Removal of halides from drinking water: technological achievements in the past ten years and research needs.

Disinfection is an essential process for drinking water supplies resulting in the formation of unintended disinfection by-products (DBPs), many of which are potentially toxic and are known as the possible or probable human carcinogens. As of now, 100⁺ DBPs were characterized while about 600⁺ others can be formed in the supply water. To protect the human health, many regulatory agencies have set the guideline values for several DBPs. Removal of halide ions and natural organic matter prior to disinfection is an important step to reduce DBPs, and the associated exposure and risks. To date, many publications have reported various methods for halide removal from drinking water. The most review about halide removal technologies, associated challenges, and future research needs was published in 2012. Since then, a number of studies have been published on different methods of halide removal techniques. This paper aims to review the state of research on halide removal techniques focusing on the development during the past 10 years (2012–2021). The techniques were clustered into six major groups: adsorption, ion exchange, coagulation, advanced oxidation, membrane separation, and combined techniques. The progress on these groups of technologies, their advantages, and limitations were examined, and the future research directions to produce the safe drinking water were identified.

Keywords: Drinking water; Disinfection by-products; Halide removal techniques; Adsorption; Combined techniques; Halide removal performance

Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s11356-022-21346-z.

Disinfection is an essential process for drinking water supplies in order to prevent waterborne diseases (Postigo et al. [39]). An unintended consequence of disinfection processes (e.g., chlorination, chlorine dioxide, chloramination, ozonation, UV, and UV/H2O2) is the formation of disinfection by-products (DBPs), many of which are potentially toxic and are known as the possible or probable human carcinogens (Richardson et al. [43]; USEPA [51]). As of now, 100+ DBPs were characterized while 600+ others can be formed in the supply water (Richardson et al. [43]; Richardson and Postigo [42]; Richardson and Kimura [41]). Among the chlorinated and ozone-chlorinated systems, the unknown DBPs were reported to be in the range of 54.9–56.6% while formation of trihalomethanes (THMs) were in the range of 25.2–27.6% and haloacetic acids (HAAs) were in the range of 14.4–15.4% (Hua and Reckhow [20]). Chowdhury ([6]) reported higher averages of THMs (17.2–59.2 µg/L) than HAAs (14.0–44.8 µg/L) for different regions of Ontario, Canada, while for Quebec, average concentrations of HAAs were higher than THMs in many regions (Chowdhury [6]). In Newfoundland, Canada, averages of THMs were reported to be lower than HAAs for different groups of supply systems (Chowdhury [7]). Further details on the formation and distribution of DBPs can be found elsewhere (Hua and Reckhow [20]; Chowdhury [6], [7]; NL-DOE [35]).

Some of the effects of DBPs were higher bladder and colorectal cancers risks, miscarriages, low birth weights, pre-term deliveries, and cardiac irregularities (King and Marrett [26]; Savitz et al. [44]; Richardson et al. [43]; USEPA [51]). To protect the human health, different regulatory agencies have published regulations or guideline values for few DBPs including THM4 (chloroform [TCM]; bromodichloromethane [BDCM]; dibromochloromethane [DBCM]; and bromoform [TBM]), HAA5 (monochloroacetic acid [MCAA], dichloroacetic acid [DCAA], trichloroacetic acid [TCAA], monobromoacetic acid [MBAA], and dibromoacetic acid [DBAA]), bromate (BrO3), and chlorite (ClO2) (USEPA [50]; Health Canada [18]; WHO [61]).

Table 1 summarizes the values of few guidelines and regulations. The USEPA has set the regulation for THM4 and HAA5 to 80 and 60 µg/L respectively (USEPA [50]). The Health Canada guideline values for THM4 and HAA5 are 100 and 80 µg/L respectively (Health Canada [18]). The World Health Organization (WHO) has the guideline values for individual THMs (TCM = 300 µg/L; BDCM = 60 µg/L; DBCM = 100 µg/L; and TBM = 100 µg/L), and BrO3 (10 µg/L) and ClO2 (700 µg/L) (WHO [61]).

Table 1 Few DBPs in drinking water, their effects, and regulatory limits

B-2 probable human carcinogen, C possible human carcinogen, D not classified, HC Health Canada, USEPA US Environmental Protection Agency, RfD reference dose (mg/kg-day), SF slope factor (mg/kg-day)−.1, WHO World Health Organization *Inhalation

