- Email: [email protected]
Toxicity of Ozone to Fish Larvae andDaphnia magna
ECOTOXICOLOGY AND ENVIRONMENTAL SAFETY
41, 176—179 (1998)
ENVIRONMENTAL RESEARCH, SECTION B ARTICLE NO.
ES981696
Toxicity of Ozone to Fish Larvae and Daphnia magna M. Leynen,* L. Duvivier,- P. Girboux,‡ and F. Ollevier* *Laboratory of Ecology and Aquaculture, Katholieke Universiteit Leuven, Naamsestraat 59, B-3000 Leuven, Belgium; -Laborelec, Rhodestraat 125, B-1630 Linkebeek, Belgium; and ‡Electrabel, Regentlaan 8, B-1000 Brussels, Belgium Received June 16, 1997
Ozone can be used as an alternative to chlorination to control biofouling in cooling water systems. The possible negative environmental impact of a discharge of ozone-containing cooling water was investigated. The acute toxicity of dissolved ozone was determined for fish larvae of three species [Cyprinus carpio (at 27°C), Leuciscus idus (at 27°C) and Clarias gariepinus (at 32°C)] and to Daphnia magna (at 21 and 27°C). The results indicate that ozone is very harmful to aquatic life. Daphnids are more sensitive to ozone than fish larvae. The mean 48-h LC50 value for the larvae amounts to about 35 lg/liter, while the 48-h NOEC for D. magna was 11 lg/liter (at 21°C). It was concluded that, to protect aquatic life, discharged cooling water should not contain any dissolved ozone. This can be achieved in practice by mixing the treated cooling water with a source of organic substances before discharge, as free ozone will react immediately with organic matter and thus disappear. ( 1998 Academic Press
INTRODUCTION
Chlorine has been used for decades to control microand macrofouling in cooling water circuits of power plants. However, it was demonstrated years ago that chlorine-treated cooling water discharges are toxic due to small amounts of free chlorine or of primary reaction products such as chloramines (EPRI, 1982). Chlorine is also responsible for the formation of toxic organohalogens (chloroform, haloacetic acids, etc.) (Bellar et al., 1974; Rook, 1974). In search of possible alternatives to chlorination, several studies have evaluated the effectiveness of ozonation in controlling biofouling (Sugam et al., 1981; Claudi and Mackie, 1994; Duvivier et al., 1996). Ozone is a stronger oxidant than chlorine and is able to remove existing biofilms (Kaur et al., 1992) and to kill veligers of zebra mussels (Dreissena polymorpha) at low concentrations (Van Benschoten et al., 1993). It is a powerful biocide and will oxidize many organic and inorganic substances. The use of ozone in cooling water treatment also allows reduction of existing COD and AOX levels (Leitzke et al., 1994), which is an ecological advantage compared with chlorine.
Ozone is a very unstable molecule and, after injection into raw water, it decomposes very rapidly. In the presence of organic impurities, the half-life of ozone is reduced to minutes (Duvivier et al., 1996). The rate of decomposition also increases with increasing pH. At low pH ((7), molecular ozone (O ) is the dominant species. As pH increases, 3 O turns into the very short-lived (microseconds) hydroxyl 3 radicals (OH°) (Rice and Wilkes, 1992). The high reactivity makes it difficult to maintain sufficient high ozone concentrations throughout the cooling water system to kill all fouling micro- and macroorganisms. On the other hand, the rapid dissipation of ozone normally ensures that there will be a small environmental impact at the discharge zone of the power plant. The present study investigates the possible adverse impact of discharged dissolved ozone in terms of acute toxicity to fish larvae (Cyprinus carpio, ¸euciscus idus, and Clarias gariepinus) and Daphnia magna. The acute toxicity curves (with corresponding LC values) were deter50 mined in laboratory tests.
