bacillus subtilis spores inactivation with a single disinfectant are presented in Figure 1 (a) for ozone and (b) for free chlorine. In the case of ozone disinfection, the C h T values for the log inactivation of Bacillus subtilis spores at pH 8.2 are smaller than those at pH 5.6, mainly because of the shortened lag phase of the inactivation curve at higher pH. When ozone was applied at a rate of less than 1.10 mg min L-1 at pH 5.6 and 0.63 mg min L-1 at pH 8.2, less than the respective maximum C h T values of the lag phase, respectively, no apparent decrease of microorganisms was observed. The reason for the smaller C h T for inactivation of Bacillus subtilis VOL. 37, NO. 10, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2135

FIGURE 1. Inactivation of Bacillus subtilis spores by a single disinfectant (a) ozone and (b) free chlorine. spores at pH 8.2 was explained by the role of •OH in previous studies (2, 19-21), where exact ozone exposure was controlled with t-BuOH, the well-known •OH scavenger (2, 21). However, in free chlorine disinfection, the C h T values for log inactivation of Bacillus subtilis spores are much smaller at low pH than those at high pH, because HOCl (pKa ) 7.4) is a more effective germicide than OCl- (7, 22). The maximum C h T values of the lag phase of the inactivation curves for inactivating Bacillus subtilis spores with chlorine were 41 mg min L-1 at pH 5.6 and 150 mg min L-1 at pH 8.2. The C hT value for 3 log inactivation at pH 5.6 was approximately one fourth of that at pH 8.2 (127 mg min L-1 versus 500 mg min L-1). Synergistic Sequential Disinfection with Ozone Followed by Free Chlorine. The results of the sequential inactivation of the Bacillus subtilis spores with ozone followed by free chlorine at pH 8.2 and 5.6 are presented in Figures 2 and 3, respectively. Ozone application to Bacillus subtilis spores as a pretreatment ranged from zero up to 4.41 mg min L-1 as quantified by the C h T values. Eleven and eight cases of ozone pretreatment were performed at pH 8.2 and 5.6, respectively. The results of Bacillus subtilis spore inactivation, where the preozonation C h T was less than C h Tlag, are presented in Figures 2(a) and 3(a). The results of Bacillus subtilis spore inactivation, where preozonation exceeded the C h T values of the lag phase of the inactivation curves, are presented in Figures 2(b) and 3(b). The difference of vertical axis in Figures 2 and 3 represents the additional inactivation due to preozonation. If no synergistic effect exists, all of the inactivation curves should have the same lag phase and slope. Figures 2 and 3 give rise to two important observations. First, the synergistic effect of ozone pretreatment was 2136

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 10, 2003

FIGURE 2. Synergistic inactivation of Bacillus subtilis spores with free chlorine at pH 8.2 after pretreatment with different C h T levels of ozone. (a) Ozone pretreatment with C h T values within the lag phase of ozone curve, (b) Ozone pretreatment with C h T values over the lag phase of ozone curve ([O3]0 ) 0.05-1.15 mg L-1, [HOCl]0 ) 2-4 mg L-1). apparent at both pHs, as previously reported in the literature (8, 12). For example, when the C h T values for 2 log reduction were compared between the cases of free chlorine disinfection only and that of free chlorine disinfection with ozone pretreatment, levels of enhancement of up to 45% and 42% were observed at pH 8.2 and pH 5.6, respectively. Second, the extent of the synergistic effect for the inactivation of Bacillus subtilis spores appears to be dependent upon the level of preozonation, the effect of which was mainly reflected in the change of C h Tlag (the intercept of the X-axis). When the preozonation C h T was less than C h Tlag (up to 0.63 mg min L-1 at pH 8.2 and 1.10 mg min L-1 at pH 5.6), the synergistic effect increased according to the level of preozonation. Sequential Disinfection with Free Chlorine Followed by Ozone. The sequential disinfection with free chlorine followed by ozone, which represents the reverse order in comparison to the previous sequential disinfection experiments, was carried out at pH 8.2 with Bacillus subtilis spores. Eight different levels of prechlorinations were carried out, from 0 to 500 mg min L-1 as quantified by the C h T values. Figure 4 indicates that the behaviors of the log reduction of Bacillus subtilis spores by ozone are quite similar regardless of the magnitude of prechlorination, within the range of C h T values of 0-500 mg min L-1. This is in stark contrast with the results shown in Figures 2 and 3, which show significant synergistic effects in the sequential disinfection with ozone followed by free chlorine.

FIGURE 5. The quantification of the synergistic effect as a function of the extent of preozonation to achieve 2 log inactivation of Bacillus subtilis spore with ozone followed by free chlorine.

FIGURE 3. Synergistic inactivation of Bacillus subtilis spores with free chlorine at pH 5.6 after pretreatment with different C h T levels of ozone. (a) Ozone pretreatment with C h T values within the lag phase of ozone curve. (b) Ozone pretreatment with C h T values over the lag phase of ozone curve ([O3]0 ) 0.05-0.65 mg L-1, [HOCl]0 ) 2 mg L-1).