In addition to chloride, source water often contains bromide and iodide in trace concentrations. Bromide and iodide are the two main halides in water, which are the precursors for halogenated DBPs. Bromide occurs naturally in groundwaters and surface waters at concentrations, reportedly ranging from 10 to 6,000 µg/L (Ding et al. [11]; Rajaeian et al. [40]). Iodine occurs naturally in water in the form of iodide, which is largely oxidized to hypoiodide (HOI) and iodate (IO3−) during water treatment. The average concentrations of iodine in seawater, rainwater, and rivers and lakes were reported to be in the ranges of 45–60, 0.5–5.0, and 0.5–20 μg/L respectively (Whitehead [58]). The mean concentration of total iodine in drinking water in the USA was reported to be 4 µg/L, and the maximum concentration was 18 µg/L, which were mainly in the form of iodide (ATSDR [1]; WHO [62]). Based on the recent data of approximately 9200 sampling locations in the USA, Sharma et al. ([45]) reported the median values of iodide ions of 12 and 13 μg/L for surface water and groundwater, respectively, and the 95th percentile values were 320 and 1300 μg/L respectively. The concentrations of iodide were probably influenced by halite basins, organic-rich shale/oil formations, saltwater intrusion, and rainfall (Sharma et al. [45]). Recent studies reported that bromide and iodide concentrations in the untreated seawater were in the ranges of 50,000–80,000 and 21–60 µg/L respectively (Duranceau [13]; Kim et al. [25]). Intrusion of seawater into surface and groundwater sources is a major pathway for bromide and iodide in drinking water sources in the coastal regions. In addition, the anthropogenic sources of bromide include the upstream discharges of bromide-containing wastes from coal-fired power utilities, discharges of hydraulic fracturing wastewater, and other industrial sources (McTigue et al [33]). The other anthropogenic sources include contamination from brominated pesticides (e.g., methyl bromide) and fuel additives (Magazinovic et al. [32]).

Chlorination of source water forms hypochlorous acid (HOCl, pKa = 7.5) and hypochlorite ion (OCl−) in water, which reacts with the natural organic matter (NOM) to form DBPs, such as THMs and HAAs. In the presence of bromide, hypobromous acid (HOBr) is formed due to rapid oxidation of bromide. HOBr (pKa = 8.7) has a higher reaction rate constant than HOCl (Kumar and Margerum [28]; Chaukura et al. [4]) leading to the preferential formation of Br-DBPs and chloro-bromo DBPs in water. In the presence of iodide, HOI is formed, which is a weak acid (pKa = 10.7) leading to the formation of iodinated DBPs (I-DBPs) and the mixture of bromo-chloro-iodo-DBPs. The brominated and iodinated THMs (Br-THMs; I-THMs) and HAAs (Br-HAAs; I-HAAs) are more physiologically active and more toxic than the corresponding chlorinated DBPs, and many of these DBPs are possible or probable human carcinogens (Richardson et al. [43]; Kim et al. [25]; Wagner and Plewa [54]; Zhang et al. [65], [67]; USEPA [51]; Chowdhury et al. [9]). Use of ozone as a disinfectant led to the formation of bromate, which is a regulated DBP and a suspected human carcinogen (Von Gunten [53]; USEPA [51]).

Removal of halide (e.g., bromide and iodide) ions (Br− and I−) prior to disinfection is important in reducing Br-DBPs and I-DBPs in supply water. Different technologies have been used for Br− and I− removal with varying degrees of success, including membrane processes, adsorption, ion-exchange resins, coagulation, and advanced oxidation. Reverse osmosis (RO) or nanofiltration (NF) is widely used for desalination, and these techniques can reduce Br− and I− to 250–600 and < 4–16 µg/L (Duranceau [13]; Kim et al. [25]), indicating more than 99% removal of Br− and I− by these processes. However, such membrane processes are not cost-effective for halide removal from drinking water sources due to the low concentrations of Br− and I− (Ding et al. [11]; Rajaeian et al. [40]). In natural waters, concentrations of chloride (Cl−) to Br− are typically in the ratios of 100:1 to 300:1, while these ratios are much higher for Cl− to I− (Magazinovic et al. [32]; Ateia et al. [2]), which makes it difficult to target Br− and I− when concentrations of Cl− are overwhelmingly high.

Watson et al. ([55]) presented a review of research for Br− and I− removal technologies from drinking water sources. The technologies were broadly classified into membrane, electrochemical, and adsorptive techniques. In membrane, RO, NF, ion exchange membranes, electrodialysis (ED), and electrodialysis reversal (EDR) techniques were included. Among these, ED showed more advantages including minimal pre-treatment of feed water and higher water recovery than the RO processes (Valero et al. [52]). However, ED/EDR processes exclusively remove ionic species (e.g., halides) while neutral and organic DBP precursors are not removed. The other limitations of ED/EDR include high cost and more energy consumption (Strathmann [49]). Future research has been recommended to improve the ED/EDR techniques. Among the electrochemical processes, capacitive deionization (CDI) was found to be robust and energy efficient while membrane capacitive deionization (MCDI) showed the removal of halides with greater efficiency than CDI. Future research has been recommended to improve the CDI techniques for widespread applications in water treatment with particular focus on optimization of deionization, commercial development of aerogels, full-scale application, and commercialization of the technique. In terms of adsorption, hydrous oxides, activated carbon, silver-doped activated carbon, carbon aerogels, ion-exchange resins, aluminum-based adsorbents, and soils were discussed. The activated carbons and magnetic ion exchange (MIEX) were used for organic matter removal rather than halide removal. Future research has been recommended for simultaneous removal of organic matter and halides in the presence of competing anions. In addition to NOM removal, MIEX had the potential for halide removal in low alkalinity waters, which needs improvement through further research (Watson et al. [55]).

