Chlorine off-flavors of tap water have caused dissatisfaction and distrust from some consumers, placing pressure on operators concerning water disinfection. Evaluating practical approaches for eliminating chlorinous off-flavors by consumers at point-of-use while avoiding production of toxic byproducts is a practical concern. Three recognized dechlorination methods: ultraviolet (UV) irradiation, ascorbic acid (AA) and hydrogen peroxide (HP), were evaluated for chlorinated and chloraminated waters. AA is the most efficient for removing free chlorine and chloramine from water samples. Three new chlorine-containing compounds were detected and identified from the reaction between AA and chlorine. High doses of UV irradiation at 254 nm virtually eliminated chlorine. HP could effectively remove free chlorine but was not effective for chloramine elimination. AA shows promise as a practical household dechlorination agent. However, to assure consumers about drinking water safety, further investigation is needed to evaluate any potential toxicity concerns for reaction products in treated water.

## INTRODUCTION

Chlorine is cost-effective to inactivate various types of pathogens and its capacity to retain a residual assists in reducing microorganism regeneration in distribution pipes while also providing a real-time measure for contaminant chlorine demand (Hrudey & Walker 2005). Chlorination of tap water has become the most common treatment to manage microbial contamination in drinking water (Puget et al. 2010) and it is a regulatory requirement in many countries to specify levels of residual chlorine to be maintained in the distribution system. However, the perception of chlorine taste and odors at the tap, because of residual chlorine or formation of odorous by-products, has generated dissatisfaction and complaints from drinking water consumers (Piriou et al. 2004; Hrudey & Hrudey 2007). Complaints about chlorine off-flavor in tap water provided the third most commonly reported problem in the USA (Suffet et al. 1996) and a survey in Quebec, Canada, found over 40% of the interviewed consumers had reported a chlorine off-flavor in their tap water (Levallois et al. 1999). Unfortunately, consumer complaints have influenced some drinking water operators to compromise their chlorine disinfection sufficiently that outbreaks of waterborne disease have occurred because of inadequate disinfection (Hrudey & Hrudey 2007). Despite the demonstrable need for disinfection of drinking water, opposition to chlorination among consumers occurs frequently (Crawford 2013), at least in part because chlorine off-flavors allow consumers to detect the use of chlorination. Consequently, the question of how to effectively eliminate chlorine off flavors from drinking water for domestic use is a matter relevant to assuring the meeting of a major goal stated for the international consensus Bonn Charter to provide: ‘Good safe drinking water that has the trust of consumers’ (IWA 2004).

Quenching the residual chlorine at the point-of-use is a possible option to deal with this aesthetic problem. To date, dechlorination has been practiced for water sample collection to prevent further reactions with chlorine (e.g. for chlorination disinfection by-products) and to reduce eco-toxicity to fish prior to discharges of chlorinated wastewater (Bedner et al. 2004). Sulfur-based compounds, including sodium thiosulfate, sodium bisulfate and sodium bisulfite, had been commonly used as reducing agents to remove residual chlorine from treated water. Adding sulfur compounds to tap water destined for human consumption is not viable because it may generate sulfur-based off-flavors. Several alternative dechlorination agents have been tested (Bedner et al. 2004; MacCrehan et al. 2005). Ascorbic acid (AA) was found to be the most reactive of the sulfur-free agents at fivefold molar excess and it has been shown that AA reacts with chlorine to generate dehydroascorbate and chloride (Ganesh et al. 2006). However, when mixing an equal mass of AA and chlorine, the dechlorination is slower and incomplete (Hermant & Basu 2013).

Hydrogen peroxide (HP) showed promising dechlorination ability but the dosage was relatively large (Hermant & Basu 2013). The products from the reaction between HP and chlorine have been identified as oxygen and chloride (Ganesh et al. 2006), but the presence of HP may influence the formation of certain disinfection byproducts when coupled with other treatments (Dotson et al. 2010).

