Page 8: Guidelines for Canadian Drinking Water Quality: Guideline Technical Document – Toluene, Ethylbenzene and the Xylenes

7.0 Treatment technology and distribution system considerations

7.1 Municipal scale

Conventional water treatment techniques (coagulation, sedimentation, filtration and chlorination) have little or no effect in reducing concentrations of VOCs in drinking water (Love and Eilers, 1982; Love et al., 1983; Lykins and Clark, 1994). Some incidental removal of VOCs may occur as a result of volatilization in open basins and open weirs (AWWA, 1991; Health and Welfare Canada, 1993).

Granular activated carbon (GAC) and packed tower aeration (PTA) have been identified as the best available technologies (BATs) for TEX removal from drinking water (U.S. EPA, 1998). Small system compliance technologies include GAC, PTA, diffused aeration, multistage bubble aeration, tray aeration and shallow tray aeration (U.S. EPA, 1998).

The selection of an appropriate treatment process for a specific water supply will depend on many factors, including the characteristics of the raw water supply, the concentration of the contaminant and the operational conditions of the specific treatment method. These factors should be taken into consideration to ensure that the treatment process selected is effective for the reduction of TEX in drinking water.

7.1.1 Activated carbon

Activated carbon is used in the water treatment process either as GAC or as powdered activated carbon (PAC). Generally, GAC and PAC are used for different purposes. GAC is used for the control of organic compounds as a barrier to occasional spikes of organics in surface water, taste and odor control as well as disinfection by-products control. In contrast, PAC controls taste and odor compounds and strongly adsorbs low levels of pesticides and herbicides. The adsorption capacity of activated carbon to remove VOCs is affected by a variety of factors, such as competition from other contaminants, preloading with natural organic matter (NOM), temperature and the physicochemical properties of the VOCs and the carbon media (Speth, 1990; AWWA, 1991). The PAC application, most suitable for conventional treatment systems treating surface waters, may remove occasional low concentrations of VOCs such as TEX when it is applied at the treatment plant, allowing sufficient contact time and proper mixing. In conventional treatment plants, the common points of PAC application are at the plant intake, during the rapid mix process and in the filter influent. Sufficient contact time is necessary, and the time required is a function of the characteristics and the concentration of the contaminant to be adsorbed (Najm et al., 1991). In general, PAC adsorption is found to be less efficient than GAC adsorption for VOCs removal largely due to: (1) its use in coagulation/sedimentation basins where the adsorption sites can be blocked due to floc formation; (2) the fact that it will not have the necessary adsorption time to reach its maximum capacity; and (3) the fact that the equilibrium liquid-phase concentration (concentration gradient driving force) goes down during the adsorption process. A conventional water treatment technique in combination with a PAC dose in the range of 8–27 mg/L demonstrated reduction rates of up to 67% for toluene, from 33% to 99% for ethylbenzene and from 60% to > 99% for total xylenes, respectively (U.S. EPA, 1991a).

Higher concentrations of VOCs are typically found in groundwater, and GAC adsorption is the most commonly used treatment process for the removal of VOCs from groundwater (Snoeyink, 1990). In the GAC process, as water passes through the GAC contactor, the contaminants diffuse into the adsorbent granules and accumulate on the inner surface within the pores. The GAC column allows more complete contact between water and the media, greater adsorption efficiency and greater process control than with PAC. Full-scale and pilot-scale data demonstrated that removal efficiency in the range of 75–99% for TEX is achievable using GAC adsorption (Miltner at al., 1987; U.S. EPA, 1990; AWWA, 1991).

The selection of GAC for removing VOCs from drinking water supplies should factor in the following process design considerations: carbon usage rate, empty bed contact time (EBCT), type of adsorbent, pretreatment of the raw water, contactor configuration (e.g., beds in series or parallel operation) and method of GAC regeneration or replacement. During the operation time, and depending on the variety of factors discussed above, organic contaminants will "break through" the carbon bed. Initial breakthrough is defined as the time when the contaminant concentration in the effluent exceeds the treatment objective. In systems with multiple beds, the individual beds can be operated beyond the time of initial breakthrough, provided the blended effluents still meet the treatment objectives. Once the GAC is exhausted, it is removed from the contactor and replaced with fresh or regenerated GAC. For practical reasons, PAC is not recovered and reactivated; thus, its carbon use rate can be high compared with that of GAC (Chowdhury et al., 2013). The regeneration and/or replacement of exhausted media are important economic considerations in achieving the contaminant removal.

