Point-of-use (POU) water treatment can improve microbial water quality and decrease diarrhoeal disease incidence in development contexts [1
]. Such techniques could reduce the risk of waterborne disease transmission for the estimated 0.8 to 1.8 billion people who lack access to safe drinking water [2
]. Some studies indicate a potential for similar effectiveness water quality improvements in humanitarian emergencies [3
], where affected populations may be more vulnerable.
Coagulant/disinfection products (CDPs) are unique relative to other available POU techniques (e.g., boiling, household chlorination, ceramic filtration, etc.) because of their capacity to achieve microbial quality improvement, turbidity reductions, and a post-treatment free chlorine residual (FCR). CDPs utilize conventional drinking water treatment processes (i.e., coagulation, flocculation, sedimentation, filtration, and disinfection) in a simplified (reduced/household) batch scale. Use of such products has been studied for both development [5
] and humanitarian relief [7
] applications. Typically, CDPs are available as sachets containing two main active ingredients (i.e., a coagulant and a disinfectant) in powdered form. It has been hypothesized that the lower chlorine demand (partially due to the simultaneous addition of a coagulant) and the physical removal of some chlorine-resistant organisms (i.e., by sedimentation and filtration) may provide a health advantage over chlorination alone [5
Most CDPs (e.g., Bishan Gari, P&G Purifier of Water, etc.) rely on calcium hypochlorite as the disinfectant. Clasen and Edmondson [10
] have contended that the differentiated chemistry of sodium dichloroisocyanurate (NaDCC) is thought to be comparatively advantageous over calcium hypochlorite in contexts where water can have high or variable chlorine demands. Although NaDCC also uses hypochlorite as the active disinfectant, only 50% of its chlorine is released, and the remainder is kept as “reservoir chlorine” (bound chlorinated isocyanurates) gradually releasing further chlorine [10
]. Therefore, a NaDCC-based CDP could have a relative advantage over similar products by providing a longer-lasting protective chlorine residual.
According to Lantagne et al. [11
], some health authorities and implementing agencies have expressed concern over the formation of unwanted disinfection by-products (DBPs) when using NaDCC as a POU disinfectant. In general, there has been little peer-reviewed research devoted to the characterization of DBP formation during POU water treatment [11
], whereas two studies included the use of CDP [11
], they were limited to trihalomethanes (THMs), and did not include haloacetic acids (HAAs), another family of DBPs of concern.
Laboratory efficacy testing can serve to highlight the performance envelopes of this type of POU water treatment product [9
]. This type of testing can also serve to highlight other relevant issues such as the product’s capability of maintaining adequate FCR levels and can indicate possible levels of DBPs that could be generated by its use. The objective of this study was to evaluate the laboratory treatment efficacy and generation of DBPs (THMs and HAAs) of a CDP with a NaDCC disinfectant under challenging conditions.
2. Materials and Methods
2.1. Product Description
The AQS 10 L is a CDP manufactured by Aquasure S.A.S. (St Just St Ramber, France) in the form of a double-layered tablet containing ferric sulfate (an active coagulant) and NaDCC (a disinfectant). According to the manufacturer, these chemicals allow for treatment in a “sequenced double action” by the initial dissolution of the coagulant followed by a slow release of the disinfectant. Although the specific proprietary formulation is not known, the AQS 10 L (referred to simply as AQS hereafter) tablets, according to the manufacturer, are also said to contain two unnamed anionic polymers as flocculants. At the time of testing, the AQS tablets were still at an experimental stage and not yet commercialized. A version of the product was available at the time of writing. Each tablet is designed to treat 10 L of water. Its usage follows similar steps to other CDPs (albeit with a slightly longer settling period) consisting of manual stirring (5 min), settling (55 min), and cloth filtration (totaling a treatment time of 60 min). For cold temperatures, the manufacturer recommended the same dosage, but with a settling period of 115 min (totaling a treatment time of 120 min).