FIGURE 4. Sequential inactivation of Bacillus subtilis spores with ozone at pH 8.2 after pretreatment with different levels of free chlorine. Quantitative Estimation of the Sequential Disinfections. The quantitative effect of preozonation on the sequential disinfection with ozone followed by free chlorine at two pHs was reconstructed in Figure 5, based on the 2 log inactivation of Bacillus subtilis spores with free chlorine. Figure 5 further emphasizes, in a more quantitative manner, the observation

previously made with regard to Figures 2 and 3. For example, Figure 5 shows that ozone pretreatment with 1.53 mg min L-1 at pH 8.2 in the sequential inactivation with ozone/free chlorine, can achieve an additional 1.65 log inactivation. 2 log inactivation of Bacillus subtilis spore requires 208 mg min L-1 with free chlorine (Figure 2(b)). If no synergistic effect existed for this sequential disinfection, 208 mg min L-1 of free chlorine would only achieve 0.35 log inactivation of Bacillus subtilis spores (Figure 1(b)). Therefore, as shown in Figure 5, an additional 1.65 log inactivation of Bacillus subtilis spores was achieved due to the synergistic effect of sequential disinfection with ozone followed by free chlorine. For other levels of preozonation, the enhanced disinfection resulting from the sequential inactivation is presented in Figure 5. The extent of the synergistic effect for inactivation of Bacillus subtilis spores increases linearly with the level of preozonation. However, when the ozone pretreatment level exceeds the C hT values of the lag phase, the chlorine inactivation curves are almost the same regardless of the level of preozonation. This implies that an optimum level of ozone pretreatment exists in the sequential disinfection with ozone followed by free chlorine. On the other hand, when the same sequential disinfection was performed in the reverse order (free chlorine first and ozone), no synergistic effect of sequential inactivation was observed, as shown in Figure 4. This difference in synergistic effect, depending on the order of disinfectant application in sequential disinfection, can be inferred from the different roles of each oxidant in the inactivation of Bacillus subtilis spores. The rigid structure of the Bacillus subtilis spores might present a substantial barrier against disinfectants such as free chlorine. However, ozone is reported to attack, oxidize, and disrupt the surface components of the cell wall, membrane, and spore coat (3, 14, 23-26) and to be partially associated with DNA damage and diffusion into the inner components of the cell (14, 22, 26, 27). Recently, Khadre et al. (14) observed the damage to the surface layer, outer spore coat, inner spore coat, and cortex in ozone treated Bacillus subtilis spores using transmission electron microscopy. In addition, the presence of •OH, induced by ozone, plays a significant role in microbial inactivation (2), largely due to the destruction of the cell membrane or wall (27), as in the case of ozone itself. On the other hand, free chlorine is not normally able to damage the outer components of the cell as much as ozone. Free chlorine is able to inactivate microorganisms by diffusing into the cell’s inner components: successfully diffused free chlorine may inactivate enzymes and damage intracellular components such as nucleic acids (10, 24, 26). VOL. 37, NO. 10, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2137

subtilis spores appears to be dependent upon the level of preozonation. However, when the ozone pretreatment level exceeds the lag phase of the ozone inactivation curve, the inactivation curves resulting from the subsequent addition of chlorine are almost identical, regardless of the level of preozonation. This indicates that an optimum level of preozonation exists in the case of sequential disinfection with ozone followed by free chlorine.

Acknowledgments This research was supported in part by the Brain Korea 21 Program (of the Ministry of Education). This support is greatly appreciated.

Literature Cited

FIGURE 6. Practical implications of the effect of sequential disinfection with ozone followed by free chlorine (pH 8.2). These differences indicate that the synergistic disinfection observed in Figures 2 and 3 may be the result of the enhanced intracellular diffusion of free chlorine, caused by the cell surface disruption induced by the primary ozone (10, 24). Furthermore, the magnitude of the synergistic effect in the sequential disinfection with ozone followed by free chlorine was proportional to the level of preozonation, up to the lag phase of the inaction curve by ozone as shown in Figure 5. It is important to note that the disruption of the cell surface, which occurred in the lag phase of the ozone inactivation of Bacillus subtilis spores, may result in cell damage without loss of viability. The presence of damaged but viable Bacillus subtilis spores contribute to the synergistic effects. Over the level of the lag phase of ozone inactivation curve, ozone reactive sites, which were related to synergistic effect, have been already consumed, and so there is no further increase in the permeability. The points of maximum synergistic effect of ozone pretreatment are 0.63 mg min L-1 at pH 8.2 and 1.1 mg min L-1 at pH 5.6. Figure 6 describes the practical implications of the effect of sequential disinfection with ozone followed by free chlorine (pH 8.2). Although Bacillus subtilis spores is not a regulated pathogen, such as Cryptosporidium parvum, they have been widely used as an indicator microorganism (2, 13), and this microorganism acts as a good surrogate in inactivation studies for inert protozoan cysts, especially for the comparison and optimization of reaction conditions. So the practical relevance was attempted. Figure 6 was reconstructed from the results of Figures 2 and 3. It shows the various options available for selecting the amount of each oxidant, in the process of sequential disinfection with ozone followed by free chlorine, in water treatment plants which need to accomplish 2 or 3 log inactivation, as compared with those where this synergistic effect is not present. For example, Figure 6 indicates that 1 mg min L-1 of preozonation and 100 mg min L-1 of free chlorination in secondary disinfection or 0.5 mg min L-1 of preozonation and 200 mg min L-1 of free chlorination are required to achieve the 3 log inactivation of Bacillus subtilis spores. In conclusion, this study reveals for the first time that the extent of the synergistic effect for the inactivation of Bacillus