Since 2012, no update has been published on the developments in context to Br− and I− removal technologies. During this period (2012–2021), many articles were published on halide removal technologies in which adsorption, ion-exchange, coagulation, advanced oxidation, membrane separation, and combined processes were investigated. In the review presented here, these techniques were discussed with their advantages, limitations, and future research needs.

Adsorption is a widely applied process that makes the use of a variety of materials as adsorbents to remove pollutants from water. In the adsorption, the important factors that affect the process include adsorption capacity, contact time, selectivity, and kinetics. Conventional and non-conventional materials are used to reduce cost by using the low-cost and easily available materials. A number of researchers have proposed different types of adsorbents for halide removal from water. These are summarized below.

The ability of silver to react with bromide and iodide, leading to the formation of low solubility silver bromide (AgBr) [Ksp of 5.4 × 10−13 mol2 dm−6] and silver iodide (AgI) [Ksp of 8.5 × 10−17 mol2 dm−6], is the main reason for developing and using the Ag-modified adsorbents. However, there are few challenges including impregnating silver with the carrier materials, possibility of silver leaching, high cost of silver, and the competition from the other halide ions. Different types of Ag-modified adsorbents were used for halide removal (Table S-1). The silver-impregnated activated carbon (SIAC) removed 85–93% of Br− in a lab-scale study (Chen et al. [5]), while the preconditioned SIAC removed up to 98% of Br− (Rajaeian et al. [40]). The silver-loaded porous carbon spheres removed up to 98% of Br− in a lab study (Gong et al. [17]), while the silver-doped polymeric cloth (Ag-Cloth) showed very good adsorption capacities for Br− and I− in which the removal efficiency was higher for I− (Polo et al. [37]). For the initial concentrations of Br− and I− of 2.5 × 10−5 M (Br− = 20 µg/L; I− = 31.7 µg/L), the adsorption capacities for ultrapure water were 1.25 mg Br−/g Ag-Cloth and 1.66 mg I−/g Ag-Cloth respectively while for synthetic water, adsorption capacities were 0.88 mg Br−/g Ag-Cloth and 1.24 mg I−/g Ag-Cloth respectively (Polo et al. [37]).

Although most of the Ag-modified adsorbents showed very good removal efficiencies for Br− and I− in the lab-scale studies (Table S-1), these adsorbents have faced limitations in removing halides from natural waters. The Cl−, NOM, and other inorganic ions were found to be the effective competitors against Br− and I− removal from natural waters (Chen et al. [5]; Rajaeian et al. [40]). The competitive species, such as Cl−, NO−3, SO42−, and NOM impeded Br− and I− adsorptions to varying extents, which were in the order of SO42− > NO−3 > Cl− (Gong et al. [17]). The NOM was found to block the pores or compete for Ag-coated surface site (Gong et al. [17]). Polo et al. ([37]) reported reduction in Br− removal efficiencies than I− removal efficiencies in the presence of Cl−. For Cl−:Br− mass ratio of 40.5 (10 mg/L of Cl− and 247 µg/L of Br−), Br− removal efficiency was 83.5%, which was reduced to 38.1% for the mass ratio of 202.5 (50 mg/L of Cl− and 247 µg/L of Br−). For Cl− concentration of 3.55 g/L, removal of Br− and I− were reported to be 24% and 83% respectively (Polo et al. [37]). Using the silver chloride superfine activated carbon composite, Ateia et al. ([2]) reported Br− removal of 97%, 96%, 90% 86%, 75%, and 65% at Cl− concentrations of 0.5, 10, 50 200, 300, and 400 mg/L respectively while I− removal was always > 90%. Kidd et al. ([24]) used two silver impregnated graphene oxides (Tour-Ag and MH-Ag) and one silver impregnated powdered activated carbon (PAC-Ag) for removing Br− from synthetic and natural waters. In synthetic water, Tour-Ag, MH-Ag, and PAC-Ag removed 91%, 82%, and 86% of Br− respectively. In natural water, the removal efficiencies were 75%, 50%, and 30% respectively indicating significant reduction in efficiencies due possibly to the presence of the competing ions.

In addition to competing ions, other factors including initial concentrations of halides, pH, and temperature showed significant effects on Br− and I− removal efficiencies. Rajaeian et al. ([40]) used the pre-conditioned SIAC for natural water matrix from a groundwater bore in Western Australia where Br− removal was 98%, 93%, and 53% for the initial concentrations of 1, 2, and 6 mg/L respectively. The lower performance was possibly due to the limited number of accessible silver ions to react with Br− at higher initial concentration of Br− (Gong et al. [17]; Watson et al. [57]; Chen et al. [5]). Gong et al. ([17]) reported increase in Br− removal efficiency with the increase of temperature. The pH beyond the range of 4–7 reduced Br− removal efficiency. The halide removal capacity of Ag-Cloth increased from pH 0.5 to 5.0, which showed a decrease in removal efficiency at pH of 7.0 (Polo et al. [37]). In addition, release of Ag into water from SIAC in the form AgCl can be an issue (Rajaeian et al. [40]).