2.2. Experimental Setup
The AQS tablets were tested in sterile polypropylene buckets (10 L test volume). A Kemwater Flocculator 2000 (Kemira, Helsingborg, Sweden) stirring paddle provided uniform mixing for the prescribed time. A J-Cloth (Associated Brands, Mississauga, ON, Canada) was used as the filtration material as it was relatively thin (2 to 3 mm) and porous. This was also in line with the adopted testing approach of evaluating the CDP’s performance under challenging conditions (relative to coagulation and disinfection) and has been adopted in previous studies on POU water treatment product evaluations [9
2.3. CDP Evaluation
Treatment performance of the AQS tablet was evaluated against established water quality guidelines [16
] and humanitarian [17
] water treatment objectives: Non detectable Escherichia coli
per 100 mL, turbidity less than 5 nephelometric turbidity units (NTU), and a FCR of at least 0.5 mg/L (after a minimum 30 min contact time and pH < 8). Bacterial log10
reductions (LRs) were assessed with regards to WHO [18
] guidelines for the evaluation of POU treatment products. This part of the study was conducted between January and April 2014 with an initial batch of AQS tablets supplied by the manufacturer.
The product’s capacity to keep a target FCR level of at least 0.2 mg/L after 24 h [19
] was also assessed in simulated storage tests. This was done under varied initial water quality conditions designed to challenge the product with regard to its underpinning treatment processes (i.e., coagulation and disinfection). Tests on each water condition were repeated three times for statistical robustness.
Separate testing characterized the DBP formation in August 2015 with a new batch of tablets supplied by the manufacturer. Two families of DBPs were evaluated, consisting of four THMs (chloroform (TCM), bromodichloromethane (BDCM), dibromochloromethane (DBCM) and tribromomethane (TBM)) and five HAAs (monochloroacetic acid (MCAA), monobromoacetic acid (MBAA), dichloroacetic acid (DCAA), trichloroacetic acid (TCAA) and dibromoacetic acid (DBAA)). The sums (on a mass basis) of these four THMs and five HAAs were evaluated against USEPA [21
] maximum contaminant limits of 0.080 mg/L and 0.060 mg/L, respectively.
2.4. Treatment Performance Trials
Treatment performance trials took into consideration previous testing of CDPs [9
] that demonstrated that alkaline pH and cold temperatures are of particular interest to challenge the microbiological efficacy of CDPs (i.e., test series A, B, and C). Such conditions were shown to affect chlorination efficiency. Furthermore, acidic pH and high turbidity were also identified as being of interest for the evaluation of turbidity reductions (i.e., test series A, D, and E) due to their effect on other underpinning treatment processes (i.e., coagulation, sedimentation, and filtration). These tests were performed on a test water matrix based on a primary settled wastewater (PSW) dilution (i.e., test series A–E) as well as on natural surface waters described below (i.e., test series F and G).
2.5. Simulated Storage FCR Experiments
Additional tests were aimed at assessing the protective FCR profiles under simulated storage conditions, as used in previous product evaluations [9
]. The AQS tablet was used as per product instructions and 1 L aliquots of treated water were collected in 1000 mL polypropylene bottles that had their caps fitted, but not tightened, and stored at ambient room temperature. FCR was determined at 1, 2, 4, and 24 h following the addition of the product to the test water. Evaluation of FCR levels in simulated storage tests was conducted in both natural (test series F and G) and synthetic (i.e., test series H and I) test waters.
2.6. Disinfection by-Product Formation Tests
DBP formation tests were performed separately and in a similar fashion to other tests as per product instructions, using natural surface waters (test series J and K) and on synthetic water (test series L). 1000 mL aliquots of treated water were collected in polypropylene bottles, as in the simulated storage FCR experiments, and stored in an incubator set at 20 °C. THMs, HAAs, pH, and FCR were sampled at 2 and 24 h after treatment.
2.7. Test Waters
Different types of test waters were used in this study: Primary settled wastewater dilution (PSWd), distilled water (DW), distilled water with humic acid (DW-HA), surface water high in organics (SW-High), and surface water low in organics (SW-Low). These are further described below and summarised in Table 1
with the corresponding test series.