Ultraviolet (UV) irradiation has been shown to be effective for destruction of chlorine and chloramine and the photodecay efficiency increased with the increase of UV dose (Watts & Linden 2007; Li & Blatchley 2009). Most experiments have been developed under controlled conditions and focused on only one or two water samples to investigate the dechlorination process. For drinking water, in order to assure water quality and consumer satisfaction, consideration must be given to obtaining comprehensive information on both laboratory-prepared and authentic tap waters.

Home water purifiers based on activated carbon are widely used in households and are known to be effective for removal of residual chlorine, organic off-flavors and other trace organic compounds. However, carbon cartridges can promote microbial growth of opportunistic pathogens by developing biofilms (Chaidez & Gerba 2004) and carbon cartridges must be routinely replaced. Other home water purifiers (e.g. distillation) will also remove residual chlorine, but they represent a substantial capital investment and require routine maintenance.

The effectiveness of three dechlorination methods that may be implemented by consumers on an as-needed basis for tap water destined for ingestion was investigated. These methods were evaluated on chlorinated and chloraminated tap water and analytical grade water. UV irradiation was regarded as a potentially consumer-attractive, chemical additive-free dechlorination method. AA and HP were selected based on their known high dechlorination efficiency and potential for low environmental impact.

## MATERIALS AND METHODS

### Reagents and chemicals

The water used for the preparation of reference test solutions was Optima® LC/MS grade and purchased from Fisher Scientific (Fair Lawn, NJ). l-Ascorbic acid (>99.0%), sodium hypochlorite solution (available chlorine 10–15%) and hydrogen peroxide (HP) were obtained from Sigma-Aldrich (Oakville, ON). The stock solutions of chlorine (28 mmol L−1 or 2 g L−1) and chloramine (200 mmol L−1 or 10 g L−1) were freshly prepared daily.

### Water samples

Commercial Optima® water and Edmonton tap water were used as synthetic water and authentic water, respectively. Optima® water was commercially obtained and stored in a 4 L amber glass bottle. Edmonton tap water was collected from an ordinary household faucet. Before sampling, the faucet was run for 10 minutes and then samples were carefully collected to avoid air bubbles. Tap water samples were also stored in a 4 L amber glass bottle at 4 °C.

To mimic the drinking water disinfection process, water samples were prepared in four groups, including chlorinated Optima® water, chlorinated tap water, chloraminated Optima® water, and chloraminated tap water. According to the Guidelines for Canadian Drinking Water Quality, concentrations of residual free chlorine in most Canadian drinking water distribution systems are up to 2.0 mg L−1 (Health Canada 2009), thus 2 mg L−1 was selected as the initial chlorine concentration although most water utilities would have lower free chlorine residuals in distribution systems. The detailed information is described in the supporting information.

Optima® water spiked with 2 mg L−1 free chlorine (nominal value) and 2 mg L−1 chloramine (nominal value) were employed as chlorinated/chloraminated synthetic water. The actual concentrations of free chlorine were determined immediately after input of chlorine at the beginning of each batch of experiments, and this actual value was regarded as the initial concentration of each water sample. Edmonton (where this research was conducted) tap water is chloraminated. Concentrations of total chlorine sampled at the tap are around 2 mg L−1 and free chlorine is below the detection limit (0.1 mg L−1), which indicates the chloramine concentration in the Edmonton tap water is also around 2 mg L−1. Actual concentrations of chloramine in the tap water were also determined before the treatments.

A pre-experiment was conducted to produce chlorinated tap water involving serial additions of sodium hypochlorite into the Edmonton tap water, which is chloraminated, using the chlorination break point (Chen & Jensen 2001). Dosage of 4.5 mg L−1 Cl2 was selected and the final concentration of free chlorine in the tap water could be maintained at around 2 mg L−1 after 6 h reaction time. The descriptions of water samples are listed in Table 1.

