Two-step chlorination: A new approach to disinfection of a primary sewage effluent
Yu Lia, Mengting Yang a,b,*, Xiangru Zhanga,*, Jingyi Jianga, Jiaqi Liua, Cie Fu Yaua, Nigel J.D. Grahamc, Xiaoyan Lid
aDepartment of Civil and Environmental Engineering, The Hong Kong University of Science and Technology, Hong Kong, China
bCollege of Chemistry and Environmental Engineering, Shenzhen University, Shenzhen 518060, China
cDepartment of Civil and Environmental Engineering, Imperial College London, London, UK
dDepartment of Civil Engineering, The University of Hong Kong, Hong Kong, China
ABSTRACT
Sewage disinfection aims at inactivating pathogenic microorganisms and preventing the transmission of waterborne diseases. Chlorination is extensively applied for disinfecting sewage effluents.The objective of achieving a disinfection goal and reducing disinfectant consumption and operational costs remains a challenge in sewage treatment. In this study, we have demonstrated that, for the same chlorine dosage, a two-step addition of chlorine(two-step chlorination) was significantly more efficient in disinfecting a primary sewage effluent than a one-step addition of chlorine (one-step chlorination), and shown how the two-step chlorination wasoptimized with respect to time interval and dosage ratio.Two-step chlorination ofthe sewage effluent attainedits highest disinfection efficiency at a time interval of 19s and a dosage ratio of 5:1. Compared to one-step chlorination, two-step chlorination enhanced the disinfection efficiency by up to 0.81- or even 1.02-log for two different chlorine doses and contact times.An empirical relationship involving disinfection efficiency, time interval and dosage ratio was obtained by best fitting. Mechanisms (including a higher overall Ct value, an intensive synergistic effect, and a shorter recovery time) were proposed for the higher disinfection efficiency of two-step chlorination in the sewage effluent disinfection. Annual chlorine consumption costsin one-step and two-step chlorination of the primary sewage effluent were estimated. Compared to one-step chlorination, two-step chlorination reduced the cost by up to 16.7%.
Keywords: Disinfection; Two-step chlorination; Disinfection efficiency; Synergistic effect; DBPs.
1. Introduction
Sewage disinfection aims at providing protection for humans from exposure to pathogenic waterborne microorganisms (Jacangelo and Trussell, 2002).Chlorine, which indicates free chlorine in this study, is extensively used as a disinfectant inmunicipal sewage treatment plantsdue to its effectiveness, low cost and ease of application (Lee and von Gunten, 2010; Drinan and Spellman, 2012; Shannon et al., 2008).In addition to inactivating pathogenic microorganisms, chlorine reacts with various reducing inorganic ions (e.g., sulfide andferrous), ammoniaand effluent organic matter present in sewageeffluents. Chloramines are generated from the reaction of chlorine with ammonia. At low Cl2:Nmass ratios (< 5:1),the dominating species of chloramines is monochloramine (White, 2011).
Various factors can influence disinfection efficiency, includingdisinfectant type,disinfectantconcentration, contact time, reaction condition, and microorganismspecies. Generally, disinfection efficiency can be expressed as:,where k is the inactivation rate constant ofa disinfectant,C is disinfectant concentration at time t,and Nt/No is the ratio of remaining microbial concentration at time t to the initial microbial concentration (Haas and Karra, 1984; Crittenden et al., 2012).The disinfectant concentration and contact time can be quantified together as the Ctvalue, which is used to determine the compliance with water regulations for inactivation of microorganisms(U.S. EPA, 1991).Fecal coliform bacteria live in the intestinal tracts of humans and other warm-blooded animals. Theyare the most common microbial contaminants in natural waters and wastewaters.Escherichia coli, a main species in the fecal coliform group, iscommonly used as an indicator organism to assess the effectiveness of water disinfection (Hu et al., 2005; Quek et al., 2006;Janjaroen et al., 2013).
