Pergamon PH: S0273-1223(97)00746-4
War. Sci. Tech. Vol. 36, No. 12, pp. 299-307, 1997. © 1997 IAWQ. Published by Elsevier Science Ltd Printed in Great Britain. 0273-1223/97 $17·00 + 0·00
EFFECT OF OZONATION FOR TREATMENT OF MICROPOLLUTANTS PRESENT IN DRINKING WATER SOURCE Joon-Wun Kang*, Hoon-Soo Park*, Rong-Yan Wang*, Minoru Koga**, Kiwao Kadokami***, Hyeon-Yeoul Kimt, Eung-Taek Lee:t: and Sung-Min Oh
Yonsei University, Korea
** University of Occupational & Environmental Health, Japan ** Kitakyushu City Institute of Environmental Health Sciences, Japan t Daewoo Corporation :t Samsung Corporation 11 LG Corporation, Korea
ABSTRACT Pilot tests were perfonned to investigate the effectiveness of ozonation for the treatment of drinking water. Prior to the experiment, four regional target water were analyzed for detennining target compounds. Various organics including pesticides were identifiedand present also in the conventionally treated tap water. The contamination of pesticides in most raw water was severe and ozonation was found to be effective to remove pesticides significantly. Other organic species were removed effectively as the order of aromatic amines, nitro compounds, ketones and ethers. On the whole, volatile organics were removed effectively than DOC, CH2Cl2 extractable organics. Accompan~ing with the increasement of AOC, aldehydes have increased after ozonation and reduced by post-GAC. Also, bromate was produced after ozonation and it was validated to suppress the production of bromate on the presence of ammonia and DOC. © 1997 IAWQ. Published by Elsevier Science Ltd
INTRODUCTION This study investigated the efficiency of ozone and PEROXONE AOP (Advanced Oxidation Process) for the treatment of various micropollutants present in various surface water in Korea with following objectives: (1) to measure trace organic contaminants in different sources, and to decide target compounds to remove; (2) to compare the efficiency of ozone and PEROXONE for micropollutant removal and oxidant doses required; (3) to evaluate the hydrodynamic behavior of reactor system; (4) to identify organic and inorganic byproducts produced from oxidation. RESEARCH METHODS Experimentations were conducted with batch or pilot scale using various water samples taken from the M treatment plant, W plant, MT plant and C plant. The pilot system was installed at M treatment plant with the treatment capacity of 12 tons/day. The pilot system was consisted with flocculation/sedimentation basin, sand filter, ozone contact column, and GAC column as shown in Fig. 1. JWST 36: 12-K
1.- W KANG et al.
I Raw water resC1"VOlr
4 03 contaclDr
3. Sand filter
6 Treated water reser-....Ol!
- + Postozonatlon
...... .. -,.. Preozonatlon
H Holdtng ank
Figure I. Schematic diagram of continuous plant (M plant).