Table 1

Descriptions of water samples

Water samplesDescriptions
Chlorinated Optima® water Adding sodium hypochlorite into Optima® water, final concentration of free chlorine at around 2 mg L−1, pH = 7.0–7.5.
Chlorinated tap water Adding sodium hypochlorite into tap water according to break point theory, concentration of free chlorine maintained at around 2 mg L−1 after 6 h reaction time, pH = 7.0–7.5.
Chloraminated Optima® water Adding chloramine into Optima® water, final concentration of chloramine at about 2 mg L−1, pH = 7.0–7.5, chloramine was freshly prepared before the experiment.
Chloraminated tap water Edmonton tap water, chloraminated, concentration of total chlorine is around 2 mg L−1, free chlorine is below the detection limit (0.1 mg L−1), pH = 7.0–7.5.
Water samplesDescriptions
Chlorinated Optima® water Adding sodium hypochlorite into Optima® water, final concentration of free chlorine at around 2 mg L−1, pH = 7.0–7.5.
Chlorinated tap water Adding sodium hypochlorite into tap water according to break point theory, concentration of free chlorine maintained at around 2 mg L−1 after 6 h reaction time, pH = 7.0–7.5.
Chloraminated Optima® water Adding chloramine into Optima® water, final concentration of chloramine at about 2 mg L−1, pH = 7.0–7.5, chloramine was freshly prepared before the experiment.
Chloraminated tap water Edmonton tap water, chloraminated, concentration of total chlorine is around 2 mg L−1, free chlorine is below the detection limit (0.1 mg L−1), pH = 7.0–7.5.

### Dechlorination methods

Comparatively high doses of low power UV light were used to determine the removal efficiency of free chlorine and chloramine in both Optima® water and tap water. A collimated beam UV (Figure S1, available online at http://www.iwaponline.com/ws/015/088.pdf) device with low power lamp was applied. The lamp system consisted of two 8 W UVC germicidal lamps (Luzchem LZC-UVC-01). UV light irradiated a 200 mL water sample, which was stored in an uncovered cylindrical crystallization dish with 11.5 cm diameter × 7.6 cm depth. The distance from the lamp to the water surface was 17.1 cm and water sample depth was 1.9 cm. The UV irradiance at the surface of the water was measured by a UVX model digital radiometer (UVP). Corrections were performed and the average UV irradiance was calculated following the procedures of Bolton & Linden (2003). UV dose (mJ cm−2) was calculated as the product of UV irradiance (mW cm−2) and exposure time (s). In this study, the target UV doses were set at 100, 200, 500, 1,000, 1,500 mJ cm−2 and the exposure periods were determined in advance depending on different water samples. Water samples were analyzed immediately after UV irradiation.

#### Ascorbic acid (AA)

The solution of AA was freshly prepared before use with a concentration of 57 mmol L−1 (10 g L−1). AA solutions were added into 200 mL water samples with molar ratios of AA to free chlorine/chloramine of 1:1 and 2:1. The reaction times were set at 10 s, 30 s, 1 min, 2 min and 5 min to examine the influence of treatment duration on the removal efficiency of residual chlorine. Water samples were analyzed immediately after treatment.

#### Hydrogen peroxide

The solution of HP was also freshly prepared before use with a concentration of 100 mmol L−1 (3.4 g L−1). HP was mixed with 200 mL of oxidant solution at a series of designated molar ratios and the samples were collected at a designated time. Water samples were analyzed immediately.

### Analytical methods

The pH of water samples was measured by Accumet model 15 pH meter. Concentrations of free chlorine and chloramine were measured by a chlorine amperometric titrator (AutocatTM 9000, HACH). The operation and calculation process is provided in the supporting information. The reaction products between AA and chlorine were identified by triple TOF mass spectrometer (TripleTOF® 5600, AB SCIEX, Concord, ON, Canada) which can measure ion mass to 0.0001 m/z, making ion elemental composition identification much more confident than is possible with low resolution mass spectrometry. The conditions of TOF mass spectrometer systems were described in the supporting information. The mixture, with a molar ratio of AA to Cl of 2:1, was freshly prepared and the mass spectra of mixtures were obtained after reaction for 10 seconds and 12 hours. Peakview® (version: 1.2.0.3, AB Sciex) was employed to process the data and track the fingerprint of chlorine in different solutions.