Using granular activated carbon (GAC) in a bench-scale study, Zhang et al. ([66]) reported 51.8–74.9% and 62.3–91.2% removal of Br− from mineral and synthetic water, respectively, for the empty bed contact time (EBCT) of 5 min. The initial concentrations of Br− were in the range of 30–600 µg/L. The GAC capability for Br− reduction was better for synthetic water than the mineral water due possibly to the more complex properties and competitive ions in the mineral water. Increase in EBCT increased Br− removal efficiency. However, in the pilot-scale study using the domestic drinking water, Br− removal efficiency decreased to 38.5–42.6% for the initial concentrations of 65.9–97.1 µg/L. The differences may be attributable to the different source waters and GAC used in the experiments (Zhang et al. [66]). In contrast, Chen et al. ([5]) reported negligible Br− removal efficiency for pre-oxidized PAC without silver impregnation. In another study, enhanced coagulation (EC) with GAC (EC/GAC) removed more than 95% I− while Br− removal performance was poor (26%) (Watson et al. [57]). Overall, GAC or PAC were not found to be a good adsorbent for removing Br− or I− from water.

Shi et al. ([46]) developed δ-Bi2O3 to remove Br− from water. Using the dosage of 2 g/L δ-Bi2O3 for the initial Br− concentrations of 50–500 mg/L, the efficiencies were 86.4%, 70%, and 56.5% at pH of 4.0, 7.0, and 9.8 respectively. For Cl− concentrations of 1000 and 5000 mg/L, the efficiencies were 55.9% and 24.1%, respectively, indicating the impeding effects of competitive ions. Nariyan et al. ([34]) developed a bismuth composite (Bi2O3–Bi2S3) synthesized via carbogenic sphere–supported strategy to remove halides from water. For an initial concentration of 0.05 mg/L of I−, the removal efficiencies were 16.6% 50.1%, and 87.7% at the adsorbent dosages of 1, 5, and 8 g/L respectively. At this time, Br− removal efficiencies were 44.2%, 71.7%, and 86.4% respectively for an initial concentration of 0.5 mg/L. In both Br− and I− removal studies, the optimum dosage was 8 g/L. It is to be noted that these dosages are relatively high, which are likely to consume large amounts of adsorbents and generate similar amounts of halide-laden waste for disposal.

Overall, most of studies to date were on Ag-modified adsorbents, which showed very good efficiencies in removing halides in the lab-scale studies. However, no study could resolve the impeding effects of the competing ions and NOM in the natural waters. Future research is warranted to improve the selectivity and efficiency of Ag-modified or other forms of novel adsorbents in the natural waters. Future studies should always include performance evaluation of newly developed sorbents under typical freshwater conditions (e.g., representative of Br−, I− levels, organic matter concentration and characteristics, and pH) for water treatment, not just studies in distilled and deionized lab waters. It would be beneficial to reduce halides and NOM simultaneously and assess the performance of the new adsorbents at different application modes (i.e., PAC or SPAC vs GAC) as well as location (at the beginning of a plant vs. after conventional clarification processes). Finally, providing a rich characterization information and details for both new adsorbents and natural waters tested will allow comparison between studies to further improve our understanding.

Ion exchange processes remove specific ions from water using resins, which exchange them with other ions of similar charge. This process is extensively used in water softening and demineralization. The regeneration properties and effectiveness for low concentrations of specific pollutants have made ion exchange an attractive technology (Hsu and Singer [19]; Kanan et al. [23]). However, the main challenge is the co-anion competition, especially from halides. A number of resins have been used in removing Br− and I− from water, while many resins have shown better affinity toward I−. Table S-2 summarizes the ion exchange studies for halide removal from water in the past decade.

The Orica Watercare's MIEX® resin was developed specifically for the removal of NOM from natural waters (Johnson and Singer [22]). MIEX® is a strong base anion-exchange resin with iron oxide integrated into a macroporous polyacrylic matrix, which provides the resin with magnetic characteristics, aiding the agglomeration and settling of the resin. It is a proven and highly successful NOM removal technology while removal of Br− is highly dependent on alkalinity and competing ions. Under optimal conditions (alkalinity = 20 mg/L; pH = 7.2), MIEX® reduced Br− from 163 to < 10 µg/L (about 93% reduction) in raw water from Manatee County, FL, whereas Br− was reduced from 94.5 to 54.1 µg/L (around 43% reduction) in raw water from Tampa, FL, under less favorable conditions (alkalinity = 91 mg/L; pH = 7.3) indicating the possible detrimental effects of higher alkalinity (Singer and Bilyk [47]; Watson et al. [55]). The MIEX-treated raw water with high alkalinity (155 mg/L) showed counter-productive results in Br− removal (Singer and Bilyk [47]). MIEX® showed the potential for halide removal in low alkalinity waters, which can be improved further through future research. Further details on research progress for MIEX® prior to 2012 can be found in literature (Johnson and Singer [22]; Watson et al. [55]).