MATERIAL AND METHODS
Experimental Setup The toxicity tests in the laboratory were carried out by means of the setup in Fig. 1, which is a modified version of systems that have been applied succesfully by other investigators (Wedemeyer et al., 1979; Asbury and Coler, 1980). Ozone is generated from pure oxygen with two aquarium ozonizators (Sander GmbH, Germany). They have a capacity of 200 mg ozone/h each. The oxygen flow through the ozonizators was held constant at 1.5 liters/min. A PVC column with a height of 2 m and a diameter of 20 cm serves as a contact chamber between water and gaseous ozone. Ozone is injected at the bottom of the column while the water flows into the column at the top (countercurrent principle). After 3 to 4 h, a constant ozone gradient is created, with the highest ozone concentration at the bottom. The contact column contains five branches. Each branch samples a different ozone concentration (according to the gradient) and leads to a set of three aquaria (2-liter volume),
176 0147-6513/98 $25.00 Copyright ( 1998 by Academic Press All rights of reproduction in any form reserved.
OZONE TOXICITY TO FISH LARVAE AND D. magna
FIG. 1.
177
Experimental setup.
in which the toxicity tests take place. The ‘‘ozone-enriched’’ water flows with a constant velocity through the test aquaria. After passing the aquaria, the test water is recovered in tank A and pumped to tank B. From tank B, the water flows by gravity back into the column. The whole setup is in fact a flow-through system with recovery of the test water. Due to the rapid dissipation of ozone, the concentration in tank B was reduced to undetectable levels. A separate setup without ozonization was used as a reference. The test conditions (temperature and water velocity) in the control experiment were the same as in the toxicity test setup. The water used in the tests was dechlorinated tap water. After the ozone gradient was created, the pH measured 8.5$0.1 during the tests. Ozone Measurement At regular time intervals, the ozone concentration was measured with a spectrophotometer by means of the indigo method (Bader and Hoigne´, 1981). Ozone decolors indigo trisulfonate under acid conditions. This decoloration can be measured at 600 nm. The detection limit is about 5 lg/liter
ozone. Compared with other methods (e.g., DPD method), the indigo method excludes all interference from other products (Gilbert and Hoigne´, 1983). The ozone concentration is calculated with the formula (APHA, 1989) mg ozone/liter"100](A!A@)/f]b]» where A!A@"difference in absorbance between sample and blank, b"path length of the cell (in cm), »"volume of the sample (in ml), and f"0.42 (correction factor). Organisms Fish larvae were collected from a hatchery (carp and ide) or obtained from this laboratory culture (African catfish). The fish were fed twice daily with commercial larval food until 2 days before the test. They were acclimated to the test water and the test temperature for at least 2 weeks in 30-liter aquaria. Mean total length ($SD) and mean total wet weight ($SD) of the test animals are given in Table 1.
178
LEYNEN ET AL.
TABLE 1 Length and Weight of the Fish Larvae
TABLE 2 48-h LC50 Values (lg/liter) for Fish Larvae Exposed to Ozonea
Mean total length (cm)
Mean total wet weight (mg)
Binomial method
1.1$0.1 1.7$0.2 1.1$0.1
17$3 50$8 11$2
Carp Cyprinus carpio ¸euciscus idus Clarias gariepinus
Ide African Catfish
The stock of D. magna used in the tests was procured from the University of Ghent, Laboratory of Animal Ecology, and has been maintained in this laboratory since February 1993. Daphnids were held at 21°C with a light/darkness period of 16/8 h. They were fed once daily with unicellular green algae (Scenedesmus acutus and Selenastrum capricornutum). Toxicity Tests Using the five branched outlets from the column with ozone-enriched water, two replicates of 10 fish larvae were exposed to each of five test concentrations of ozone. The test duration was 48 h and the temperature was held constant at 27°C for carp and ide and at 32°C for African catfish. Dead animals were removed regularly. The experiment was carried out twice with carp and ide. The fish larvae were not fed during the test. Neonates (less then 24 h old) of D. magna were exposed to five concentrations of ozone during 48 h (two replicates of 10 daphnids each). The neonates used in the test came from the second or a later brood of females kept in stock culture. The test was carried out at two different temperatures, 21 and 27°C. Test animals and stock culture were not acclimated at 27°C. Daphnids were not fed during the test. Where possible, LC values with 95% confidence limits 50 were calculated with the binomial method and by means of probit analysis (Stephan, 1977). Given an experimental design, it should be noted that the 95% confidence limits refer to the within-experiment repeatability rather than the interexperimental error. The latter can be estimated from the comparison of average values obtained in different experiments. RESULTS
Fish Larvae Forty-eight-hour LC values for the fish larvae are pro50 vided in Table 2. No mortality occurred in the control groups. Calculations used measured levels of dissolved ozone in the test aquaria. Two different calculation methods were reported. Overall, the 48-h LC values are very sim50 ilar for the different fish species, ranging between 30 and 45 lg/liter.