### Quality assurance/quality control

The control group consisted of the same water samples with the same initial chlorine concentrations, but was stored in the dark at room temperature. The samples for analysis were collected at the designated time and concentrations of free chlorine or chloramine were measured by the same device. All the glassware and containers used in this study were soaked in a dilute bleach solution (1 mL commercial bleach solution to 1 L of water) for 2 h and rinsed thoroughly at least three times with Optima® water. The pre-cleaned glass and containers were baked at 180 °C for 1 h prior to use. All experiments were performed in triplicate and the mean values and standard deviations were reported.

## RESULTS AND DISCUSSION

### Chlorine elimination by UV irradiation

Figure 1 represents the variation of residual chlorine contents during the UV treatment of four sets of samples (A) chlorinated Optima® water; (B) chlorinated tap water; (C) chloraminated Optima® water; and (D) chloraminated tap water. The results demonstrate that UV254 irradiation can reduce the free chlorine and chloramines in the tested samples. In general, approximately 50% of the initial chlorine content was removed when UV doses reached 500 mJ cm−2. When the samples were treated with a very high UV254 dose of 1,500 mJ cm−2, free chlorine was removed to undetectable level. However, under the same UV treatment conditions, more than 15% of the initial chloramine was detected in the chloraminated tap water samples. UV irradiation is apparently more effective for removing free chlorine than chloramine. These UV doses are impractically high for use in a full-scale water treatment plant, but for the purposes of in-house consumer use on an as-needed basis, such high UV doses may be feasible.

Figure 1

Effect of UV light irradiation on chlorine elimination in different water samples (n = 3). The small figure represents nonlinear fits of eq. 1 to the data. (a) Chlorinated Optima® water; (b) Chlorinated tap water; (c) Chloraminated Optima® water; (d) Chloraminated tap water.

Figure 1

Effect of UV light irradiation on chlorine elimination in different water samples (n = 3). The small figure represents nonlinear fits of eq. 1 to the data. (a) Chlorinated Optima® water; (b) Chlorinated tap water; (c) Chloraminated Optima® water; (d) Chloraminated tap water.

Quantum yield is a measure of the effect of absorbed light to the target compound so it can represent efficiency of photodegradation (Skillman 2008). The values of quantum yields (calculation process described in the supporting information) for UV-induced removal of free chlorine and chloramine are summarized in Table 2. The quantum yields of free chlorine and chloramine in Optima® water were 0.96 ± 0.13 Es mol−1 and 0.55 ± 0.10 Es mol−1, respectively. These values were in accordance with the quantum yields for similar photoreactions reported in previous studies (Feng et al. 2007; Li & Blatchley 2009). However, the quantum yields obtained from tap water (0.52 and 0.35 Es mol−1) were much lower than those from Optima® water (0.96 and 0.55 Es mol−1), which implied that the matrix in the tap water could influence the efficiency of photodegradation of free chlorine and chloramine, making higher UV doses likely necessary for a practical application with authentic conditions. In both water samples, the quantum yields of free chlorine were higher than those of chloramine which further confirmed that free chlorine was easier to eliminate from water by UV irradiation than were chloramines.

Table 2

Quantum yields for photodecay of free chlorine and chloramine in different water samples

Water samplesQuantum yield (Es mol−1)
Chlorinated Optima® water 0.96 ± 0.13
Chlorinated tap water 0.52 ± 0.11
Chloraminated Optima® water 0.55 ± 0.10
Chloraminated tap water 0.35 ± 0.10
Water samplesQuantum yield (Es mol−1)
Chlorinated Optima® water 0.96 ± 0.13
Chlorinated tap water 0.52 ± 0.11
Chloraminated Optima® water 0.55 ± 0.10
Chloraminated tap water 0.35 ± 0.10

The photodecay of free chlorine by UV irradiation could lead to the generation of chloride, chlorite, chlorate ions and several intermediate products (Vogt & Schindler 1991; De Laat et al. 2010). Li & Blatchley (2009) reported that the monochloramine could decompose to nitrate, nitrite, nitrous oxide and ammonia under UV254 irradiation. De Laat et al. (2010) reported nitrite formed at 0.37 mol/mol of NH2Cl decomposed and nitrate formed at 0.073 mol/mol of NH2Cl decomposed. Although the amounts of these photodegradation products was small, the potential health effects on long term consumption of household drinking water will need to be considered on a comparable basis in relation to any of the dechlorination options.