In addition to MIEX and its other versions (e.g., MIEX-Br, MIEX-Gold, MIEX-DOC), past studies have applied different ion exchange resins for halide and NOM removal (Table S-2). These include chitosan anion exchanger (Goncalves et al. [16]), Purolite (Phetrak et al. [36]), Purolite-Br/PPA860S (Soyluoglu et al. [48]; Kanan et al. [23]; MacKeown et al. [31]), Amberlite IRA-400/410/458/910/958 (Phetrak et al. [36]; Soyluoglu et al. [48]; MacKeown et al. [31]) and others (Table S-2). The Purolite-Br removed 93.5 ± 4.5% Br− from synthetic water (Soyluoglu et al. [48]), and 95% Br− and 100% I− from surface water (Kanan et al. [23]). The Amberlite IRA-400 and IRA-910 removed 90% and 84% Br−, respectively, from a eutrophic surface water (Phetrak et al. [36]). The MIEX-Br removed 89% Br− and 100% I− from surface water, and 70% Br− from wastewater (Kanan et al. [23]). The mixture of two modified MIEX (MIEX-Br + MIEX-Gold; and Purolite-Br + MIEX-Gold) showed similar halide removal performance in comparison to the individual resins (e.g., Purolite-Br or MIEX-Br). From the list of the ion exchange resins, Purolite-Br, MIEX-Br, and Amberlite IRA-400/IRA-910 showed better removal efficiencies (Table S-2). In addition to halide removal, ion exchange resins were also used for the removal of THM precursors characterized through formation potential (FP) experiments. The Amberlite IRA-958 and IRA-410 removed I-THMs by 76% and I-THMs-FP by 76–96% respectively (Table S-2).

The performances of the ion exchange resins were affected by several factors including initial concentration of target compound, pH, temperature, mixing speed, and competing ions. Increase of MIEX doses and mixing speed increased halide removal efficiencies. The Br− removal increased from 78.6 to 98.5% when the MIEX dosage was increased from 0.5 to 5 mL/L (Ding et al. [11]). When the agitation speed was changed from 50 to 100 rpm, Br− adsorption rose from 3.23 to 4.56 mg/mL. When pH was increased from 3 to 6, Br− adsorption improved from 2.34 to 4.53 mg/mL. This might be explained by the fact that HCl is used to lower the pH of bromide solution, which introduces Cl− in the solution leading to the reduced exchange behavior of Br− with Cl− used as exchange reaction on MIEX resin surface. However, Br− adsorption was decreased from 4.53 to 1.56 mg/mL when pH was increased from 6 to 11, which might be explained by the competitive adsorption of hydroxyl ions (OH−) on MIEX resin with Br−. Increase in temperature from 285 to 333 K decreased Br− adsorption from 4.62 to 4.36 mg/mL. In the absence of competing ions, Br−adsorption was 4.65 mg/mL, which were reduced to 3.50, 2.10, and 1.05 mg/mL in the presence of Cl−, CO32−, and SO42− respectively (Ding et al. [11]). Li et al. ([30]) applied four different ion exchange resins (D205, D213, NDMP-3, and M80) in combination with chlorination for removing DBPs and organic micropollutants (OMPs) from different waters with Br− concentration up to 425 µg/L. The M80 resin showed the best performance for OMP removal. Reduction of THMs were 42.7%, 37.6%, 32.1%, and 0% by D205, D312, NDMP-3, and M80 respectively while HAA reduction were 34.0%, 31.2%, 23.0%, and 17.9% respectively. The D205 resin reduced Br-THMs by 64.6–74.2% while removal of Br-HAA was higher. However, the presence of humic acid reduced Br− removal efficiency.

Overall, these studies indicated that the performances of ion exchange resins were affected by the competing ions. Despite the performances of MIEX, Purolite-Br, MIEX-Br, Amberlite IRA-400, and IRA-910 for halide removal, there is still a need to further develop halide selective resins to minimize the impeding effects of competing ions. Further, applications of multiple resins with focus on the removal of halides both in fixed-bed and in suspended reactor conditions are likely to enhance the performance of the ion-exchange techniques, as compared to single resin systems, but research is needed. Finally, some recently emerging compounds of interests (e.g., per- and polyfluoroalkyl substances and microcystins) are also likely to be removed with ion-exchange, providing an opportunity to address multiple treatment challenges while removing halides from water. Another area of important research need is also the treatment and management of regenerant brines. Liu et al. ([29]) identified several research opportunities for brine management including resin selection, design of cycle length, segmented regeneration, chemical-free regeneration, and alternative regenerants, such as bicarbonate salts for anion exchange resins and potassium salts for cation exchange resins for reduced environmental impact. Future research is recommended for environment-friendly ion-exchange water treatment processes.

