ZHANGJinsong ^1*, DONG Wenyi ²,LI Ji ¹, FAN Jie ¹
(1. Shenzhen WaterGroup, Shenzhen, 518031, China; 2. School of Municipal and EnvironmentalEngineering, Harbin Institute of Techonlogy, Harbin, 150008, China)

ABSTRACT:Ozonation andbiological activated carbon (BAC) process has become one of the most popularadvanced treatment technologies to ensure biological and chemical safety ofdrinking water, especially in the southern China where source water isseriously contaminated. Four safety-associated problems, namely biological stabilitymaintenance, turbidity and bacteria control, disinfection by-product precursorsremoval and mutagenicity abatement, were studied in this paper. Experimentalresults indicate that ozonation led to the increase of assimilable organic carbon (AOC)concentration; however, BAC filtration following ozonation was effective toreduce AOC and maintain biological stability. Bottomsand layer beneath BAC bed trapped part of the flakes of biomass falling fromBAC surface, and ensured the low turbidity for effluent from BAC filter.Ozonation process was capable of removing trihalomethane formation potential(THMFP) and haloacetic acid formation potential (HAAFP); BAC had considerableremoval efficiency for the HAAFP, but only limited effect on THMFP. Mutagenicity increasedafter raw water treated by conventional process; however, the ozonation and BACfiltration process reduced the mutagenic activity remarkably, and afterdisinfection, mutagenicity maintained a stable level. Inconclusion, ozonation-BAC process makes full use of both ozone and BAC, andforms a cooperative system that is effective to remove pollutants and ensurethe drinking water safety.
KEYWORDS Ozonation, biological activated carbon, drinking water safety

1　INTRODUCTION

Conventionaltreatment process for drinking water aims at removing turbidity and bacteriafrom source waters. However, increasingly polluted water resources make conventionalprocess difficult to meet stringent standards in terms of chemical andbiological safety for drinking water quality, and therefore, necessitate theadoption of advanced technology.

Nowadays,ozonation–BAC treatment process, which integrates strong oxidation anddisinfection capability of ozone and high adsorption and biodegradationcapacity of BAC, has become one of the most popular advanced technologies andbeen put into practice rapidly over the world. In China, several drinking waterplants in cities such as Shanghai, Hangzhou and Shenzhen, have adoptedozonation–BAC process or plan to adopt in the near future ^{[[1],[2],[3]]}.However, with the development of analytical methods and toxicologic studies,some problems were raised, such as ozonation associated by-products andbiological stability, and BAC associated biological safety. These problems havebeen a great concern in recent years.

This researchfocused on biological stability maintenance, turbidity and bacteria control, disinfectionby-product precursors removal, and mutagenicity abatement in ozonation-BACtreatment process for drinking water. It aimed at providing technical supportinformation for this advanced process application to Bijiashan Drinking WaterTreatment Plant in Shenzhen City, China.

2　 EXPERIMENTS

2.1 Treatment Process

This researchwas carried out on pilot scale in Shenzhen City in the southern China. Thesource water was obtained from a reservoir in this city, characteristics ofwhich are summarized in Table 1.

Table 1.　Characteristics of Source Water

The flow sheet of the process in the experiments is shown in Fig. 1.

Fig. 1.　 Flowsheet ofTreatment Process

Key parametersfor the pilot-scale experiment are summarized as follows:

TreatmentCapacity: 3m³/h.

Pre-ozonationReactor: Counter-current reactor with a diffuserwas used; height = 3.0m; available water depth = 2.5m; inner diameter = 300mm;HRT = 4.5 min; Ozone generator type was CFS-1A (Ozonia Ltd, Switzerland); Ozonedosage =1.5mg/L.

Mechanicalmixing tank: mixing time = 6 seconds; Aluminum poly-chloridesolution (10% Al₂O₃) was used as coagulant.

Cyclonereaction tank: reacting time = 23 min.

Tube settler: HRT = 36 min.

Sand filter: Uniform silica sands were used as filter media; filtering velocity= 10 m/h.

Intermediate OzonationReactor: Same as pre-ozonation reactor except:tower height = 6.0m; available water depth = 5.7 m; inner diameter = 400mm; HRT= 16 min; Ozone dosage =2.0mg/L.