### Chlorine elimination by HP

The effects of reaction times on residual chlorine elimination were first tested. For the chlorinated water, HP is efficient for free chlorine removal, as over 85% of initial free chlorine can be removed within 60 s, and the removal efficiency could reach around 95% in both of the water samples when reaction time increased to 120 and 300 s (Figure S3(a), available online at http://www.iwaponline.com/ws/015/088.pdf). The rate of dechlorination in the chlorinated Optima® water is faster (removal efficiency is about 93% at 60 s reaction time) than that in the chlorinated tap water (removal efficiency is about 86% at 60 s reaction time), which means the reactions in the tap water may be more complex. For the chloraminated water, HP appears to be not an effective agent for removing chloramine because about 35% of initial chloramine can still be detected in the solutions after 10 min reaction (Figure S3(b), available online at http://www.iwaponline.com/ws/015/088.pdf). Increasing reaction time has little effect on chloramine removal in the chloraminated Optima® water but in the chloraminated tap water, the removal efficiency increased from 37 to 62% when the reaction time increased from 60 to 600 s. Similar to the chlorinated water, the dechlorination rate is slower in the tap water, which further suggested that the dechlorination reactions may involve several steps when other substances are present in the tap water. The reaction between HP and hypochlorite could be fast because of the occurrence of radical reactions and the formation of radical species (Castagna et al. 2008), which may explain why the reaction between HP and free chlorine is faster than the reaction between HP and chloramine. In order to achieve the maximum chlorine removal efficiency in the four water samples, 120 and 600 s were chosen as the reaction times for the HP dosage test in the chlorinated and chloraminated water, respectively.

As shown in Figure 2, the removal efficiency of free chlorine and chloramine could be improved with increasing HP dosage in a certain range. For the chlorinated water, the removal efficiency could achieve maximum (about 92%) when the molar ratio of HP to free chlorine was 1.25:1, and this value showed no significant difference when adding HP at twofold molar excess (Figure 2(a)). Hermant & Basu (2013) similarly reported that chlorine could be effectively eliminated at an equal stoichiometric ratio (on a weight basis) of HP to chlorine. For the chloraminated water, the required dosage of HP was much higher as the maximum removal could only be detected when the molar ratio of HP to chloramine was 4:1, but the removal efficiency was not substantially different than achieved with the molar ratio of HP to free chlorine of 3:1 (Figure 2(b)).

Figure 2

Effect of addition of HP at different ratios on chlorine elimination in different water samples (n = 3). The reaction times are 120 seconds for free chlorine ((a) and (b)) and 600 seconds for chloramine ((c) and (d)). (a) Chlorinated Optima® water; (b) Chlorinated tap water; (c) Chloraminated Optima® water; (d) Chloraminated tap water.

Figure 2

Effect of addition of HP at different ratios on chlorine elimination in different water samples (n = 3). The reaction times are 120 seconds for free chlorine ((a) and (b)) and 600 seconds for chloramine ((c) and (d)). (a) Chlorinated Optima® water; (b) Chlorinated tap water; (c) Chloraminated Optima® water; (d) Chloraminated tap water.