For a particular disinfection unit in a sewage treatment plant, increasing the chlorine dosemay increase the disinfection efficiency, but this will consume more chlorine and increase the operational cost, as well as cause the formation of more halogenated disinfection byproducts (DBPs)in the chlorinated sewage effluent(Rebhun et al., 1997; Fabbricino and Korshin, 2005; Krasner et al., 2009; Sun et al., 2009; Sedlak and von Gunten, 2011; Tang et al., 2012; Ding et al., 2013;Gong and Zhang, 2013; Richardson andTernes, 2014; Gong and Zhang, 2015; Gong et al., 2016; Cai et al., 2016)and possible adverse effects to theecosystem of the receiving water body(Watson et al., 2012; Yang and Zhang, 2013, 2014; Liu and Zhang, 2014).Thus, it is of great importance to develop and optimize the sewage disinfection processwitheco-friendly and cost-effective features.
Recently, Verma et al.(2013) reportedthat, for the same chlorine dose,a two-step addition of chlorine (two-step chlorination)showeda greater disinfection efficiency than a one-step addition of chlorine (one-step chlorination). The result indicated that two-step chlorination hadthe potential to reduce the chlorine dose for the same level of E. coli removal. However, this study of step-wise chlorination was still at a preliminary stage, as only one condition was reported (i.e., a chlorine dose of 2.5+2.5 mg/L as Cl2was applied witha 5-min time interval, for a total contact time of 20 min) and only one effluent sample from a secondary sewage treatment plantwascollected and tested.Two otherstudies (Graham and Hong, 2001; HKDSD, 2015a) also reported similar tests withpreliminary results.However, it has remained unclear whether two-step chlorinationcan also perform better for a primary sewage effluent than one-step chlorinationand whether the better performance of two-step chlorination was reproducible.Additionally, it has remained unclear what the mechanismsarefor the potentially higherdisinfection efficiency of two-stepchlorination. Such mechanisms should be of great scientific value.
Accordingly, this study aimed to compare the disinfection efficiencies of two-step chlorination and the commonly used one-step chlorination of a primary sewage effluent. Furthermore, the two-step chlorination process of the primary sewage effluent was optimized in terms of time interval (the time period between the first and second steps of chlorination) and dosage ratio (the mass ratio of chlorine dosed at the first step to the second step). The mechanisms for the observed greater disinfection efficiency of two-step chlorination were also explored.
2. Materials and methods
2.1. Chemicals and reagents
All the chemical solutions were prepared from chemicals of reagent grade or higher. Ultrapure water (18.2 MΩ·cm) was provided by a Cascada I Laboratory Water Purification System (PALL). A NaOClstock solution was prepared by diluting a reagent grade sodium hypochlorite solution (Allied Signal). A preformed monochloramine stock solution was freshly prepared by mixing of the NaOCl stock solution and an ammonium chloride solution at a chlorine to ammonium mole ratio of 0.8:1. Both the NaOCl stock solution and the preformed monochloramine stock solution were standardized by the DPD colorimetric method(APHA et al., 2012).Thioacetamideand sodium thiosulfate were purchased from Sigma–Aldrich.
2.2.Collection and characterization of effluent samples from a primary sewage treatment plant
A total of19undisinfected effluent samples were collected from a primary sewage treatment plantover a 9-month period from May 7, 2015 to February 15, 2016, whereeach sample was collected ona given day.The sewage treatment plant has a treatment capacity of 2.0×106 m3/d and is one of the largest primary sewage treatment plants in the world.Each collected sample was transferred immediately to the laboratory in an ice cooler to minimize changes in its constituents. A major portion of the sample was used immediately for conducting disinfection, and the remaining sample was used for characterization. Dissolved organic carbon (DOC) and ammoniawere measured with a total organic carbon analyzer (Shimadzu, Japan) and a flow injection analysis system (8500 Series, Lachat, USA), respectively.Chemical oxygen demand (COD) and total suspended solids (TSS) wereobtained from the Hong Kong Drainage Services Department (HKDSD, 2015b; 2016), and they were determined following the Standard Methods(APHA et al., 2012). The ultraviolet (UV) absorbance was measured at 254 nm with a 1-cm quartz cuvetteby using a spectrophotometer (Lambda 25, Perkin Elmer, USA).