BiologicalActivated Carbon (BAC) filter: The filter wasconsist of two identical units; Total height = 4.9m; BAC bed thickness = 2m ifnot specified in following text; empty bed contact time = 10min; filteringvelocity = 12m/h. The carbon type was ZJ-15 column activated carbon (1-1.5 mmin diameter, 3-5 mm in length) provided by Ningxia Taixi Activated CharcoalFactory.

2.2 Analytical Methods

Turbidity was measured by a HACH turbiditmeter (model2100) in nephelometric turbidity units (NTU).
pH value was measured by electrometric method witha pH meter (ThermoOrion model 868).
Potassium permanganate index was analyzedfollowing the standard method ^[[4]].
NH₃-Nwas determined by Nessler’s reagent colorimetric method with a UV-Vis spectrophotometerprovided by Beijing Purkinje General Instrument Co. LTD, China (model:PU-1901).
NO₂^--N was determined bycolorimetric method ^[[5]] with the UV-Visspectrophotometer.
Algae were counted using a microscope.
Total bacteria count was measured using standardplate count method ^[4].
AOC was analyzed by the standard method ^{[[6],[7],[8],[9],[10]]}. The maximum growth of P. fluorescens P17 andSpirillum sp. strain NOX was converted to the amount of AOC.
Chlorine wasmeasured using a Hach chlorine analyzer (model CL17)
Mutagenicactivity was assayed by using Salmonella typhimurium strains TA98 and TA100 byfollowing the Ames method ^[5 ,[11]].
Formationpotentials were used to quantify the concentrations of trihalomethanes (THMs)and haloacetic acids (HAAs) precursors. The chlorine dose for samples was determinedby maintaining 3~5mg/L residual chlorine at the end of incubation, which wasperformed in the dark at 25 ± 2℃ for 5 days. A phosphate buffer wasadded to maintain the pH at 7.0. THMs were measured using EPA Method 551 ^[[12]], and HAAs were measured usingEPA Method 552.0 ^[12].

3 RESULTS ANDDISCUSSIONS

3.1 Biological stability Maintenance

Biologicalstability refers to a condition wherein the treated water quality does notenhance biological regrowth in the distribution system ^[[13]].Maintenance of biological stability requires removing nutrients from the waterprior to distribution, especially biodegradable organic matters. Since theconcentrations of biodegradable organic compounds in drinking water wasrelatively low and difficult to quantify by chemical analytical methods, assimilableorganic carbon (AOC) was defined as a surrogate parameter for the measurement of potential bio-regrowthand biological stability ^[6].

Table 2 givesthe variation of AOC concentration in the process.

Table 2. Variation of AOC Concentrations in Treatment Process *

* Pre-ozonation was not in operation.

Table2 shows that AOC concentration was only slightly affected by conventional treatmentprocess (flocculation, sedimentation and sand filtration), probably because theAOC fraction was composed of small molecular weight, non-humic compounds thatare not amenable to coagulation ^[[14]]. Hem and Efraimsen found AOCin NOM with molecular weight less than 1000. Therefore, conventional treatmentprocess mainly removing large molecular organic matters was capable of a smallreduction of AOC level ^[14,[15]].

AOC increasedsharply by about 80% after ozonation process and resulted in biologicalinstability. Ozone can oxidize organic matters to aldehydes, aldoketoacids, carboxylicacids, and other AOC forming matters that cause microbial regrowth ^[[16],[17]]. However,the oxidizing ability was subjected to the characteristics of raw waters.Literatures showed that AOC level was about 2 to 4 times higher after ozonation^{[16,[18],[19]]}. In this study, only about afactor of maximum 2.3 was observed (Table 3). Ozone dosage was another possiblekey factor to influence the AOC formation. Table 3 presents effects of ozone dosageon AOC concentration.

Table 3. Effects of Intermediate Ozone Dosage on AOC Concentration

Underozone dosage of 1.0mg/L, AOC concentration doubled. However, when ozone dosageraised to 3.0mg/L, only additional 32mg/Lincrease of AOC concentration was observed. Raising ozone dosage to 4mg/L, AOCconcentration began to descend instead of increasing. This result shows thatthere is a maximum AOC level during ozonation. This might be explained by tworeasons. Firstly, with the increase of ozone dosage, the organic matters thatare easy to be oxidized decrease, therefore, the AOC formation rate becomeslower. Secondly, when high ozone dosage is applied, part of the AOC might befurther oxidized and mineralized.