### Chlorine elimination by AA

In this study, removal efficiencies of residual chlorine for two different dosages of AA and several reaction durations were tested and reaction products were identified. As shown in Figure 3, AA is a very efficient dechlorination agent with the reactions completed within 10 seconds in all the tested water samples. Removal efficiency was enhanced with increasing AA dosage. When the molar concentration of AA was equal to the chlorine in water samples, over 80% of initial free chlorine could be removed from the chlorinated water samples (Figures 3(a) and 3(b)), and about 75–90% of initial chloramine could be eliminated from the chloraminated water samples (Figures 3(c) and 3(d)). When AA dosage doubled (molar ratio 2:1), almost all the free chlorine and chloramine were eliminated rapidly from the water samples, suggesting that a molar excess of AA is necessary for efficient removal.

Figure 3

Effect of addition of AA on chlorine elimination in different water samples (n = 3). ND means not detected and the detect limitation of free chlorine is 0.1 mg L−1. (a) Chlorinated Optima® water; (b) Chlorinated tap water; (c) Chloraminated Optima® water; (d) Chloraminated tap water.

Figure 3

Effect of addition of AA on chlorine elimination in different water samples (n = 3). ND means not detected and the detect limitation of free chlorine is 0.1 mg L−1. (a) Chlorinated Optima® water; (b) Chlorinated tap water; (c) Chloraminated Optima® water; (d) Chloraminated tap water.

For routine household use, adding ∼10 mg of AA into a liter of water and then stirring vigorously for 10 seconds may be a practical option to eliminate residual chlorine from tap water to well below odor detection thresholds. This added quantity of AA is well below the US and Canadian recommended daily allowance (RDA) of 75 and 90 mg (female and male respectively) for vitamin C. Because of its reaction with chlorine, only the excess AA (∼5 mg, 6–7% of RDA) is likely to be consumed as vitamin C.

Figure S4 (available online at http://www.iwaponline.com/ws/015/088.pdf) shows the high resolution mass spectra of chlorine-containing compounds in untreated, pure chlorine solution, pure AA solution and a reaction mixture of AA and chlorine, when a quadruple time of flight mass spectrometer (Triple TOF) was used to analyze the water samples. Three peaks containing chlorine were only detected in the reaction mixture, but not in blank, free chlorine alone, and AA alone. Compound A consists of molecular ions of m/z 208.9840, compound B with a molecular ion of m/z 226.9947, and compound C with a molecular ion of m/z 241.0102 in the reaction mixture after 10 s. These peaks were still detectable in the reaction mixtures after 12 h reaction. The peak intensity shows that compound B is the major product with compounds A and C as minor products.

To identify the structures of these compounds, tandem mass spectrometry (MS/MS) spectra (fragments of the parent ions) were obtained using the high resolution Triple TOF mass spectrometer. High resolution mass measurements of the fragment ions of a compound provide structural information on the compound. Figure S5 (available online at http://www.iwaponline.com/ws/015/088.pdf) shows the fragment ions of the compounds A, B, and C. The isotopic patterns of fragment ions matching those of 35,37Cl confirm that these are chlorine-containing products. Using the built-in software for searching chemical structures, possible molecular formulae and structures are obtained and listed in Tables S1 and S2 (available online at http://www.iwaponline.com/ws/015/088.pdf). The compound with a nominal m/z value of 173 was detected as a product ion of all three chlorine-containing compounds (Table S2). This compound has been identified as dehydroascorbic acid and it is a recognized oxidation product of AA (Bradshaw et al. 2011), which supports a hypothesis that the chlorine-containing compounds detected in this study are formed from AA via variable pathways (Figure 4).

Figure 4

Possible pathway of generation of chlorine-containing products from the reaction between AA and chlorine.

Figure 4

Possible pathway of generation of chlorine-containing products from the reaction between AA and chlorine.