2138

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 10, 2003

(1) Joret, J. C.; Mennecart, V.; Robert, C.; Compagnon, B.; Cervantes, P. Water Sci. Technol. 1997, 35, 81-86. (2) Cho, M.; Chung, H.; Yoon, J. Ozone Sci. Eng. 2002, 24, 145-150. (3) Facile N.; Barbeau, B.; Prevost, M.; Koudjonou, B. Water Res. 2000, 34, 3238-3246. (4) Finch. G. R.; Black, E. K.; Labatiuk, C. W.; Gyurek, L.; Belosevic, M. Appl. Environ. Microbiol. 1993, 59, 3674-3680. (5) Hunt, N. K.; Marin ˜ as, B. J. Water Res. 1999, 33, 2633-2641. (6) Wickramanayake, G. B.; Rubin, A. J.; Sproul, O. J. J. Water Pollut. Control Fed. 1984, 56, 983-988. (7) Driedger A. M.; Rennecker, J. L.; Marin ˜ as, B. J. Water Res. 2000, 34, 3591-3597. (8) Lewin, N.; Craik, S.; Li, H.; Smith, D. W.; Belosevic, M. Ozone Sci. Eng. 2001, 23, 411-420. (9) Liyanage, L. R. J.; Finch, G. R.; Belosevic, M. Ozone Sci. Eng. 1997, 19, 49-423. (10) Montgomery, J. M. Water Treatment Principles and Design; John Wiley & Sons: 1985. (11) Bukhari, Z.; Marshall, M. M.; Korich, D. G.; Fricker, C. R.; Smith, H. V.; Rosen, J.; Clancy, J. L. Appl. Environ. Microbiol. 2000, 66, 2972-2980. (12) Rennecker, J. L.; Driedger, A. M.; Rubin, S. A.; Marin ˜ as, B. J. Water Res. 2000, 34, 4121-4130. (13) Driedger A. M.; Staub, E.; Pinkernell, U.; Marin ˜ as, B. J.; Koster, W.; von Gunten, U. Water Res. 2001, 35, 2950-2960. (14) Khadre, M. A.; Yousef, A. E. Food Sci. Technol. 2001, 71, 131138. (15) Nakayama, A. Y.; Yano, Y.; Kobayashi, S.; Ishikawa, M.; Sakai, K. Appl. Environ. Microbiol. 1996, 62, 3897-3900. (16) APHA, AWWA, WPCF. Standard Methods for the Examination of Water and Wastewater, 20th ed.; Washington, DC, 1998. (17) Bader, H.; Hoigne´, J. Water Res. 1981, 15, 449-456. (18) Driedger A. M.; Rennecker, J. L.; Marin ˜ as, B. J. Water Res. 2001, 35, 41-48. (19) Hoigne´, J.; Bader, H. Ozone Sci. Eng. 1979, 1, 73-85. (20) Wolfe, R. L.; Stewart M. H.; Scott K. N.; McGuire M. J. Environ. Sci. Technol. 1989, 23, 744-745. (21) Cho, M.; Chung, H.; Yoon, J. Appl. Environ. Microbiol. 2003, 69, in press. (22) Sobsey, M. D.; Fuji, T.; Shields, P. A. Water Sci. Technol. 1988, 20, 385-391. (23) Fisher, C. W.; Lee, D.; Dodge, B.; Hamman, K. M.; Robbins, J. B.; Martin, S. E. Appl. Environ. Microbiol. 2000, 66, 1405-1409. (24) Hass, C. N.; Engelbrecht, R. S. J. Water Pollut. Control. Fed. 1980, 52, 1976-1989. (25) Langlais, B.; Perrine, D. Techniques Sciences MeAˆ thodes 1989, 84, 215-219. (26) Watts, R. J.; Kong, S.; Orr, M. P.; Miller, G. C.; Henry, B. E. Water Res. 1995, 29, 95-100. (27) von Sonntag. Water Supply 1986, 4, 11-18.

Received for review September 9, 2002. Revised manuscript received March 3, 2003. Accepted March 13, 2003. ES026135H