BAC filtrationshowed good efficiency for AOC removal, reaching 47.9%. This demonstrated thatBAC filter was indispensable to counteract the negative effects caused byozonation and maintain biological stability. GAC was very effective to remove AOC because of its high adsorptioncapacity for small molecular organic matters ^[[20]].Total AOC removal efficiency of more than 80% was reported ^{[19 ]}. In this experiment,the granular activated carbon (GAC) in the filter had been put in operation formore than one year, and then off-stream for a period of time, its adsorptioncapacity had reduced significantly – the iodine value was only 70% of virginGAC. Consequently, AOC reduction in BAC filter was mostly achieved bybiological degradation instead of activated carbon adsorption. This was alsoapproved by experiment results in non-GAC biofilters, in which similar removalefficiencies were observed, e.g., 50% in a biologically active dual media filter(sand and anthracite, EBCT = 6min) ^[[21]] and 45% ina bio-ceramic bioreactor ^[19].

EBCT is a keyparameter that determines AOC removal efficiency ^[[22]]. The impact of BAC EBCT on AOCconcentration at ozone dosage of 3.0 mg/L is shown in Table 4.

Table 4. Effect of EBCT on AOC in BAC Filter

* Ozone dosage 3.0 mg/L.

Table4 shows that AOC concentration was noticeably high after ozonation. When EBCTwas 7.5min, AOC concentration decreased about 34.5%, and raised the EBCT to 10min, a much greater reduction (58.7%) was observed. However, when EBCTcontinued to increase, AOC maintained at a relatively stable level. This wassimilar to the kinetic modeling results by Zhang and Huck ^{[22 ]}. They defined a dimensionless EBCT parameter,and found this parameter determined AOC removal, which increased convexly withincreasing EBCT. Namely, beyond a certain value of EBCT, further increaseachieved little improvement.

3.2 Turbidity and Bacteria Control

BAC filter washelpful for removing organic matters and the improving biological stability;however, it also exerted adverse impact on turbidity, especially from theflakes of biomass falling from BAC surface. This dilemma is difficult toresolve by merely optimizing BAC filter.

To meet thestrict stringent of turbidity – the goal of turbidity for the projectedBijiashan Plant is 0.1 NTU – bottom sand layer (thickness: 0.3m) beneath theactivated carbon layer (thickness: 2.2m) was tested in this research. Two typesof carbon were chosen, namely column GAC as stated above and crushed GAC(Filtrasorb 300, 8 × 30 US mesh, provided by Calgon Co., U.S.). After twomonths’ operation, biofilm had been formed on carbon surface and GAC had transferredinto BAC. Table 5 gives the turbidity of effluent from both BAC bed and sandlayer.

Table 5.　 Effects ofBottom Sand Layer on Turbidity (NTU)

From Table 5, it can be seen that an average 0.02 NTU reduction wasachieved by sand layer. This proves the positive effect of sand layer onturbidity control for BAC effluent.

Total bacterialcount was also monitored for both BAC bed and sand layer, as shown in Table 6.

Table 6. Effect of Bottom Sand Layer on Bacteria Control　 (Unit: count/mL)

Table 7. Bacteria Control Before and After BAC Filter Backwash(Unit: count/mL)

* In operation for 3 weeks

Table 6 and 7 show that the sand layer’s ability to restrainbacteria impaired after a 3 weeks’ operation, but it recovered after backwash-- the sand layer recuperated its ability to trap the bacteria. Therefore, itwas important to control backwash cycle for controlling turbidity and bacteriain the BAC filter effluent.

Total bacterialcount in effluent from sand filter was fairly low, and after ozonation, itdecreased to below detection limit. However, after BAC filtration, totalbacteria count increased more or less. To testify the impact of bacteriaincrease on chlorine consumption, effluents from ozonation, BAC bed and sandlayer were chlorinated (Cl₂ dosage: 1.25mg/L) and residual chlorinewas tested.