Common operational issues when using GAC adsorption contactors can include biological growth and a concurrent increase in heterotrophic plate counts in the treated water as well as clogging and fouling of the carbon adsorber by chemicals, bacterial precipitants and background organic matter (AWWA, 1991). Operational considerations may also include the need to ensure a proper backwash, maintain the bed depth and bed density after backwashing and control the flow rate. To prevent the bed from clogging, pretreatment of the water before it enters the GAC contactor is often required (Snoeyink, 1990; Speth, 1990; AWWA, 1991; Crittenden et al., 2005).

When designing a GAC system, relevant information and operational parameters are obtained from pilot plant experiments, rapid small-scale columns test (RSSCT) and vendor experience. Mathematical model can also be used to predict process performance and select the optimum process design as an additional or complementary approach to pilot plant and RSSCT experiments.

7.1.1.1 Toluene

Full-scale data demonstrated that two GAC adsorbers operating in series with a hydraulic loading rate of 2.2 gpm/ft2 (5.4 m/h) and an EBCT of 27 minutes were capable of achieving greater than 99% reduction of a toluene concentration in groundwater. However, the influent concentration of 220 µg/L has been reported as a total BTX (benzene, toluene and xylenes) concentration (AWWA, 1991).

A pilot-scale study indicated that a GAC column operating with a hydraulic loading rate of 2.0 gpm/ft2 (4.9 m/h) and an EBCT of 1.0 minute was capable of reducing a toluene concentration of 24.55 µg/L in groundwater to 6.1 µg/L, achieving a carbon usage rate of 0.066 lb/1000 gal (0.008 kg/m 3) and a bed life of 1.5 months (Miltner et al., 1987; U.S. EPA, 1990). The results from the pilot-scale study were applied to predict the carbon usage rate of a full-scale GAC contactor. The GAC contactor, designed with a bed depth of 4.05 ft (1.23 m), a hydraulic loading rate of 2.0 gpm/ft 2(4.9 m/h) and an EBCT of 15 minutes, could reduce an influent concentration of 500 µg/L to 5 µg/L. Under these conditions, a carbon usage rate of 0.33 lb/1000 gal (0.04 kg/m3) and a bed life of 4.5 months would be required to achieve a 99% reduction of toluene (Miltner at al., 1987).

Operational data from a groundwater remediation site demonstrated that a full-scale GAC system with a bed life of 3.5 months was capable of reducing toluene concentrations ranging from 4.4 to 55.0 mg/L to 5 µg/L. The breakthrough of the carbon bed was defined as the time when the benzene concentration in the effluent exceeded 5 µg/L (369 917 L of treated water) (Giffin and Davis, 1998).

7.1.1.2 Ethylbenzene

Data from a pilot-scale study demonstrated that the influent concentration of ethylbenzene of 4.5 µg/L in groundwater was reduced to 1.1 µg/L using a column with a hydraulic loading rate of 2.0 gpm/ft2 (4.9 m/h) and an EBCT of 1.0 minute, resulting in a carbon usage rate of 0.071 lb/1000 gal (0.009 kg/m3) and a bed life of 1.4 months (Miltner et al., 1987; U.S. EPA, 1990). Using the input parameters from this study, Miltner et al. (1987) predicted the carbon usage rate of a full-scale GAC adsorber. The GAC adsorber, designed with a bed depth of 4.05 ft (1.23 m), a liquid loading rate of 2.0 gpm/ft2 (4.9 m/h) and an EBCT of 15 minutes, could reduce an influent ethylbenzene concentration of 100 µg/L to 10 µg/L (90% reduction). The resulting carbon usage rate would be 0.24 lb/1000 gal (0.03 kg/m3), and the bed life would be 6.1 months under these conditions (Miltner at al., 1987).

Operational data from a groundwater remediation site demonstrated that a full-scale GAC system with a bed life of 3.5 months was capable of reducing ethylbenzene concentrations in the range of 0.54–3.9 mg/L to below 1 µg/L. The breakthrough of the carbon bed was defined as the time when the benzene concentration in the effluent exceeded 5 µg/L (369 917 L of treated water) (Giffin and Davis, 1998).