31 44 36 39 35
(26—36) (41—61) (27—48) (33—46) (21—64)
Probit analysis 31 (28—33) 44 (43—45) n.d. n.d. 37 (29—46)
a Ninety-five percent confidence intervals (in parentheses) were calculated separately for each experiment. n.d., could not be determined.
Daphnids Tables 3 and 4 indicate the mortality of D. magna neonates in the presence of different ozone concentrations at test temperatures of 21 and 27°C, respectively. At the highest concentration tested, 140 lg/liter, all daphnids were killed within 1 h of exposure. After 24 h, 100% mortality was obtained with 21 lg/liter, whereas 11 lg/liter had no effect on mortality during 48 h of exposure. Because it was impossible with the experimental setup to create a gradient between 11 and 21 lg/liter the 48-h NOEC was determined instead of the 48-h LC . At 21°C, the 48-h NOEC for 50 ozone for D. magna was 11 lg/liter. Very similar results were obtained at 27°C (Table 4) and the resulting 48-h NOEC was 16 lg/liter. DISCUSSION
When ozone is used as an alternative to chlorination in power plant cooling systems, it should be kept in mind that dissolved ozone is very toxic to aquatic organisms. Fortyeight-hour LC and NOEC values are indeed low 50 (10—45 lg ozone/liter), 0.3 mg/liter ozone kills catfish larvae in 1 h (Leynen, personal observation) and 0.14 mg/liter over 1 h is enough to kill neonates of D. magna. According to Wedemeyer et al. (1979), death of fish following acute
TABLE 3 Mortality Data of Daphnia magna Neonates Exposed to Ozone at 21°C O concentration 3 (kg/liter) Control 11$3a 21$5 57$16 93$25 140$19 a $SD.
Exposure period (h)
Mortality (%)
48 48 24 4 2 1
0 0 100 100 100 100
OZONE TOXICITY TO FISH LARVAE AND D. magna
TABLE 4 Mortality Data of Daphnia magna for Ozone at 27°C O concentration 3 (lg/liter) Control 16$5a 23$5 42$10 97$17 137$21
Exposure period (h)
Mortality (%)
48 48 24 4 2 1
0 0 100 100 100 100
179
temperatures because the temperature in the discharge zone of the cooling water is higher than the usual standard test temperature. REFERENCES
a $SD.
exposure to ozone is most likely due to severe gill lamellar ephithelial tissue destruction accompanied by massive hydromineral imbalance. From the results, it is clear that the ozone concentration in the water at the discharge site should be diluted to 0.015 lg/liter or less. This initial PNECwater value (predicted no-effect concentration) is based on acute toxicity test data. Since only few data are available from acute tests, EEC legislation (EEC, 1996) requires an assessment factor of 1000 (degree of uncertainty) on the lowest LC value avail50 able, yielding a PNEC value of 0.015 lg/liter. 8!5%3 In practice, one should make sure that the ozone-treated cooling water does not contain any dissolved ozone when it is discharged. This is relatively easy to achieve: when the ozone-containing water is mixed with a source of organic matter before discharge, free ozone reacts immediately with the organic matter and disappears. It should, however, be noted that this may not solve all problems associated with the use of ozone as an agent against biofouling. Not only is dissolved ozone harmful to aquatic life, but byproducts formed by the reaction of ozone with natural water components might be toxic too. Compounds such as aldehydes and peroxides formed during ozonation are of concern for human health (Weinberg et al., 1993). CONCLUSION