Table 8.　 Variation ofResidual Chlorine (mg/L)

Table8 shows that residual chlorines in effluent from BAC filters (both BAC bed orbottom sand layer) were slightly lower than that from ozonation, demonstratingthat the possible bacteria increase after BAC filtration led to only a smallsurplus chlorine consumption. However, this surplus chlorine was negligiblecompared to the residue chlorine, especially in the long-term – residualchlorines in all samples were almost at the same level after 5 hours. Theseresults proved that BAC filtration did not led to noticeable surplus chlorineconsumption, and disinfection at normal dosage was effective to remove bacteriareleased from BAC filter, and ensure the biological safety without surpluschlorine demand

3.3 Disinfection Precursors

Generally, it isbelieved that chlorination-derived disinfection by-products (DBPs) precursorsremoval before disinfection favors DBPs control. The oxidation of DBPs by ozoneis slow because of their high degree of halogenation, which leads to a lowelectron density at the carbon centers. Even with OH radicals, only iodinatedTHMs show a high reactivity. Hence, ozonation is not a suitable process for theremoval of DBPs formed during chlorination ^[[23]].Removal effectiveness by activated carbon is determined by the molecular sizeof DBP precursors – small and middle sized precursors is easy for adsorption,but the big sized are difficult to disperse into the microspores of GAC.

THMs and HAAsare typical DBPs. Their formation potentials in the treatment process wereanalyzed in this research (Table 9).

Table 9. The variation of THMFP and HAAFP in treatmentprocess

From Table 9, itcan be seen that pre-ozonation plus sedimentation has a similar percentageremoval for both THMFP (12%) and HAAFP (10.1%). The former one was a littlehigher than the latter. This results were different from Liang and Singer’sresults ^[[24]].They found that HAAs precursors have a higher aromatic content than THMs precursors,and coagulation generally removed more HAAFP than THMFP.

THMFP increasedafter filtration, which was possibly due to the accumulated pollutants such asalgae, a kind of THMFP, in sand filter. However, HAAFP level continued to declineafter filtration. The overall removal efficiency of HAAFP in conventionaltreatment was higher than THMFP.

Ozonationachieved satisfying removal efficiency for both THMFP (55.1%) and HAAFP(42.4%). BAC also showed good capability of reducing HAAFP (33.9%), but haslimited impact on THMFP. One reason for this may be due to the declinedadsorption capacity of GAC. GAC removes THMs precursors largely by adsorption;however, adsorption capacity of GAC used in this study was reducedsignificantly as stated above. Another reason may be the influence ofpollutants accumulated in the BAC filter such as algae.

In conclusion,ozonation-BAC process was effective to control the disinfection by-productsprecursors.

Influence ofsand filter before and after backwash on THMFP was shown in Table 10. It wasevident that the effect of sand filter on THMFP differed before and afterbackwash. Attention should be paid to the fluctuation of THMFP in the effluentof sand filter.

Table 10.　 Effect ofSand Filter Backwash on THMFP (mg/L)

3.4 Mutagenicity Abatement

The bacteriareversed mutation assay (Ames Test) is most commonly used to evaluate themutagenic properties ^[5]. It was found that 83% mutagenscould be identified by TA98 alone and 93% by both TA98 and TA100. The test usesa strain of Salmonella bacteria that require histidine in the mediumbecause of a defect in a gene necessary for histidine synthesis. Mutagens cancause a further change in this gene that reverses the defect, creatingrevertant bacteria that do not require histidine. To increase the sensitivityof the test, the bacteria also have a defect in their DNA repair machinery thatmakes them especially susceptible to agents that damage DNA. Values ofmutagenicity ratio (MR) were obtained by dividing the number of revertantcolonies in sample plates by that in the blank control. An MR≥2 with a clear dose response wasdefined as mutagenic, whereas an MR < 2 was defined as a lack ofmutagenicity.

Table 11 showsresults of Ames test for the treatment process in this study.

Fromtable 11, it can be seen that raw water was sensitive to TA98, and positivemutagenic effect was obtained with merely 1 liter of raw water; TA100 was muchless sensitive, and no positive effect was found with the largest dosage inthis research. This demonstrates that the mutagenicity of raw water was mostlycaused by frame-shift mutation.