2.3.Disinfection of the primary sewage effluent
Initially, a preliminary test was conducted to choose appropriate disinfection conditions which were capable of meeting the sewage effluent discharge standard. Aliquots of a sewage effluent sample were chlorinated by dosing 1.0‒6.0 mg/L NaOClas Cl2. After a 15- or 30-min contact time, each aliquot was dechlorinated with 105% of the requisite stoichiometric amount of 0.10 M Na2S2O3. Two representative chlorination scenarios (i.e., dosing 4.0 mg/L of NaOCl as Cl2 for a 30-min contact time and dosing 6.0 mg/L of NaOCl as Cl2 for a 15-min contact time) for one-step chlorination were selected for further comparison as they could fulfill the disinfection goal (Fig.S1andSupplementary Information (SI)).
To compare the disinfection efficiencies of two-step chlorination and one-step chlorination, aliquots ofanundisinfectedprimary sewage effluent sample were chlorinated.For one-step chlorination, 100-mL aliquots ofthe sewage effluent sample were dosed with4.0 mg/L of NaOCl as Cl2 for a 30-min contact time; for two-step chlorination, 100-mL aliquots ofthe sewage effluent sample were chlorinated with the same total chlorine dose (4.0 mg/L as Cl2), but the dose was split into two portionsat different ratios (first step: second step = 1:1, 3:1, 5:1 or 8:1).The two portions of each ratio were dosed at different time intervals (first step at 0 s, and second step at 5, 10, 19, 38, 75, 150, 300 or 600 s). It should be mentioned that different dosage ratios and time intervals were examined to study their effects on the disinfection efficiency.After disinfection with thorough mixingfor a total 30-min contact time, the chlorine residual in each aliquot was dechlorinated with 105% of the requisite stoichiometric amount of 0.10 MNa2S2O3.To further compare the disinfection efficiencies of two-step chlorination and one-step chlorination, the experiments above were repeated except that a total chlorine dose of 6.0 mg/L as Cl2and a total contact time of 15 min were used.
To study the mechanism of two-step chlorination, the disinfection efficiencies of one-step chlorination and one-step chloramination of the primary sewage effluent were compared. Aliquots of an undisinfected primary sewage effluent sample were disinfected by dosing 4.0 mg/L chlorine or monochloramine as Cl2. After disinfection with thorough mixingfor a 30-min contact time, the chlorine residual in each aliquot was dechlorinated with 105% of the requisite stoichiometric amount of 0.10 M Na2S2O3.
2.4.Determinationof disinfection efficiency
A membrane filter procedure(APHA et al., 2012) was used to quantify the E. coliconcentrations before and after disinfection. The detailed procedure for the determination of disinfection efficiency is described in the SI.
2.5. Determination of residual disinfectants and the Ct values
Owing to the high concentration of ammonia (19.3‒27.4 mg/L NH4+ as N) in the primary sewage effluent, chlorine reacted rapidlywith ammonia to form monochloramine. In the theory of chemical kinetics, the reaction of chlorine and ammonia to form monochloramine is very fast, with a rate constant of 5.1×106 M−1s−1 (Morris, 1967). In practice, however, the overall reaction rate of this reaction in a reactor is controlled by mixing which transportsthe reactants to each other, and thus chlorine might exist for a short time before converted to monochloramine. Wolfe et al. (1984) indicated that for sub-optimal treatment conditions, such as insufficient mixing,chlorine may co-exist with monochloramine for a short time (usually for a few minutes) prior to complete monochloramine formation. Accordingly, the chlorine concentration obtained from a thermodynamics or kinetics calculation,or measured bya stopped-flow system,might notbe representativefor the reaction conditionsin the disinfection efficiency test. To accurately evaluate the Ct values, the concentrations of residual chlorine and monochloramine were monitored with time. Because a high concentration of monochloramine (> 0.5 mg/L as Cl2) may break into chlorine and seriously interfere with the detection of chlorine, a modified procedure was usedin this study for accurately measuring chlorine, involvingthe addition of athioacetamide solution (APHA et al., 2012). Adding thioacetamide can completelyterminate further reaction with monochloramine in the chlorine test (APHA et al., 2012).