Results indicate that ozone is very toxic to aquatic organisms. Neonates of D. magna are more susceptible to ozone than are fish larvae. No major difference in toxicity of ozone for daphnids was observed at the different test temperatures, and all three species of fish larvae are similarly sensitive to ozone. The 48-h LC for fish larvae [C. carpio (at 27°C), ¸. 50 idus (at 27°C), and C. gariepinus (at 32°C) ranges between 30 and 45 lg/liter, and the 48-h NOEC for D. magna is 11 lg/liter at 21°C and 16 lg/liter at a test temperature of 27°C. Most of the experiments were carried out at elevated
American Public Health Association (APHA), American Water Works Association, Water Pollution Control Federation (1989). Standard Methods for the Examination of ¼ater and ¼astewater, 17th ed. APHA Washington, DC. Asbury, C., and Coler, R. (1980). Toxicity of dissolved ozone to fish eggs and larvae. J. ¼ater Pollut. Control Fed. 52, 1990—1996. Bader, H., and Hoigne´, J. (1981). Determination of ozone in water by the indigo method. ¼ater Res. 15, 449—456. Bellar, T. A., Lichtenberg, J. J., and Kroner, R. C. (1974). The occurence of organohalides in chlorinated drinking water. J. Am. ¼ater ¼orks Assoc. 66, 703—706. Claudi, R., and Mackie, G. L. (1994). Practical Manual for Zebra Mussel Monitoring and Control. Lewis, Boca Raton, FL. Duvivier, L., Leynen, M., van Damme, A., and Ollevier, F. (1996). Fighting zebra mussel fouling with ozone. In Proceedings of the EPRI PSE Service ¼ater Systems Reliability Improvement Seminar—Ninth in the S¼SRI Series. EPRI, Daytona Beach, FL. EEC (1996). Technical Guidance Documents in support of the Commission Directive 93/67/EEC on Risk Assessment for New Substances and the Commission Regulation (EC) No. 1488/94 on Risk Assessment for Existing Substances. Office for Official Publications of the European Communities, Luxembourg. EPRI (1982). ¹he Effects of Chlorine on Freshwater Fish under »arious ¹ime and Chemical Conditions: ¹oxicity of Chlorine to Freshwater Fish. Project 1435, Palo Alto, CA. Gilbert, E., and Hoigne´, J. (1983). Messung von Ozon in Wasserwerken; Vergleich der DPD- und Indigo-Methode. gwf-¼asser/Abwasser 124, 527—531 Kaur, K., Bott, T. R., and Leadbeater, B. S. C. (1992). Effects of ozone as a biocide in an experimental cooling water system. Ozone Sci. Eng. 14, 517—530. Leitzke, O., Schwammlein, K., Baresel, M., Wenzel, W., and Bradsch, K. (1994). Ozonation of Cooling ¼ater Circuits. Report WEDECO, Herford. Rice, R. G., and Wilkes, J. F. (1992). Fundamental aspects of ozone chemistry in recirculating cooling water systems: data evaluation needs. Ozone Sci. Eng. 14, 329—365 Rook, J. J. (1974). Formation of haloforms during chlorination of natural waters. J. ¼ater ¹reat. Exam. 22, 234—243. Stephan, C. (1977). Methods for calculating an LC . ASTM (American 50 Society for Testing and Materials) Special Technical Publication 634 pp. 65—84. Sugam, R., Singletary, J. H., Sandvik, W. A., and Guerra, C. R. (1981). Condenser biofouling control with ozone. Ozone Sci. Eng. 3, 95—107. Van Benschoten, J. E., Jensen, J. N., Brady, T. J., Lewis, D. P., Sferrazza, J., and Neuhauser, E. F. (1993). Response of zebra mussel veligers to chemical oxidants. ¼ater Res. 27, 575—582. Wedemeyer, G. A., Nelson, N. C., and Yasutake, W. T. (1979). Physiological and biochemical aspects of ozone toxicity to rainbow trout (Salmo gairdneri). J. Fish. Res. Board Can. 36, 605—614. Weinberg, H. S., Glaze, W. H., Krasner, S. W., and Sclimenti, M. J. (1993). Formation and removal of aldehydes in plants that use ozonation. J. Am. ¼ater ¼orks Assoc. 85, 72—85.