First, chlorine addition was expected on the carbon-oxygen double bonds in the five-membered ring to form chlorinated AA compounds. This chlorinated compound (A) was detected at a nominal m/z of 209 (Figure S4 and Table S1). Second, a hydrolysis reaction may occur to add a hydroxyl into the ring and this compound (B) was detected at a nominal m/z of 227. Third, further reaction was conducted to attach a methyl to the side chain to form the compound (C) with a nominal m/z of 241 and this methyl might be derived from the impurities in sodium hypochlorite and AA solutions, which implied that substances in drinking water might influence the formation of reaction products and generate byproducts during AA treatment. Moreover, these reaction products might be quite stable as they could still be detected in the solutions after 12 h reaction (Figure S5, available online at http://www.iwaponline.com/ws/015/088.pdf). In order to assure consumers about the safety of the dechlorinated water, further investigation is suggested to evaluate the toxicity of these reaction products and the drinking water after AA treatment.

### Discussion

This research has provided a proof-of-concept for the point-of-use application of three otherwise, well-established methods of dechlorination. Based on these findings, practical implementation of the AA or HP methods for dechlorination at point-of-use will require further study to develop methods appropriate to consumer use. In particular, a practical means for consumers to deliver a small volume of HP solution would be needed for this technique to viable.

The costs of three approaches were estimated to an order of magnitude approximation. Under UV the conditions described in this study, the power consumption for elimination of residual chlorine in 5000 L water (on a drinking water tap) would be about 300 kWh, which would cost about $24 (assuming$0.08/kWh). This cost would exceed $100 when a lamp requires replacement. For AA, 50 g of vitamin C costs about$10, which is sufficient for treating 5,000 L water. If using HP to quench residual chlorine in 5,000 L water, approximately 20 ml of HP (30%) is required to eliminate free chlorine (assuming free chlorine is 2 mg/L, molar ratio of HP to free chlorine is 1.25:1) and about 50 ml of HP (30%) is required to eliminate chloramine (assuming chloramine is 2 mg/L, molar ratio of HP to chloramine is 3:1). 100 ml HP costs about $10, thus the cost for the treatment of 5,000 L water is less than$5. Any one of these approaches could be optimized for practical application such that operating cost would not be a primary barrier to adoption.

This work should be replicated over a wider range of authentic tap waters to assure similarity of reaction by-products (specifically for AA). Dehydroascorbate is a relatively innocuous by-product, similar to AA itself, but chlorine substituted by-products should be confirmed and evaluated for any toxicity concerns before AA could be confidently recommended to consumers for point-of-use reduction of chlorinous off flavors. This will require synthesis of these compounds in sufficient quantity to allow a range of toxicity testing to be performed.

The quantitative results obtained in this research demonstrate feasibility for some point-of-use options for consumers to eliminate chlorinous off-flavors without having to invest in major home water treatment devices. These concepts may allow water utilities to inform consumers about a pragmatic solution if they find chlorinous off-flavors to be a problem. There is clearly some research and development work necessary to provide confidence in recommending practical, wide-spread adoption of any of these options.

## CONCLUSION

All the dechlorination methods can remove residual chlorine as previously recognized. For the specific application of household dechlorination, AA may be the easiest dechlorination agent to use at the consumer's point-of-use among those evaluated in this study. Free chlorine and chloramine, with the initial concentrations of around 2 mg L−1, could be removed within 10 s by 2:1 molar excess dose of AA. UV irradiation was also effective for removal of residual chlorine from water, but the reaction time was longer than AA. HP can remove free chlorine effectively, the optimum molar ratio of HP to free chlorine was 1.25:1 and the optimum reaction time was 120 s, but it was not effective for removing chloramine.

Because reaction products were detected after AA dechlorination treatments, further evaluation of potential toxicity of AA treated water is indicated for this option. Consumer dissatisfaction with chlorine off-flavors is a pervasive reality. We believe that water utilities should be able to explain to their consumers, who are ultimately the community experts in tap water aesthetic quality, that at least in principle they can reliably deal with chlorinous off-flavor concerns at their own tap without the expense and maintenance of home water treatment devices or the expense and nuisance of bottled water.

## ACKNOWLEDGEMENTS

This work has been funded by a Natural Sciences and Engineering Research Council Discovery Grant to Steve E. Hrudey, a grant from Alberta Innovates-Energy and Environment Solutions (Water Resources), and by infrastructure funding from Alberta Health.