A 1:5 dilution (in dechlorinated tap water) of a primary settled wastewater from a local wastewater treatment plant was used as the test water (PSWd) for test series A to E. This was in line with the WHO [18
] POU water treatment evaluation recommendations, simulating a grossly polluted drinking-water source. PSW samples were collected on a weekly basis and stored in a cold room (<4 °C) until use and had an average chemical oxygen demand of 116.8 mg/L, a suspended solids concentration of 45.4 mg/L, and a turbidity of 44.2 NTU, as reported previously [9
]. Test water quality was adjusted according to desired initial conditions (i.e., turbidity, pH, and temperature). Turbidity was adjusted using a kaolin clay (Sigma-Aldrich, St. Louis, MO, USA) to ±10% of the target turbidity. The pH was modified using solutions of either H2
or NaOH to 5.0, 7.0 or 9.0. A crushed ice jacket around the 10 L vessel kept the test water at 5 ± 1 °C for cold temperature trials, as it was shown previously that cold temperatures can affect CDP performance [15
]. All other trials were conducted at a target temperature of 20 °C (in ambient conditions).
AQS was tested (treatment performance and simulated storage trials) on natural surface waters (i.e., test series F and G) with average total organic carbon (TOC) contents that were relatively high (SW-High1: Chaudière River—averaging 8 mg/L) and low (SW-Low: Saint Lawrence River—averaging 4 mg/L). It was thought that these would be more representative of field use conditions than the synthetic PSWd test water used to characterise the product’s performance envelope with regards to different chlorine demands. As treatment performance trials were conducted between January and April 2014, a slight modification was made with regard to the natural surface water source used for series G in comparison to previous studies on CDPs. The source used previously [9
] to represent a natural surface water high in organics (i.e., Marais du Nord) was completely frozen at the time of performance testing, and thus was unavailable for treatment performance trials. SW-High1 and SW-Low samples were collected from local drinking water treatment plants and stored in a cold room (<4 °C) until use (within a week) and brought to room temperature and homogenized before each test. DBP formation tests on natural surface waters were also conducted with natural waters with TOC levels that were relatively high (SW-High2: Marais du Nord—averaging 11 mg/L) and low (SW-Low: Saint Lawrence River—averaging 4 mg/L).
Synthetic test waters (i.e., DW and DW-HA) were used in simulated storage tests (test series H and I) and in DBP formation tests. DW was the chlorine demand free water used to assess the maximum FCR attainable. Humic acid (Sigma-Aldrich) was added to DW to benchmark CDP performance if future comparisons with other products are envisioned. It was added at a concentration of 30 mg/L to make up DW-HA1 to be used in the simulated storage tests. A second synthetic test water (DW-HA2) was prepared for DBP formation tests by adding humic acid (11.5 mg/L) to DW, corresponding to a water with relatively low TOC that would still result in a 24 h FCR [23
]. These synthetic test waters were prepared daily and could serve for benchmarking purposes between products.
2.8. Analytical Methods
Turbidity, pH, and FCRs were measured in duplicates using a 2100 P turbidimeter, HQ40d pH meter, and Pocket Colorimeter™, respectively, as specified by the manufacturer (HACH, Loveland, CO, USA). Triplicate enumeration of naturally occurring E. coli
in the PSWd and natural test waters was performed with the Colilert Quanti-tray/2000 system (IDEXX Laboratories, Markham, ON, Canada). A detailed description of the THM and HAA analyses are available elsewhere [24
2.9. Statistical Analysis
An analysis of variance (ANOVA) was used to compare bacterial LRs and turbidity reductions. Where significant differences were noted, post-hoc analyses using Dunnett’s test were performed. To this end, Series A (i.e., 100 NTU, pH 7, and 20 °C) was defined as the “reference” (i.e., control) test water condition for comparisons between treatment series using the same test water matrix. Geometric means were used for microbiological data, whereas arithmetic means were used for all other parameters. A value of 0.5 MPN/100 mL was used for the calculation of bacterial geometric means and LRs with regards to non-detects (i.e., <1 MPN/100 mL). Statistically significant differences were defined at a significance level of α = 0.05.