41, 176—179 (1998)
ENVIRONMENTAL RESEARCH, SECTION B ARTICLE NO.
ES981696
Toxicity of Ozone to Fish Larvae and Daphnia magna M. Leynen,* L. Duvivier,- P. Girboux,‡ and F. Ollevier* *Laboratory of Ecology and Aquaculture, Katholieke Universiteit Leuven, Naamsestraat 59, B-3000 Leuven, Belgium; -Laborelec, Rhodestraat 125, B-1630 Linkebeek, Belgium; and ‡Electrabel, Regentlaan 8, B-1000 Brussels, Belgium Received June 16, 1997
Ozone can be used as an alternative to chlorination to control biofouling in cooling water systems. The possible negative environmental impact of a discharge of ozone-containing cooling water was investigated. The acute toxicity of dissolved ozone was determined for fish larvae of three species [Cyprinus carpio (at 27°C), Leuciscus idus (at 27°C) and Clarias gariepinus (at 32°C)] and to Daphnia magna (at 21 and 27°C). The results indicate that ozone is very harmful to aquatic life. Daphnids are more sensitive to ozone than fish larvae. The mean 48-h LC50 value for the larvae amounts to about 35 lg/liter, while the 48-h NOEC for D. magna was 11 lg/liter (at 21°C). It was concluded that, to protect aquatic life, discharged cooling water should not contain any dissolved ozone. This can be achieved in practice by mixing the treated cooling water with a source of organic substances before discharge, as free ozone will react immediately with organic matter and thus disappear. ( 1998 Academic Press
INTRODUCTION
Chlorine has been used for decades to control microand macrofouling in cooling water circuits of power plants. However, it was demonstrated years ago that chlorine-treated cooling water discharges are toxic due to small amounts of free chlorine or of primary reaction products such as chloramines (EPRI, 1982). Chlorine is also responsible for the formation of toxic organohalogens (chloroform, haloacetic acids, etc.) (Bellar et al., 1974; Rook, 1974). In search of possible alternatives to chlorination, several studies have evaluated the effectiveness of ozonation in controlling biofouling (Sugam et al., 1981; Claudi and Mackie, 1994; Duvivier et al., 1996). Ozone is a stronger oxidant than chlorine and is able to remove existing biofilms (Kaur et al., 1992) and to kill veligers of zebra mussels (Dreissena polymorpha) at low concentrations (Van Benschoten et al., 1993). It is a powerful biocide and will oxidize many organic and inorganic substances. The use of ozone in cooling water treatment also allows reduction of existing COD and AOX levels (Leitzke et al., 1994), which is an ecological advantage compared with chlorine.
Ozone is a very unstable molecule and, after injection into raw water, it decomposes very rapidly. In the presence of organic impurities, the half-life of ozone is reduced to minutes (Duvivier et al., 1996). The rate of decomposition also increases with increasing pH. At low pH ((7), molecular ozone (O ) is the dominant species. As pH increases, 3 O turns into the very short-lived (microseconds) hydroxyl 3 radicals (OH°) (Rice and Wilkes, 1992). The high reactivity makes it difficult to maintain sufficient high ozone concentrations throughout the cooling water system to kill all fouling micro- and macroorganisms. On the other hand, the rapid dissipation of ozone normally ensures that there will be a small environmental impact at the discharge zone of the power plant. The present study investigates the possible adverse impact of discharged dissolved ozone in terms of acute toxicity to fish larvae (Cyprinus carpio, ¸euciscus idus, and Clarias gariepinus) and Daphnia magna. The acute toxicity curves (with corresponding LC values) were deter50 mined in laboratory tests.