## REFERENCES

REFERENCES
Bedner
M.
MacCrehana
W. A.
George
R. H.
2004
.
Water Research
38
,
2505
2514
.
Bolton
J. R.
Linden
K. G.
2003
.
Journal of Environmental Engineering (ASCE)
129
,
209
215
.
M. P.
Barril
C.
Clark
A. C.
Prenzler
P. D.
Scollary
G. R.
2011
.
Critical Reviews in Food Science and Nutrition
51
,
479
498
.
Castagna
R.
Eiserichb
J. P.
M. S.
Stipac
P.
Crossb
C. E.
Proiettie
E.
Voss
J. C.
Grecic
L.
2008
.
Atmospheric Environment
42
,
6551
6554
.
Chaidez
C.
Gerba
C. P.
2004
.
International Journal of Environmental Health Research
14
,
253
260
.
Chen
W. L.
Jensen
J. N.
2001
.
Water Environment Research
90
,
721
731
.
Crawford
T.
2013
Chilliwack residents fight to save water supply from being chlorinated deadly E. coli bacteria has been found in drinking water but residents are fighting the fraser health authority's plan to chlorinate. Vancouver Sun. 17 February 2013. http://www.vancouversun.com/health/health/7964738/story.html.
De Laat
J.
Boudiaf
N.
Dossier-Berne
F.
2010
.
Water Research
44
,
3261
3269
.
Dotson
A. D.
Keen
V. S.
Metz
D.
Linden
K. G.
2010
UV/H2O2 treatment of drinking water increases post-chlorination DBP formation
.
Water Research
44
,
3703
3713
.
Feng
Y.
Smith
D. W.
Bolton
J. R.
2007
.
Journal of Environmental Engineering and Science
6
,
277
284
.
Ganesh
R.
Wang
L.
Lawrence
Y. C.
Tikkanan
M. W.
Peterka
G. J.
2006
Dechlorination. Handbook of Environmental Engineering: Advanced Physicochemical Treatment Processes
, Vol.
l
,
The Humana Press
,
Totowa, NJ
, pp.
441
442
.
2009
Guidelines for Canadian Drinking Water Quality, Guideline Technical Document, chlorine
.
Ottawa, Ontario
.
Hermant
B. M.
Basu
O. D.
2013
.
Journal of Environmental Engineering-ASCE
139
,
522
529
.
Hrudey
S. E.
Walker
R.
2005
Walkerton – 5 years later tragedy could have been prevented
.
Opflow
31
(
6
),
1
,
4
6
.
Hrudey
S. E.
Hrudey
E. J.
2007
.
Water Science and Technology
55
,
239
247
.
IWA
2004
The Bonn Charter for Safe Drinking Water
.
International Water Association
,
London
,
Levallois
P.
Grondin
J.
Gingras
S.
1999
.
Water Science and Technology
40
,
135
139
.
Li
J.
Blatchley
E. R.
2009
.
Environmental Science and Technology
43
,
60
65
.
MacCrehan
W. A.
Bedner
M. B.
Helz
G. R.
2005
.
Chemosphere
60
,
381
388
.
Piriou
P.
Mackry
E. D.
Suffet
I. H.
Bruchet
A.
2004
Chlorinous flavor perception in drinking water
.
Water Science and Technology
49
,
321
328
.
Puget
S.
Beno
N.
Chabanet
C.
Guichard
E.
Thomas-Danguin
T.
2010
.
Water Research
44
,
956
964
.
Skillman
J. B.
2008
.
Journal of Experimental Botany
59
,
1647
1661
.
Suffet
I. H.
A.
Chou
D.
McGuire
M. J.
Butterworth
S.
1996
AWWA taste and odor survey
.
Journal American Water Works Association
88
,
168
180
.
Vogt
R.
Schindler
R. N.
1991
.
Journal of Photochemistry and Photobiology A-Chemistry
66
,
133
140
.
Watts
M. J.
Linden
K. G.
2007
.
Water Research
41
,
2871
2878
.