For the comparison of preozonation and postozonation process, pipe lines for transItIOn between dual processes were installed. Ozone contact column had a height of 300 cm, i.d. of 16 cm and water capacity of about 60 L. Disc type diffuser for ozone sparging was be installed and water was induced from up to down with a countercurrent flow. Corona arc type ozone generator was used for ozone generation with pure oxygen and ozone concentration was checked continuously with Ozone monitor (PCI HC-NEMA 12). Ozone doses for variation of water quality were adjusted with concentration near ozone demand. GAC columns had a height of 180cm and a i.d. of 30cm. Activated carbon was filled in column with 124 L capacity. EBCT of GAC was 14-15 min when water flow of plant was 8L/min. The details of capacity, contact time for each process unit were described in Table 1. Table 1. Capacity and retention time for individual unit process of pilot plant Raw water Mixing reserior basin
Coaguration Sedimentation Sand filter basin baSin
Capacity (L) 1000 34.7 320 Retention time 121 4 39 (min) - Treatment capacity of one day = 12 ton/day(S.3 Llnun)
Tracer tests were designed to predict the hydrodynamic behavior of our reactor and to analyze the effect of certain factors in the reactor. Physical parameters such as dispersion number and ozone utilization rate were measured to characterize the ozone column. The data of residence time distribution (RTD) were obtained from several tracer tests. NaCl was used as tracer with the amount of 5g per one injection. At the point of outlet, the detection of tracer was performed with a conductivity meter (HANNA inst co). To decide the target compounds to oxidize, water samples were analyzed by GC/MS after using various concentration techniques, such as purge and trap, liquid-liquid extraction with derivatization, and XAD resin accumulation, etc. Tenax GC column was used for purge & trap method and adsorbent was Tenax GR 60/80 mesh. Before using, preconditioning was performed with 300°C, Ihr and water of 25ml was purged with pure nitrogen (99.999%). Hewlett Packard HP-5890 GC, JEOL JMS-DX 303 Mass spectrometer and J&W DB-WAX column were used for purge and trap method. Injector, detector temperature were 220°C, 300°C and column temperature was 35°C-200°C with 5°C degree per minute. For liquid-liquid extraction method, water of lOOOrnl was extracted with dichloromethane of 150ml. For the first time, the sample was extracted with dichloromethane of lOOml and water layer after separation was extracted with dichloromethane of 50ml for the second time. This combined dichloromethane was dehydrated with an hydrous Na2So4 and then concentrated to Iml with K.D evaporator. l~L of concentrate was injected into Ion trap GC/MS. Varian 3400 GC, Finnigan Mat ITS 40 mass spectrometer and J&W DB-5ms column were used for LLE method.
Treatment of micropollutants
Injector, detector and transfer line temperature were 250°C, 260°C and 280°C and column temperature was 50°C-300°C with 8°C degree per minute. For XAD accumulation method, column was made with XAD-4 resin (Aldrich chemical co.) and stainless-steel pipe (250mrn length, 1/2inch I.D.) manually. Resin was used after preconditioning with methanol, dichloromethane in soxhlet extractor. Water of about 40-50 L was accumulated through XAD column with teflon head pump (Masterflex 50). After accumulation, resin was discarded and extracted with dichloromethane and soxhlet extractor. Dichloromethane was dehydrated with anhydrous Na2S04 and then concentrated to lml with K.D. evaporator. IIlL of concentrate was injected into Ion trap GelMs. GCIMS condition had such a same condition as LLE method. Aldehydes were analyzed for the identification of organic by-products after ozonation. Aldehydes wee known as a surrogate for the determination of AOC(Assimilable organic carbon).PFBO (0-(2,3,4,5,6pentafluorobenzyl)-hydroxylamine hydrochloride, Aldrich)-derivatization method allows the quantification of CI-ClO straight-chain aliphatic aldehydes, glyoxal and methylglyoxal in the raw water of four treatment plants before and after ozonation. Water sample of 5ml was used for the derivatization with addition of O.5ml PFBOA solution (lmg/ml). After the derivatization for 2 hours, the sample was extracted with nHexane containing internalstandard (DFBP, 400llglL) and luL of concentrate was injected into GC with ECD. HP-5890 GC, J&W DB-I column were used for the analysis of aldehydes. Injector, detector temperature were 120°C, 300°C respectively and AOC, analysis of AOC was performed with Pseudomonas fluroescens P 17, Spirillum NOX and Acinetobacter calcoaceticus found newly. Water sample for analysis of AOC was filtered with nylon filter having a O.21lm pore and cultivated. Bromate ion (Br03-) was