Using three different coagulants (compound ferrous, ferric chloride, and poly alum chloride), Zhu et al. ([68]) reported Br− removal in the ranges of 50.6–67.6% and 44.8–55.2% for synthetic and raw water from Songhua river (China) respectively. However, when the Ag-amended alum-based coagulation was performed, Br− removal was > 85% in surface and groundwater for Ag+:Br− molar ratio of > 50, while removal of I− was up to 95% (Gan et al. [14]). This study used DIW, SW, GW, and tertiary-treated wastewater (TWW) samples from Arizona and South Carolina. The Br− and I− ions were spiked to 200 and 20 µg/L, respectively, for DIW, SW, and GW. For TWW, dilution was needed to lower the Br− and I− levels to 200 and 20 µg/L respectively. The samples had the Cl−: Br− ratios of 15–866 mg Cl/mg Br. The dosage of 15 mg alum with Ag of 1–100 molar ratios (molar Ag:molar Br) was used. In this study, increase in Cl− significantly lowered the halide removal efficiency (Gan et al. [14]), indicating the implications of competing ions (Table S-3). This study showed comparable halide removal efficiencies for SIAC and Ag-amended coagulation for the low Cl− waters, such as bromide-rich industrial discharges. However, effects from the competing anions (Cl−, NO3−, NO2−, SO42−, PO43−, etc.) were also the major issues in these processes.

The advanced oxidation processes use highly reactive hydroxyl/oxygen radicals (·OH or O2·) for removing pollutants from water. The ·OH can be generated by several methods including ozone, hydrogen peroxide, ultraviolet radiation, Fenton's processes, or photocatalytic oxidation or by combination of one or more of these methods. These processes are suitable for selected contaminants and water disinfection while few metals can also be removed by this process. In general, these processes are costly and often need pre-treatment, which limit their applications. To date, ferrate [Fe(VI)] and silver nanoparticle-hydrogen peroxide (AgNP/H2O2) processes have been used for Br− and Cl− removal from synthetic and natural waters (Jiang et al. [21]; Polo et al. [38]). The Fe(VI) could oxidize only 16.4% Br− for the initial concentration of 0.1–16 mg/L Br− indicating poor performance in reducing Br− (Jiang et al. [21]). Polo et al. ([38]) applied AgNP/H2O2 process for removing Br− and Cl− using synthetic water with 0.03 mM/L of Br− and Cl−. At pH of 3, 7, and 12, Br− removal was 63, 70, and 100% respectively while Cl− removal was 50, 65, and 91% respectively for the AgNP and H2O2 dosages of 5 × 10−4 and 0.06 mol/L respectively. Increase in pH from 7 to 12 increased the performance of AgNP/H2O2 process significantly while in ion-exchange process, Br− adsorption was decreased when pH was increased from 6 to 11. However, the causes of enhanced removal efficiencies at higher pH were not explained. Further details can be found in Table S-4.

Membranes are permeable or semi-permeable barriers that filter or remove selected pollutants from water and wastewater while allowing the passage of certain constituents. Membranes generally show high performance, are simple to operate, and consume less space. However, fouling and replacement of membranes, and high energy consumption, are several issues in the membrane-based processes. Wisnieswki and Kliber ([59]) used two different anion-exchange membranes (Selemion AMV and Neosepta ACS) in a lab-scale study to remove Br− from natural water enriched with Br− (500 μg/L) with NaCl concentration of 100 mM/L in the receiver. Removal of Br− by the Selemion AMV and Neosepta ACS membranes was 86% and 90% respectively (Wisnieswki and Kliber [59]; Wisnieswki et al. [60]). However, NaCl of 300 mM/L was required in the receiving water for the Selemion AMV to achieve similar performance to Neosepta ACS membrane (Table S-5). Low concentration of salt and current density of 50 A/m2 showed the best results for this process (Wisnieswki et al. [60]).

In the recent years, more studies have investigated the performances of combined approaches in removing Br−, I−, NOM, and DBPs. Most of the studies applied MIEX in combination with other methods for the combined techniques (Xu et al. [63], [64]; Watson et al. [56]; Gibert et al. [15]). MIEX with a high-performance coagulant (HPAC) removed Br− and DOC by 55–91% and 35–50% respectively (Xu et al. [63]), which were much higher than the HPAC alone (Br−: 0–15%; DOC: 14–25%). HPAC is a composite coagulant with 10% Al2O3 made from polyaluminum chloride with other organic and inorganic additives (hydroxycitric acid and activated silica, enhancing both the charge neutralization and bridging abilities of polyaluminum chloride). Xu et al. ([64]) applied MIEX, ferric coagulant, and their combination in removing Br− and DOM. For the initial Br− concentration of 500 µg/L, removal efficiencies were in the range of 55–91% at the MIEX dose of 8 mL/L. When MIEX dosage was less than 4 mL/L, Br− removal was much lower. This can be explained by the fact that DOM was mainly removed at low MIEX dosages and Br− could compete with DOM for exchange sites at higher MIEX dose. Combination of MIEX with ultrafiltration (UF) removed Br− and DOC in the ranges of 16–37% and 32–46% respectively (Gibert et al. [15]). Watson et al. ([56]) reported up to 53% and 78% removal of Br− and I−, respectively, for MIEX alone.