The residual chlorine and monochloramine in two pairs of chlorinated primary effluent samples were monitored with time. Each pair included a one-step chlorination sample and atwo-step chlorination samplewitha 19-s time interval and a 5:1 dosage ratio, at which the disinfection efficiency of the two-step chlorination was relatively high as shown in the Results and Discussion. The first pair was with a total chlorine dose of 4.0 mg/Las Cl2 anda total contact time of 30 min, and the second pair was with a total chlorine dose of 6.0 mg/Las Cl2 anda total contact time of 15 min.To initiate the residual chlorine measurement, a series of 100-mL aliquots of an unchlorinated primary sewage effluent sample, a phosphate buffer‒DPD (v/v, 1:1) mixed solution, and a0.25% thioacetamide solution (w/w)were freshly prepared prior to the test. Then, a 100-mL aliquot of the sewage effluent sample was chlorinated by dosing a given dose of NaOCl (one-step chlorination) or by dosing two portions of the same dose of NaOCl with the given time interval (two-step chlorination). To simulate an extreme case of the insufficient mixing condition, mixing was not applied to all the sewage effluent samples after each addition of chlorine. After a given contact time (controlled by a timer), 10mL of the buffer‒DPD mixed solution and 0.5mL of the 0.25% thioacetamide solution were successively added to the aliquot with thorough mixing. The absorbance of the produced pink color was read immediately by a spectrophotometer at 515 nm and the chlorine concentration was quantified by a calibration curve as shown in SI Fig. S2a. To determine residual monochloramine concentrations, a series of 100-mL aliquots of an unchlorinated primary sewage effluent sample, a phosphate buffer‒DPD (v/v, 1:1) mixed solution and a 5.0 g/L potassium iodide solution were also freshly prepared. A 100-mL aliquot of the sewage effluent sample was chlorinated by dosing a given dose of NaOCl or by dosing two portions of the same dose of NaOCl with the given time interval. After a given contact time, 10mL of the buffer‒DPD mixed solution and 0.20 mL of a potassium iodide solution were successively added to the aliquot. The absorbance of the produced pink color was read by a spectrophotometer at 515 nm and thetotal chlorineconcentration was quantified by a calibration curve as shown in SI Fig. S2b. The monochloramine concentration was calculated by deducting the chlorine concentration from the total residual chlorine concentration.
Two pairs of curves of the residual disinfectant versus contact time, one for chlorine and the other for monochloramine, were obtained. The Ct value for each disinfectant was obtained by integrating the area below the curve.
2.6.Preparation of aneffluent-isolated E. coli stock solution and evaluation of the synergistic effect of chlorine and monochloramine
Due to the presence of ammonia in the primary sewage effluent, effluent-isolated E.coli bacteria were selected for evaluating the synergy of chlorine and monochloramine. The E. coli species was isolated from thesame primary sewage effluentby separating an E. coli colony on the m-FC medium. Then, it was transferred to a typtone soya broth (Oxoid, UK) and incubated with aeration at 37ºCfor 18 h, and sub-cultured for four times to reach a stationary phase. Bacterial cells were harvested from the broth by repeating centrifugation at 1200×ɡfor 10 min at 4ºC and washing with a phosphate-buffered saline solution (PBS, 150 mM at pH 7.4). The concentration of the E. coli stock solution (109colony forming units(CFU) per 100 mL) was determined by measuring the optical density at 600 nm (Zhao et al., 2014).