Table 11.　 Results of Ames Test in Treatment Process

Afterpre-ozonation, flocculation and sand filtration, compounds causing frame-shiftmutation did not decrease, but increased instead. Compounds causing base-pairsubstitution also increased considerably. Conventional treatment process led tomutagenicity increase possibly due to the variation of the characteristics oforganic matters and accumulation of pollutants such as algae in the sandfilter.

Ozonation andBAC filtration significantly reduced compounds causing base-pair substitutionand frame-shift mutation, reaching a maximum reduction of 60%.

Ozonation was regardedas an effective method to reduce mutagenicity ^[[25]], however,the effectiveness will depend on the water sample in question, and a specificdosage is required to generate a low level of mutagenic activity for a givensource water ^[[26]].

It is generallybelieved that ozonation does not add to positive effect, but reduce thepositive mutagenicity level, however, it was also reported that ozonationincreased the mutagenicity from negative to positive. The possible reason mightbe associated with characteristics of source water. Some ozonation by-productshave been identified as mutagenic compounds, such as acetaldehyde,formaldehyde, glyoxal, glyoxylic acid and methylglyoxal^[[27]]. Activated carbon was efficientin reducing most of the mutagenic matters, however, some of them such as glyoxalincreased after filtration ^[27]. Therefore, further researches are necessaryabout the effects of ozonation-BAC process on mutagenicity of drinking water.

Mutagenicitymaintained at a stable level after disinfection, demonstrating that disinfectionfollowing ozonation and BAC treatment has little effects on mutagenic activity.

4 CONCLUSIONS

The followingconclusions are drawn from this study:

(a) Ozonation ledto the increase of AOC concentration; however, BAC filtration followingozonation was effective to reduce AOC and maintain biological stability.

(b) Bottom sandlayer beneath BAC bed ensures the effluent of low turbidity and control thetotal bacterial count; Backwash interval of BAC filter was a key factor tocontrol bacteria in effluent. Though the bacteria in BAC filter effluentincreased, they could be effectively inactivated by chlorine disinfection at anormal dosage. This ensures the biological safety of drinking water with littlesurplus chlorine demand.

(c) Ozonationprocess was capable of reducing both THMFP and HAAFP; BAC had considerableremoval efficiency for the HAAFP, but only limited effect on THMFP.

(d) Mutagenicityincreased after raw water treated by conventional process; however, theozonation and BAC filtration process reduced the mutagenic activity remarkably.After disinfection, the mutagenicity maintained a stable level.

Authors

ZHANGJinsong: Chief Engineer, Shenzhen Water Group, Shenzhen, China.

Address: Water Building 1019Shennan Zhong Road, Shenzhen, 518031, China

Tel: 086-755-82137919;Email: zhangjinsong@waterchina.com

DONG Wenyi: Ph. D Candidate, School of Municipal andEnvironmental Engineering, Harbin Institute of Techonlogy, Harbin, China.　

Email: dwy1967@21cn.com

LI Ji:Post doctor, Shenzhen Water Group, Shenzhen, China.

Email: jli@waterchina.com

FANJie: Senior Engineer, Shenzhen Water Group, Shenzhen, China.

　　 Email:fanjie@waterchina.com

REFERENCES

* Corresponding author. Email: zhangjinsong@waterchina.com

[1] Zhou Y, Luo QD. OzoneOxidation – Activated Carbon Process in Zhoujiadu Waterworks. Water &Wasterwater Engineering, 2003, 29(9): 5-8 (in Chinese)

[2] Wang RH. TechnicalReform and Expansion of Nanxingqiao Waterworks in Hangzhou. Water &Wasterwater Engineering, 2003, 29(9): 9-12. (in Chinese)

[3] Huang NL, Liao FJ.Advanced Treatment of Ozone Oxidation-Activated Carbon Process in MeilinWaterworks, Shenzhen City, Water & Wasterwater Engineering, 2003, 29(9):12-15.

[4] Minstry of Health ofChina. Standard Examination Methods for Drinking Water (GB 5749-85). (inChinese)

[5] Standard Methods forthe Examination of Water and Wastewater, 20th ed.; American Public HealthAssociation: Washington, DC, 1998.