analyzed for the identification of formation after ozonation of bromide-containing water. Having the draft MCL (Maximum Contamination Level) of 101lglL, the determination of bromate in drinking water needs more precise analysis method. Flow injection method based on the oxidation of chloropromazine by bromate ion allowed the detection of bromate having a low concentration and was very useful because of automatic analysis. Using this method, the detection limit of bromate could be low to l~gIL at least. Chloropromazine (Fluka, 99%+) is reacted with bromate at acidic condition and colorized to blue in a visible range. One four-channel peristaltic pump (Gilson, minipuls 3), one reaction coil and one HPLC UVNIS detector (SOMA Optical Co.) were used for the flow injection system of bromate analysis. Concentrations of chloropromazine and HCI were 100mglL, 1.32 M in total flow respectively. The optimum wavelength for the analysis of colorized chloropromazine was 530nm and absorbance peaks were recorded continuously using a Young-In D520B integrator. Table 2 Classification of micropollutants identified in several raw water w Cia ssification
Aliphatic compounds Bmz.cnes Polycyclic compounds Ethers Ketones Phenols Phthalates Aromatic amines Quinoline Nitro compounds Nitrosoamines Phosphoric esters Pesticides others
20 2 12 2 3 S 7 4 1 2 1 3 14 7
23 S 20 1 4 8 6 S 1 2
20 3 17 2 S 9 6 S
The No. oloJ'lanJo
August 22 3 13 2 4 10 6 8 1 2
22 49 22 2 2 11 7 10 1 2
23 3 13 1 3 S 7 2 1 2
4 35 7
2 22 8
2 14 7
2 45 9
3 15 3
8 2 3 1 7 1 1 2
DATA & DISCUSSION Before the verification for the effect of ozonation and PEROXONE AOP treatment, target water was analyzed and target organics was identified. The samples for analysis were defined ~y four regional .raw water and some raw water were analyzed on the considerations of rainy or non-ramy season. VarIOUS
micropollutants such as pesticides, aliphatic hydrocarbons, aromatic hydrocarbons, phenolic compounds, and a tasteand odor causing compound were identified and quantified in raw and treated water. Table 2 shows the species and number of organic identified in raw water. The maximum number of organic and pesticide were 139 and 45 respectively. The number of organics was increased at August, a rainy season. Of this compounds, pesticides' contamination was the most significant. The concentration level of pesticide was particularly high in August exceeding .EC limit of 500 ppt in some water.
Figure 2. Comparison of total sum of pesticides for target raw water (3, 4, 8 indicate March, April, August respectively).
Figure 2 shows the comparison of total concentration of pesticide in different raw water. The raw water of Han river M(8) and Nak-dong river C(8) had the exceeding value over 500ppt. In the same raw water, the concentration of pesticide at August had a higher value than March or April because the use of pesticide is increased for extermination of vermin. Isoprothiolane, diazinon, tricyclazole, iprobenfos, fenobucarb and nitrofen were found to have more higher concentration than other pesticides. These pesticides also had the trend that the concentration level at August shows higher than March, April. The total sum of CH 2Cl 2 extractable organics shows the most highest level in the raw water of Nak-dong C(8) at August. The contamination of phthalate was sever in the most raw water. Particularly, the raw water of Nak-dong C(8) was analyzed as a result of having a most severe contamination of phthalates. Dimethyldisulfide was found to be the odor causing compound in the raw water of Han river at March. Tracer test was performed for the hydrodynamic behavior of fluid in the continuous ozone contactor in M drinking water plant. The capacity of ozone contactor was about 60 L and the flow rate of water flow could be adjusted from 3.5 to 11.3 Umin. Therefore, the mean contact time of flow was varied from 6 to 19 min. Gas flow rate was varied from 0 to 10 Umin and the operation of counter-current flow was performed. In this situation, tracer test was performed as each condition was changed. Principally, the axial dispersion model and the tanks-in-series model are two typical models widely used to characterize the fluid flow in a real reactor. Tanks-in-series model was adapted to simulate this tracer test because of its simplicity and another advantage. This model permits to introduce the new parameters such as the dead zone, the shortcircuit or the recycle assumed to be existed in a real reactor. In this case, the tanks-in-series model was calculated with the sum of reactors having equal size, J as follows:
Ozone contractor was found to be the CSTR in cascade having no dead zone (well-mixed tanks in series). The number of the reactor in series J depends on the gas and liquid flow rates.