A series of aliquots of E. coli working solution (106 CFU/100 mL) were prepared by adding0.10 mL of the E. coli stock solution to 100 mL PBS. Then, the working solutions were individually or sequentially disinfected with 0.20 mg/L chlorine for a contact time of 1, 2, 4, 6, 8 or 10 s, and 4.0 mg/L monochloramine for a contact time of 1.0 min, to simulate a low level of chlorine and a high level of monochloramine during chlorination of the primary sewage effluent. For the sequential disinfection, after a given contact time, the remaining chlorine was converted to monochloramine by adding anappropriate amount of ammonium chloride based on the residual chlorine level. Additional preformed monochloramine was added to give the desired 4.0 mg/L monochloramine. The disinfectant concentrations and contact times were selected based on the values achievable in practice yet capable of producing a measurable inactivation. After a given contact time, the sample was quenched with 105% of the requisite stoichiometric amount of 0.01 or 0.10 M Na2S2O3. The same membrane filter procedure(APHA et al., 2012) was used to quantify the E. coliconcentrations before and after disinfection (SI), except that measurement was conducted in tetraplicate for each sample. The synergy was quantified per the following equation (Gyürék et al., 1997).
(1)
where is the log reduction of the E. coli survival ratio aftera sequential disinfection; and are the log reduction of E. coli survival ratios after chlorination individually and chloramination individually, respectively.
3. Results and discussion
3.1. Characteristics of the primary sewage effluent
Undisinfected effluent samples from the primary sewage treatment plant were collected over a span of 9months covering different seasons. Table 1 lists the characteristics of the samples, which showed some variability of the effluent: pH 7.05±0.10, NH4+ 23.9±2.7 mg/L as N, DOC 23.7±3.9 mg/L as C, UV254 0.21±0.05 cm−1, COD 41.8±10.5 mg/L, and TSS 40.0±4.3mg/L. SI Fig. S3presents the mean value and the range of each parameter. The E. coliconcentration in each undisinfected sewage effluent sample was also measured, and it ranged from 0.9×106 to 22.7×106 CFU/100mL with a geometric mean of 11.0×106 CFU/100mL.
3.2. Comparison of disinfection efficienciesof two-step chlorination and one-step chlorination
Fig. 1 shows the comparative disinfection efficiencies of two-step and one-step chlorination with a total chlorine dose of 4.0 mg/L as Cl2 and a total contact time of 30 min.This figure includes 21 charts for the primary effluent samples collected on different days and for the two-step chlorination with different time intervals and different dosage ratios. Although the effluent characteristics varied with the sampling date, the disinfection efficienciesof the two-step chlorination in nearly all the charts were significantly higher than thoseof the one-step chlorination.Compared to one-step chlorination, the two-step chlorinationelevated the disinfection efficiency (E. coli reduction) up to a maximum of 0.81-log(Fig. 1g), corresponding toa 19-s time interval and a 5:1 dosage ratio. The only exception, where theone-step chlorination was superior to the two-step chlorination,was for a time interval of 600 s (the longest time interval examined in this study), indicating that too long a time intervalwas undesirable and should be avoided (Fig. 1p and 1u).
It should be noted that, even for the same disinfection setting (with a chlorine dose of 4.0 mg/L as Cl2 and a contact time of 30 min), the disinfection efficiency of one-step chlorination varied with the sampling date. Gehret al. (2003) also observed that, for the same sewage effluent and the same disinfection setting, the disinfection efficiency varied significantly with the sampling date. For better comparison of the disinfection efficiencies of two-step and one-step chlorination, the disinfection efficiency of two-step chlorination of the effluent sample collected on a given day was normalizedby adding the difference between the disinfection efficiency of one-step chlorination on that day and the average disinfection efficiency of one-step chlorination on all days. With the normalized data, a relationship of the disinfection efficiency of two-step chlorination with time interval and dosage ratiowas obtained by fitting the data with the software Matlab, as shown in Equation(2). The regression coefficient of this equationwas0.95, indicating that the disinfection efficiency of two-step chlorination could be well predicted with time interval and dosage ratio, even though the physical meaning of thisempirical equation remains unclear.