MATERIAL AND METHODS
Experimental Setup The toxicity tests in the laboratory were carried out by means of the setup in Fig. 1, which is a modified version of systems that have been applied succesfully by other investigators (Wedemeyer et al., 1979; Asbury and Coler, 1980). Ozone is generated from pure oxygen with two aquarium ozonizators (Sander GmbH, Germany). They have a capacity of 200 mg ozone/h each. The oxygen flow through the ozonizators was held constant at 1.5 liters/min. A PVC column with a height of 2 m and a diameter of 20 cm serves as a contact chamber between water and gaseous ozone. Ozone is injected at the bottom of the column while the water flows into the column at the top (countercurrent principle). After 3 to 4 h, a constant ozone gradient is created, with the highest ozone concentration at the bottom. The contact column contains five branches. Each branch samples a different ozone concentration (according to the gradient) and leads to a set of three aquaria (2-liter volume),
176 0147-6513/98 $25.00 Copyright ( 1998 by Academic Press All rights of reproduction in any form reserved.
OZONE TOXICITY TO FISH LARVAE AND D. magna
FIG. 1.
177
Experimental setup.
in which the toxicity tests take place. The ‘‘ozone-enriched’’ water flows with a constant velocity through the test aquaria. After passing the aquaria, the test water is recovered in tank A and pumped to tank B. From tank B, the water flows by gravity back into the column. The whole setup is in fact a flow-through system with recovery of the test water. Due to the rapid dissipation of ozone, the concentration in tank B was reduced to undetectable levels. A separate setup without ozonization was used as a reference. The test conditions (temperature and water velocity) in the control experiment were the same as in the toxicity test setup. The water used in the tests was dechlorinated tap water. After the ozone gradient was created, the pH measured 8.5$0.1 during the tests. Ozone Measurement At regular time intervals, the ozone concentration was measured with a spectrophotometer by means of the indigo method (Bader and Hoigne´, 1981). Ozone decolors indigo trisulfonate under acid conditions. This decoloration can be measured at 600 nm. The detection limit is about 5 lg/liter
ozone. Compared with other methods (e.g., DPD method), the indigo method excludes all interference from other products (Gilbert and Hoigne´, 1983). The ozone concentration is calculated with the formula (APHA, 1989) mg ozone/liter"100](A!A@)/f]b]» where A!A@"difference in absorbance between sample and blank, b"path length of the cell (in cm), »"volume of the sample (in ml), and f"0.42 (correction factor). Organisms Fish larvae were collected from a hatchery (carp and ide) or obtained from this laboratory culture (African catfish). The fish were fed twice daily with commercial larval food until 2 days before the test. They were acclimated to the test water and the test temperature for at least 2 weeks in 30-liter aquaria. Mean total length ($SD) and mean total wet weight ($SD) of the test animals are given in Table 1.
178
LEYNEN ET AL.
TABLE 1 Length and Weight of the Fish Larvae
TABLE 2 48-h LC50 Values (lg/liter) for Fish Larvae Exposed to Ozonea
Mean total length (cm)
Mean total wet weight (mg)
Binomial method
1.1$0.1 1.7$0.2 1.1$0.1
17$3 50$8 11$2
Carp Cyprinus carpio ¸euciscus idus Clarias gariepinus
Ide African Catfish
The stock of D. magna used in the tests was procured from the University of Ghent, Laboratory of Animal Ecology, and has been maintained in this laboratory since February 1993. Daphnids were held at 21°C with a light/darkness period of 16/8 h. They were fed once daily with unicellular green algae (Scenedesmus acutus and Selenastrum capricornutum). Toxicity Tests Using the five branched outlets from the column with ozone-enriched water, two replicates of 10 fish larvae were exposed to each of five test concentrations of ozone. The test duration was 48 h and the temperature was held constant at 27°C for carp and ide and at 32°C for African catfish. Dead animals were removed regularly. The experiment was carried out twice with carp and ide. The fish larvae were not fed during the test. Neonates (less then 24 h old) of D. magna were exposed to five concentrations of ozone during 48 h (two replicates of 10 daphnids each). The neonates used in the test came from the second or a later brood of females kept in stock culture. The test was carried out at two different temperatures, 21 and 27°C. Test animals and stock culture were not acclimated at 27°C. Daphnids were not fed during the test. Where possible, LC values with 95% confidence limits 50 were calculated with the binomial method and by means of probit analysis (Stephan, 1977). Given an experimental design, it should be noted that the 95% confidence limits refer to the within-experiment repeatability rather than the interexperimental error. The latter can be estimated from the comparison of average values obtained in different experiments. RESULTS
Fish Larvae Forty-eight-hour LC values for the fish larvae are pro50 vided in Table 2. No mortality occurred in the control groups. Calculations used measured levels of dissolved ozone in the test aquaria. Two different calculation methods were reported. Overall, the 48-h LC values are very sim50 ilar for the different fish species, ranging between 30 and 45 lg/liter.