Treatment of micropollutants
Table 3 Summary of tracer tests for different conditions (M plant)
N° 1 2 3 4 5 6 7 8 9 10
Ug mJh 0 6.8 20.4 34 0 6.8 20.4 34 0 6.8 20.4 34
G I/mio 0 2 6 10 0 2 6 10 0 2 6 10
15 9 7 5 20 7 5 6 30 9 9 7
0.46 0.30 0.26 0.23 0.58 0.43 0.40 0.36 0.62 0.50 0.38 0.38
1.11 1.11 1.26 0.95 1.20 1.41 1.63 1.52 1.00 1.33 1.33 1.38
Table 3 summarizes the results of tracer test from RTD model. The number of tanks-in-series depends on the ratio of gas and liquid flow rate. Having the flow rate of gas over 2 Umin, the number of J have results under 9. It means,that the reactor is close to plug flow when the flow rate of gas is increased at the fixed flow rate of water. And the'number of J increased as an the increasement of liquid flow rate for a given flow rate of gas. The term T 10ft' T 90/t (These time parameters indicate when the tracer concentration is detected as 10% and 90% of the total amount on outlet area) indicate that the increasement of gas flow improved the mixing behavior of contactor. Considering the result of the tracer test, it was found that the flow rate of gas and liquid can affect the treatment efficiency of target water. Ozonation was found to be very effective to remove pesticides, aromatic amines, and nitro compounds, but ineffective to remove aliphatic compounds. Figure 3 shows the order of removal trends of different organic species as a average. Removal of pesticides by ozonation was more significant than other organic species. Pesticides had the removal percent of 90% in most raw water.
:;2 > 0
60.0 50.0 40.0 30.0 20.0 10.0 0.0
e " .S!
Figure 3. Removal trend of organic species by ozonation (Ozone dose: 0.8-1.5mgIL).
J.-W. KANG etal.
250 200 ..... 150 0. 0.
Figure 4. Comparison of chlorinated tap water and ozonation for removal of pesticides (Target water: Han river M raw water, Ozone dose: ImgIL).
Figure 4 shows the degradation of pesticides after ozol'}ation comparing with conventional treatment process containing chlorination. The effectiveness of ozonation was improved in the full treatment process as shown in Fig. 5.
endosulfasulfate tricyclazole iprobenfos
Figure 5. Reduction of pesticides through the process containing post-ozonation.
pencycuron i so f enphos oxon carbaryl (HAC) cnlorfenvinpnos iprobenfos( IBP. Ki tuin P) f.nobucarb isoprolniolan. di az inon
Figure 6. Removal of pesticides during 03 process or 031H202 process (Target water: Han river M raw water at March, 03 dose O.8-1.4mgIL, H202.03 ratio+O.3).
Treatment of micropollutants
The primary breakdown of nitro compounds, aromatic compounds and benzene show the trends of rapid removal efficiency over 70%. Other aromatic compounds such as polycyclics, phenols, phthalates had the trend of lower removal than the precedents. Dimethyldisulfide, the compound causing odor, was reduced rapidly in 15 min when ozone was induced with the rate of 0.28mglL/min as the test batch. Most volatile organics identified with purge & trap method were removed very effectively except the saturated hydrocarbons. But the removal rate of DOC, CH 2Cl 2 extractable organs were more lower than volatiles. Figure 6 shows that the process of ozone only is sufficient to remove various pesticides in the raw water of Han rive. PEROXONE AOP is very effective process to enhance destruction rate of pollutants, but in some water, the process of ozone alone was as effective as PEROXONE AOP due to promoters present in raw water. Aldehydes are increased as a result of ozonation, with acetaldehyde being the most prevalent. AGe had the trend of increasement after ozonation and Fig. 7 shows the increasement of AOe determined with Acinetobacter calcoaceticus after ozonation. Therefore, the good correlation between aldehydes and Ape were shown in the most raw water. Once formed, aldehydes can persist in treatment lane, but decreased to some extent after GAC filter. 800
500 400 300
I_ Acetate eq. _ Oxalate eq·1
Figure 7. Increasement of assimilable organic carbon by increasement of ozone dose (Target water: Youngsan river MT raw water).