The enhanced coagulation (EC) and EC followed by PAC (EC/PAC) were effective in removing NOM (70% removal) and DBPs (80–95% removal), while these were ineffective in removing Br− (Kristiana et al. [27]). For EC, Br− was reduced by 7.7% while for EC/PAC, Br− was increased by 1.8% (Kristiana et al. [27]). Watson et al. ([56]) also reported poor performance of EC/GAC on halide removal. On the other hand, EC followed by SIAC (EC/SIAC) removed more than 95% of Br− and I− from synthetic water. The EC followed by GAC (EC/GAC) removed more than 95% of I− ions while Br− removal was only 26 ± 4% (Watson et al. [57]). Adsorption of Br− were 114.9 and 27.7 μg/g of EC/SIAC and EC/GAC respectively. Cohen et al. ([10]) employed the hybrid physical adsorption (HPA) and capacitive deionization (HPA-CDI) for Br− removal and recovery, while Br− removal was 3.15–3.51 mM/g of AC-Cloth for the 1st–3rd cycles. Dorji et al. ([12]) applied MCDI for Br− and I− removal from a synthetic water mixture with Br− and I− concentrations of 0.5, 2.0, and 8.0 mM/L. In this study, Br− removal was 95–99%, 60–98%, and 8–31%, respectively, while I− removal was 95–99%, 84–99%, and 43–70% respectively. In natural water, Br− and I− removal were 91.2 and 54.1% respectively. Further details on the combined techniques can be found in Table S-6.

In addition to the above-noted processes, past studies have also investigated the electrochemical techniques, layered double hydroxides, carbon aerogels, and soils (Watson et al. [55]). However, no further progress was reported on these processes beyond 2012.

As reviewed in this paper, since the review of Watson et al. ([55]), numerous studies were published in the past 10 years. The majority of them have focused on sorption-based processes, primarily developing and/or testing sorbents, mainly silver-impregnated activating carbons and ion exchange resins, while a few studies have also examined coagulation, advanced oxidation, and membrane separation processes. Furthermore, combined processes have also been studied for removing halides and NOM from drinking water.

From application readiness perspective, ion exchange appears to be still the most promising method for the control of halides. A number of ion exchange resins showed up to 95 to 100% removal of halides (Table S-2). At the same time, simultaneous removal of organic DBP precursors as well as some other emerging contaminants presents additional advantages. However, the presence of competing co-anions, especially chloride and sulfate, was found to have the detrimental impact on halide removal efficiencies. In natural waters, concentrations of Cl− to Br− are typically in the mass ratios of 100:1 to 300:1, and Cl− to I− ratios are likely to be higher (Magazinovic et al. [32]; Ateia et al. [2]), which make it difficult to target Br− and I− when concentrations Cl− are overwhelmingly high (Table 2).

Table 2 Main findings of different techniques for halide removal

Several attempts have been used in the laboratory studies to produce and develop silver-based AC or other sorbents. Halide removal ranged from 53 to 98% in these studies (Table S-1). However, like ion-exchange resins, the performances of silver-based sorbents were reduced significantly due to the presence of co-anions (Cl−, SO42−, etc.) and NOM in natural water. The high ratio of Cl−:Br− showed detrimental effects on the performances of SIAC (Chen et al. [5]; Rajaeian et al. [40]; Kidd et al. [24]). The NOM was found to block the pores or compete for Ag-coated surface sites in SIAC (Polo et al. [37], [38]). Therefore, increasing the sorbent selectivity for both SIAC and ion-exchange type sorbents in the presence of competing background ions remains the main challenge to further enhance effectiveness of these applications. Combined processes such as EC followed by application of MIEX or SIAC or use of MCDI were found to remove more than 90% of halide ions (Table S-6). More studies have been reported using mixed resins or combined processes, which may provide some new approaches to control such competitions.

In terms of application feasibility, SIAC and MIEX are commercially available, which is an additional advantage in context to practical application and availability. However, the dosages of SIAC and other Ag-doped adsorbents were relatively high, indicating possible high cost and large amount of waste generation for the large plants. The high dosages of adsorbents can result in higher levels of Ag leaching into effluent while the possible toxic effects of the fine particles of Ag are yet to be assessed. For silver-based sorbents, it is important to assure the stability of silver on the surface to minimize its leaching into water.