31 44 36 39 35
(26—36) (41—61) (27—48) (33—46) (21—64)
Probit analysis 31 (28—33) 44 (43—45) n.d. n.d. 37 (29—46)
a Ninety-five percent confidence intervals (in parentheses) were calculated separately for each experiment. n.d., could not be determined.
Daphnids Tables 3 and 4 indicate the mortality of D. magna neonates in the presence of different ozone concentrations at test temperatures of 21 and 27°C, respectively. At the highest concentration tested, 140 lg/liter, all daphnids were killed within 1 h of exposure. After 24 h, 100% mortality was obtained with 21 lg/liter, whereas 11 lg/liter had no effect on mortality during 48 h of exposure. Because it was impossible with the experimental setup to create a gradient between 11 and 21 lg/liter the 48-h NOEC was determined instead of the 48-h LC . At 21°C, the 48-h NOEC for 50 ozone for D. magna was 11 lg/liter. Very similar results were obtained at 27°C (Table 4) and the resulting 48-h NOEC was 16 lg/liter. DISCUSSION
When ozone is used as an alternative to chlorination in power plant cooling systems, it should be kept in mind that dissolved ozone is very toxic to aquatic organisms. Fortyeight-hour LC and NOEC values are indeed low 50 (10—45 lg ozone/liter), 0.3 mg/liter ozone kills catfish larvae in 1 h (Leynen, personal observation) and 0.14 mg/liter over 1 h is enough to kill neonates of D. magna. According to Wedemeyer et al. (1979), death of fish following acute
TABLE 3 Mortality Data of Daphnia magna Neonates Exposed to Ozone at 21°C O concentration 3 (kg/liter) Control 11$3a 21$5 57$16 93$25 140$19 a $SD.
Exposure period (h)
Mortality (%)
48 48 24 4 2 1
0 0 100 100 100 100
OZONE TOXICITY TO FISH LARVAE AND D. magna
TABLE 4 Mortality Data of Daphnia magna for Ozone at 27°C O concentration 3 (lg/liter) Control 16$5a 23$5 42$10 97$17 137$21
Exposure period (h)
Mortality (%)
48 48 24 4 2 1
0 0 100 100 100 100
179
temperatures because the temperature in the discharge zone of the cooling water is higher than the usual standard test temperature. REFERENCES
a $SD.
exposure to ozone is most likely due to severe gill lamellar ephithelial tissue destruction accompanied by massive hydromineral imbalance. From the results, it is clear that the ozone concentration in the water at the discharge site should be diluted to 0.015 lg/liter or less. This initial PNECwater value (predicted no-effect concentration) is based on acute toxicity test data. Since only few data are available from acute tests, EEC legislation (EEC, 1996) requires an assessment factor of 1000 (degree of uncertainty) on the lowest LC value avail50 able, yielding a PNEC value of 0.015 lg/liter. 8!5%3 In practice, one should make sure that the ozone-treated cooling water does not contain any dissolved ozone when it is discharged. This is relatively easy to achieve: when the ozone-containing water is mixed with a source of organic matter before discharge, free ozone reacts immediately with the organic matter and disappears. It should, however, be noted that this may not solve all problems associated with the use of ozone as an agent against biofouling. Not only is dissolved ozone harmful to aquatic life, but byproducts formed by the reaction of ozone with natural water components might be toxic too. Compounds such as aldehydes and peroxides formed during ozonation are of concern for human health (Weinberg et al., 1993). CONCLUSION