The CDP achieved bacterial LRs greater than 4 for test series A–E, which is the minimum default value to be considered a “highly protective” POU treatment option, as defined by the WHO [18
]. Test series F and G did not allow for such an assessment to be made, as the attained bacterial LRs were censored by the relatively lower initial E. coli
concentrations of the natural surface waters that were tested. Test series at initial pH 9 and at cold temperatures (<5 °C) did not yield LRs with differences from the reference condition that were statistically significant, as observed with other CDPs tested in similar conditions [9
]. In general (except in series C—pH 5 test water), finished water pH was typically lower than initial target pH. This slight pH drop is thought to have occurred due to ferric sulfate alkalinity consumption. Finished water pH in alkaline test water conditions was below 8 (Figure 4
), above which hypochlorite ion speciation (a less effective disinfectant) is favored. This is indicative of sufficient buffering capacity of the tested CDP for alkaline conditions where it is perhaps most needed. Tests at cold temperatures had longer prescribed disinfectant contact times, which are thought to have overcome the cold temperature effects on disinfectant efficiency. Otherwise, LRs observed in this study were similar to other such studies and are consistent with findings of other laboratory evaluations [25
]. Notably, the relatively low FCR values observed (i.e., some close the method detection limit of 0.02 mg/L—Table 3
) in most trials may signify that bacterial LRs could have been higher if a stronger FCR concentration was present. The overall protective capacity of the AQS tablets could not be assessed, as viral and protozoan indicators were not tested for. According to the bacteriological water quality risk categories defined by Lloyd & Helmer [28
], finished water E. coli
levels could be considered to be of “very low” (<1 MPN/100 mL) or “low” risk (1 to 10 MPN/100 mL). It is worth noting that such laboratory efficacy results may differ from performance in the field due to a variety of factors (i.e., end-user training/habits, implementation strategy, variable water quality, etc.). Such differences (i.e., laboratory versus field performance) have been observed previously with CDPs [25
] and other POU water treatment techniques [30
The turbidity reductions (Figure 5
) of test series A to E with initial turbidity levels varying between 100 and 800 NTU were similar to those of previous studies in similar test conditions [9
]. Although final turbidity levels were mostly within 5 and 10 NTU, these may not necessarily be correlated with the visual acceptability of the finished water by potential users. Tests series F and G on natural surface waters had relatively low initial turbidity and resulted in higher final turbidity levels (Table 3
). Such increases have been observed elsewhere in low initial turbidity conditions [12
]. Relatively low turbidity reductions may also be due to the use of a relatively thin porous cloth (intended to simulate challenging conditions). Use of thicker filtration materials can improve such reductions [9
], but can also increase the treatment time, which could be a deterrent to the sustained use of POU interventions [33
Overall, the performance of the AQS with regard to FCR levels after treatment (60 min after addition) and after 24 h was poor. Apart from the trials on demand free water (DW—series H), average FCR concentrations above 0.5 mg/L were only achieved in test series F with the SW-Low test water matrix (St. Lawrence River with a relatively low average TOC of 4 mg/L). Other CDPs tested in similar conditions with a PSWd test water matrix also did not achieve 0.5 mg/L FCR after treatment [9
]. In tests with natural surface and synthetic test waters, the AQS tablet had a relatively worse performance in comparison to the product tested by Marois-Fiset et al. [9
]. However, that product only achieved the target FCR after 24 h of simulated storage when tested with DW test water. These results suggest that AQS tablet formulation may warrant improvement. According to the manufacturers, the tested version had a “slow release” NADCC disinfectant. The NADCC release kinetics were not measured, but it is possible that any disinfectant that had not been totally dissolved was filtered out once the settling period was over. It is also possible that the initial FCR levels were not sufficient to overcome the chlorine demand of the natural surface waters tested. Based on these results with the tested product formulation, it is possible that this product may not attain recommended FCR levels (after 30 min and 24 h) in field conditions.