• Improvement of SIAC
• As demonstrated in the past studies, high cost of AC, impregnation of Ag onto AC, detrimental effects of competing anions, and halide laden waste disposal are some of the major issues of SIAC. Alginate is a sea weed, which is widely available at a very low cost, and commonly used in food industry and medical treatment. Several past studies impregnated Fe onto alginate beads to remove arsenic from aqueous solution (Chowdhury et al. [8]), which had the advantage of regeneration. Impregnation of Ag onto alginate beads might assist in reducing the cost of the adsorbents, which can be further reduced through regeneration. In addition, a low-cost functionalized AC was developed using jute stick, which was an agricultural byproduct and showed an excellent performance in removing Pb2+ from aqueous solution. Impregnation of Ag onto jute stick AC (JSAC) might assist in reducing the cost of the SIAC while regeneration can reduce the cost further. The highly porous and high surface area of JSAC (1142.4 m2/g) is likely to hold more Ag particles (e.g., more adsorption sites) than many commercially available SIAC (Aziz et al. [3]), in which additional halide and competitive ions can be adsorbed. Appropriate functionalization of JSAC might assist in reducing NOM and metal ions (e.g., Pb2+) from water simultaneously, which can be an additional advantage of using JSAC. Another option is to develop low-cost polymeric membranes onto which Ag particles/Ag nanoparticles can be impregnated. Future research is needed in these directions to improve the performance of Ag-modified adsorbents in presence of competitive anions and NOM in natural water.
• Improvement of ion-exchange methods
• There is a need to improve the performance of ion-exchange resins in context to ion selectivity to reduce the effects of competing anions. In addition, the cost of ion-exchange resins is typically high, which often restricts their applications to the small- and medium-scale water treatment processes. Research is needed to reduce the cost of the resins for large-scale and commercial applications. Further, combination of multiple ion-exchange resins can have an added advantage, which can be investigated further for improved performance. Future research is warranted in these directions.
• High initial concentration setback
• The halide removal methods were mostly tested for the lab-scale experiments using synthetic water samples. In these experiments, halide removal efficiencies were decreased significantly with the increase of initial concentrations of halides. The source waters in many coastal regions often have higher concentrations of halides, competing anions, Br-DBPs, and I-DBPs as well as NOM. Consequently, the methods are likely to show poor performance in halide removal for these waters. Research is needed to improve the halide removal performances for high initial concentrations of halides in the presence of competing anions.
• Halide laden waste disposal
• Following the applications of halide removal methods, halide laden waste and/or membrane rejects are accumulated. Unsafe disposal may contaminate the nearby water sources, which can have negative consequences on the product water quality. No study to date discussed the safe disposal strategies for the waste considering the protection of the source waters.
• Combination of halide removal methods
• Due to complexity of the problems, a single method may not be enough to address all issues. Application of MIEX in the EC-treated synthetic water reduced Br− and I− in the ranges of 0–53% and 4–78% respectively while SIAC reduced Br− and I− in the ranges of 95 ± 4% and > 95% respectively. The EC followed by GAC removed Br− and I− in the ranges of 26±4% and >95% respectively. These experiments were conducted using the synthetic water samples. Appropriate combination of the methods should be selected to achieve improved performance for halide removal in presence of competing anions and NOM in natural water.
• Commercialization of halide removal methods

The development and application of halide removal techniques need to be commercialized for large-scale applications. The hybrid techniques using the combination of multiple approaches might assist in achieving the ultimate goal of removing halides, competing anions and NOM. Adaptation for commercial production and application of ion exchange resins, minimizing the effects of NOM and competing anions during the adsorption processes, and development of improved membranes with low energy requirement are some of the necessary research directions for future. Further, cost reduction is essential for the small and medium scale treatment plants. Research is needed toward cost reduction for halide removal techniques.

This paper investigated the state of research on the techniques for halide removal from drinking water to date with particular focus since 2012. The techniques were divided into six major cluster: adsorption, ion exchange, coagulation, advanced oxidation, membrane separation, and combined techniques. Among these groups, adsorption, ion exchange, and combined techniques were found to be more promising while their applications in natural waters with competing anions and NOM often reduced their performances. Selectivity and effectiveness of such sorbents in challenging water matrices remain as the main challenge that need to be addressed. The combined approaches, such as enhanced coagulation followed by selected adsorption techniques (e.g., SIAC) or ion exchange (e.g., MIEX) or multiple mixed resin systems (e.g., Purolite-Br and MIEX-Gold; MIEX-Br and MIEX-Gold) were found to show better performance in the lab-scale studies. However, limited studies were tested for natural water with halide, competing anions, and NOM. More investigations in this area may provide useful solutions as the removal or managing of halides in competitive water matrices for DBP control will not likely rely on the performance of a single process.

Dr. Chowdhury: Review article planning, collection of published articles, writing, editing, and overall management.

Engr. Koyappathody: Article collection, data tabulation, draft paper writing.

Dr. Karanfil: Writing, editing, innovative/critical idea accumulation.

This work received financial support provided by the Deanship of Research Oversight and Coordination (DROC) at King Fahd University of Petroleum & Minerals (KFUPM) through project No. DF 191013.

All the data and related materials are available with the corresponding author (Shakhawat Chowdhury).

The authors declare no potential conflicts of interest associated with this article preparation and publication. The research does not include human participants and/or animals.

Not applicable

Not applicable

The authors declare no competing interests.

Below is the link to the electronic supplementary material.

Graph: Supplementary file1 (DOCX 49 KB)

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.