Results indicate that ozone is very toxic to aquatic organisms. Neonates of D. magna are more susceptible to ozone than are fish larvae. No major difference in toxicity of ozone for daphnids was observed at the different test temperatures, and all three species of fish larvae are similarly sensitive to ozone. The 48-h LC for fish larvae [C. carpio (at 27°C), ¸. 50 idus (at 27°C), and C. gariepinus (at 32°C) ranges between 30 and 45 lg/liter, and the 48-h NOEC for D. magna is 11 lg/liter at 21°C and 16 lg/liter at a test temperature of 27°C. Most of the experiments were carried out at elevated
American Public Health Association (APHA), American Water Works Association, Water Pollution Control Federation (1989). Standard Methods for the Examination of ¼ater and ¼astewater, 17th ed. APHA Washington, DC. Asbury, C., and Coler, R. (1980). Toxicity of dissolved ozone to fish eggs and larvae. J. ¼ater Pollut. Control Fed. 52, 1990—1996. Bader, H., and Hoigne´, J. (1981). Determination of ozone in water by the indigo method. ¼ater Res. 15, 449—456. Bellar, T. A., Lichtenberg, J. J., and Kroner, R. C. (1974). The occurence of organohalides in chlorinated drinking water. J. Am. ¼ater ¼orks Assoc. 66, 703—706. Claudi, R., and Mackie, G. L. (1994). Practical Manual for Zebra Mussel Monitoring and Control. Lewis, Boca Raton, FL. Duvivier, L., Leynen, M., van Damme, A., and Ollevier, F. (1996). Fighting zebra mussel fouling with ozone. In Proceedings of the EPRI PSE Service ¼ater Systems Reliability Improvement Seminar—Ninth in the S¼SRI Series. EPRI, Daytona Beach, FL. EEC (1996). Technical Guidance Documents in support of the Commission Directive 93/67/EEC on Risk Assessment for New Substances and the Commission Regulation (EC) No. 1488/94 on Risk Assessment for Existing Substances. Office for Official Publications of the European Communities, Luxembourg. EPRI (1982). ¹he Effects of Chlorine on Freshwater Fish under »arious ¹ime and Chemical Conditions: ¹oxicity of Chlorine to Freshwater Fish. Project 1435, Palo Alto, CA. Gilbert, E., and Hoigne´, J. (1983). Messung von Ozon in Wasserwerken; Vergleich der DPD- und Indigo-Methode. gwf-¼asser/Abwasser 124, 527—531 Kaur, K., Bott, T. R., and Leadbeater, B. S. C. (1992). Effects of ozone as a biocide in an experimental cooling water system. Ozone Sci. Eng. 14, 517—530. Leitzke, O., Schwammlein, K., Baresel, M., Wenzel, W., and Bradsch, K. (1994). Ozonation of Cooling ¼ater Circuits. Report WEDECO, Herford. Rice, R. G., and Wilkes, J. F. (1992). Fundamental aspects of ozone chemistry in recirculating cooling water systems: data evaluation needs. Ozone Sci. Eng. 14, 329—365 Rook, J. J. (1974). Formation of haloforms during chlorination of natural waters. J. ¼ater ¹reat. Exam. 22, 234—243. Stephan, C. (1977). Methods for calculating an LC . ASTM (American 50 Society for Testing and Materials) Special Technical Publication 634 pp. 65—84. Sugam, R., Singletary, J. H., Sandvik, W. A., and Guerra, C. R. (1981). Condenser biofouling control with ozone. Ozone Sci. Eng. 3, 95—107. Van Benschoten, J. E., Jensen, J. N., Brady, T. J., Lewis, D. P., Sferrazza, J., and Neuhauser, E. F. (1993). Response of zebra mussel veligers to chemical oxidants. ¼ater Res. 27, 575—582. Wedemeyer, G. A., Nelson, N. C., and Yasutake, W. T. (1979). Physiological and biochemical aspects of ozone toxicity to rainbow trout (Salmo gairdneri). J. Fish. Res. Board Can. 36, 605—614. Weinberg, H. S., Glaze, W. H., Krasner, S. W., and Sclimenti, M. J. (1993). Formation and removal of aldehydes in plants that use ozonation. J. Am. ¼ater ¼orks Assoc. 85, 72—85.