[6] Zhou LH, Liu WJ, ZhangLP, et al. A Study on the Methods of Measuring the Assimilable Organic Carbon.Journal of Jimei University (Natural Science). 1999, 4(3): 19-23. (in Chinese)

[7] Dirk Van Der Kooij, A Visser,W A M Hijnen. Determining the concentration of easily assimilable organic carbonin drinking water. J AWWA, 1982,74(10):540-545.

[8] Mark W Lechevallier, NancyE Shaw, Louis A Kaplan. Development of a rapid assimilable organic carbon methodfor water[J]. Appl Environ Micrbiol,1993,5(59):1526-1531.

[9] Dick Van Der Kooij. Assimilableorganic carbon as an indicator of bacterial regrowth. J AWWA 1992:57-65.

[10] Louis A K, Thomas L B, DonaldJ Reasoner. Evaluation and simplification of the Assimilable organic carbon nutrientbioassay for bacterial growth in drinking water. Appl Environ Microbiol, 1993,5(59):1532-1539.

[11] Maron D, Ames B. RevisedMethods for the Salmonella mutagenicity test. Mutation Research, 1983,113:173-215.

[12] Methods for the Determinationof Organic Compounds in Drinking Water, Supplement 1. (EPA/600/4-90020). EnvironmentalMonitoring Systems Laboratory, Office of Research and Development, U.S. EPA:Cincinnati, OH, 1990.

[13] US EPA. AlternativeDisinfectants and Oxidants Guidance Manual. EPA 815-R-99-014. April 1999.

[14] Christian Volk, KimberlyBell, Eva Ibrahim, et al. Impact of enhanced and optimized coagulation onremoval of organic matter and its biodegradable fraction in drinking water.Water Research, 2000,34(12):3247-3257.

[15] Lars J. Hem, Harry Efraimsen.Assimilable organic carbon in molecular weight fractions of natural organicmatter. ter Research, 2001,35(4):1106-1110.

[16] Lechevallier MW, Becker WC,Schorr P, Lee RG. AOC reduction by biologically active filtration. Revue desSciences de l‘Eau, 1992:113-142

[17] van der Kooij D, Hijnen WAM,Kruithof JC. Effects of ozonation, biological filtration and distribution onthe concentration of easily assimilable organic carbon (AOC) in drinking water.Ozone: Science and Engineering, 1989, 11(3):297-311

[18] Nieminski EC, Bradford SM.IMPACT OF OZONE TREATMENT ON SELECTED MICROBIOLOGICAL PARAMETERS. Ozone-Science& Engineering, 1991,13(2): 127-145.

[19] Hu JY, Wang WJ, et al. TheEffect of Water Treatment Processes on the Biological Stability of PotableWater. Water Research, 1999,33(11): 2587-2592.

[20] Van Der Hoek, et al.Minimizing Bromate Formation During Ozone-activated Carbon Treatment. WaterSupply, 1996, 14(3-4):371-380.

[21] Richardson SD, Thruston ATJr, Caughran TV, et al. Identification of New Ozone Disinfection Byproducts inDrinking Water. Environ. Sci. Technol. 1999, 33(19),3368-3377

[22] Zhang SL, Huck PM. Removal ofAOC in Biological Water Treatment Process: A Kinetic Modeling Approach. Wat. Res.1996, 30(5):1195-1207

[23] Urs von Gunten. Ozonation ofdrinking water: Part I. Oxidation kinetics and product formation. WaterResearch, 2003,37():1443–1467.

[24] Liang L, Singer PC. Factorsinfluencing the formation and relative distribution of haloacetic acids andtrihalomethanes in drinking water. Environmental Science & Technology. 2003,37(13):2920-2928.

[25] Kainulainen TK, Tuhkanen TA,Vartiainen TK, et al. Removal of Residual Organics from Drinking-Water byOzonation and Activated Carbon Filtration - A Pilot-Plant Study. Ozone- Science& Engineering, 1995:17 (4): 449-462.

[26] Langlais B, Reckhow DA, BrinkDR. Ozone in Water Treatement – Application and Engineering. Lewis Publishers,Michigan, US, 1991.

[27] Kainulainen TK, Tuhkanen TA,Vartiainen TK, et al. Removal of Residual Organics from Drinking-Water byOzonation and Activated Carbon Filtration - A Pilot-Plant Study. Ozone- Science& Engineering, 1995:17 (4): 449-462.