Notably, the THM levels resulting from the lower TOC test water (SW-Low) resulted in relatively higher DBP levels in comparison to the SW-High2 (i.e., Marais du Nord) test water for which higher THM levels were expected. Although the testing design considered organic content based on an aggregate parameter (i.e., TOC), it did not consider the possible differences resulting from the organic fractionation of the surface waters tested. This may explain the higher THM values in waters with lower TOC content. It seems that the formulation of the batch used for the DBP tests differed from the one used in the treatment performance trials. Target FCR levels (at 0.5 and 24 h) were achieved in all waters tested (Figure 3
). As such, it is possible that higher bacterial LRs could have been observed if this parameter had been tested for in these trials. The pH values of test waters in DBP trials were in the neutral to mildly acidic range (Figure 3
). The addition of the AQS resulted in a slightly reduced pH after 2 h, which tended to increase back towards original values after 24 h. Lower pH values typically result in lower THM concentrations and increased HAA concentrations [34
]. This may also have explained why test water SW-Low (with higher pH) resulted in higher THM concentrations than the SW-High2 test water.
The total THM values after 24 h of CDP dosing in the present study varied between 0.027 and 0.056 mg/L. Lantagne et al. [11
] found that total THM concentrations after 24 h varied between <0.05 to between 0.020–0.025 mg/L. In another study [12
], this range was of approximately 0.020–0.110 mg/L. However, given the variability of water sources, water treatment products evaluated, and test conditions, a direct comparison between studies is perhaps not meaningful. To achieve this, a direct comparison between products would need to be done under similar conditions. This could also highlight any potential advantages between products with regard to their formulation (i.e., calcium hypochlorite versus NADCC) and benefits in terms of improved FCR levels and reduced DBP formation, as has been speculated elsewhere [10
]. However, general trends of increasing THM formation with time are in accordance with the literature. Chloroform (TCM) was the main THM species formed in accordance with other studies considering DBP formation following use of CDPs [11
]. Furthermore, these studies have also demonstrated that CDPs result in fewer DBPs after treatment when compared to chlorination alone. It is highly plausible that the CDP tested in this study would present such an advantage with regard to DBP formation. Other studies have not considered HAA formation from chlorine-based POU water treatment interventions in similar contexts. Given that this type of product is intended to be the sole form of treatment (i.e., no pre-treatment), it is difficult to operationally mitigate DBP formation considering that the coagulant and disinfectant are added at the same time. Further research could elucidate whether changes to CDP formulation could result in better DBP precursor removal and reduced overall DBP formation.
For the tested conditions in this study, both families of DBPs were below the USEPA [21
] maximum contaminant limits for THMs (0.080 mg/L) and HAAs (0.060 mg/L). THMs observed in this study were below the WHO [16
] additive toxicity guideline value (i.e., the sum of the actual values of the four THMs divided by their respective guideline values should be less than 1). It is worth noting that such regulatory and guideline values are based on chronic health risks assuming exposure periods that are longer than the intended use of CDPs, such as the one evaluated here. It would be perhaps more relevant to consider DBP-related health risks of CDPs from a sub-chronic exposure perspective (i.e., with exposure between 30 days to a year). However, sub-chronic exposure risks of DBPs are not as well characterized nor as well regulated. There is emerging but limited literature on the sub-chronic toxicity of DBPs [35
], and further research is necessary to characterize the risks and to determine health-based guidance values. This could help better inform relevant guidelines for humanitarian contexts [17
], which preconize that short-term use of chemical contaminants (interpreted here to include DBPs from the use of disinfectants) and assessments show no significant probability of negative health effects. However, any such guidelines should be formulated in light of the primary objective of CDPs, which is the production of drinking water that is microbially safe.