The degradation of three aldehydes after post-GAC was shown in Fig. 8. ppb
Raw • Formaldehyde
Figure 8. Aldehyde formation after ozonation and reduction by granular activated carbon (GAC).
A flow injection method has been explored for the detennination of bromate ion in ozonated wat~r, and an effect of ammonia on the bromate production was investigated. Bromate formations were detect~d 10 several raw water containing bromide ion. Bromate production was reduced under presence of arnmoma and DOC but it was not reduced proportionately by increasing of ammonia and DOC concentration. Figure 9 shows that the formation of bromate canbe suppressed on the presence of ammonia and DOC in a most raw water. 180 160
100 80 60 40 20 0
« Gi iG ~
en Gi iG ~
Figure 9. Bromate formation in various source water (Initial bromide: 250ppb Ozone dose: ImglL, CT: 10 min).
CONCLUSIONS The results from the pilot test and several others could demonstrate the effectiveness of ozonation or PEROXONE AOP for the treatment of drinking water. Comparing with a conventional system of drinking water treatment, the application of ozonation gave the evident improvement for the breakdown of pesticides and other organic species. And ozonation also showed the ability to remove a odor-causing compound clearly. Organic and inorganic byproducts could be suspected as the fault of ozonation, on the other hand, it was validated that the capability of removal by GAC can be enhanced and the degradation of ammonia by bromate canbe possible. ACKNOWLEDGEMENT These results were attained from the works funded by the Korea Institute of Construction Technology under G7 project planning. REFERENCE Aieta. E. M., K. M. Reagon, J. S. Lang, L. McReynolds, 1. W. Kang and W. H. Glaze (1988). AdvancedOxidation Processes for Treating Groundwater Contaminated with TCE and PCE: Pilot Scale Evaluations. Journal of American Water Works Association, 80(5), 64. Le Sauze, N., Laplanche, A., Orta de Velasquez, M. T., Martin, G., Langlais, B. and Martin, N. (1992). The Residence Time Distribution of the Liquid Phase in a Bubble Column and its effect on Ozone Transfer. Ozone Sci. & Eng., 14,245-262. Marinas, B. 1., Liang, S. and Aieta, E. M. (1993). ModelIing Hydrodynamics and Ozone Residual Distribution in a Pilot-scale Ozone Bubble-diffuser Contactor. J. Am. Water Works Assoc., 85, 90-99. Roustan, M., R. Y. Wang and D. Wolbert (1996). Modeling Hydrodynamics and Mass Transfer Parameters in a Continuous Ozone Bubble Column. Ozone Sci. Eng, 18,95-115. Yamada,H. and I. Somiya (1989). The Determination of Carbonyl Compounds in Ozonated Water by the PFBOA method. Ozone Sci. Eng.• II, 127-141. Staehelin, J. and J. Hoigne (1985). Decomposition of Ozone in the Presence of Organic Solutes Acting as Promoters and Inhibitors of Radical Chain Reactions. Envir. Sci Technol., 19, 1206. Kadokami, K., Sato, K., Hanada, Y., Shinohara, R., Koga, M. and Shiraishi, H. (1995). Simultaneous Determination of 266 Chemicals In Water at pt Levels by GC-lon Trap MS. Analytical Sciences, June, 11,375-384.
Treatment of micropollutants
Gordon, G., B. Bubnis, D. Sweetin and C. Y. Kuo (1994). A flow injection, Non-Ion Chromatographic Method for Measuring Low Level Bromate Ion in Ozone Treated Waters. Ozone Sci. ~ng., 16,79-87. von Gunten, U. and J. Hoigne (1993). Bromate fonnation during Ozonation of Bromide-containing waters. Proceedings of Eleventh Ozone World Congress San Fransisco, volume I, California.