7.1.1.3 Xylenes

Full-scale data demonstrated that two GAC adsorbers operating in series with a hydraulic loading rate of 2.2 gpm/ft2 (5.4 m/h) and an EBCT of 27 minutes were capable of achieving a greater than 99% reduction in the concentration of total xylenes in groundwater. The influent concentration of 220 µg/L was reported as a total BTX concentration (AWWA, 1991).

A pilot-scale study using a GAC column indicated that the influent concentration of m-xylene of 5.45 µg/L in groundwater could be reduced to 1.36 µg/L using a hydraulic loading rate of 2.03 gpm/ft2 (4.9 m/h) and an EBCT of 1.0 minute, resulting in a carbon usage rate of 0.069 lb/1000 gal (0.008 kg/m3) and a bed life of 1.4 months. Under the same operating conditions, the concentration of o/p-xylene of 9.0 µg/L was reduced to 2.25 µg/L, achieving a carbon usage rate of 0.172 lb/1000 gal (0.02 kg/m3) and a bed life of 0.6 month (Miltner et al., 1987; U.S.EPA, 1990). Based on the pilot-scale results, Miltner et al. (1987) predicted the carbon usage rates of a full-scale GAC adsorber. The GAC contactor, designed with a bed depth of 4.05 ft (1.23 m), a hydraulic loading rate of 2.03 gpm/ft2 (4.95 m/h) and an EBCT of 15 minutes, could reduce influent concentrations of 1000 µg/L for each of m-xylene and o/p-xylene to 50 µg/L. Under these conditions, a 95% reduction could be achieved, resulting in a carbon usage rate of 1.9 lb/1000 gal (0.22 kg/m3) and a bed life of 0.8 month for m-xylene and a carbon usage rate of 1.2 lb/1000 gal (0.14 kg/m3) and a bed life of 1.3 months for o/p-xylene (Miltner et al., 1987).

Operational data from a groundwater remediation site demonstrated that a full-scale GAC system with a bed life of 3.5 months was capable of reducing a concentration of total xylenes ranging from 2.7 to 23.0 mg/L to below 3.0 µg/L The breakthrough of the carbon bed was defined as the time when the benzene concentration in the effluent exceeded 5 µg/L (369 917 L of treated water) (Giffin and Davis, 1998).

7.1.2 Air stripping

Air stripping is a well-established technology for removing VOCs from drinking water (Cummins and Westrick, 1990; U.S. EPA, 1991a; Dyksen, 2005; WHO, 2011). A variety of configurations exist with respect to air stripping equipment; however, PTA provides an optimum system for the removal of VOCs from drinking water. Removal efficiencies in the range from 94% to > 99% for TEX in drinking water are considered to be achievable using a PTA column (Hand et al., 1986; U.S. EPA, 1990; AWWA, 1991). PTA application allows for greater air-to-water ratios than with other aeration systems (i.e., diffuser aerator, multiple tray aerator, spray aerator, mechanical aerator). In concurrent PTA, the contaminated water flows downward by gravity over a bed of packing material, while air is introduced into the tower below the packed bed and flows upward, countercurrent to the water flow. As PTA transfers VOCs from water to air, treatment of the stripping tower off-gas to reduce the contaminant concentrations prior to discharge into the atmosphere may be necessary (Crittenden et al., 1988; Adams and Clark, 1991).

Several factors affect the stripping rate of VOCs: air-to-water ratio (A:W), available area of mass transfer, hydraulic loading rate, temperature of the water and air, gas pressure drop, types and size of packing materials and the physical and chemical properties (e.g., liquid-phase mass transfer coefficients) of the contaminant (AWWA, 1991; Crittenden et al., 2005; Dyksen et al. 2005). Diffused aeration, multistage bubble aerators, tray aeration and shallow tray aeration have been identified as alternative air stripping treatment technologies for the reduction of toluene, ethylbenzene and xylenes in drinking water for small systems (U.S. EPA, 1998).

A common operating problem is scaling and fouling of the column. The main causes of fouling are calcium carbonate and/or calcium sulphate scale, iron oxidation, microbial growth and NOM. Methods to prevent fouling of the column include pH suppression of the influent, using scale inhibitors or iron removal prior to the PTA application. Algal growth can also be a problem in locations where light could be introduced into the tower. Post treatment, such as the use of a corrosion inhibitor, may also be required to reduce corrosive properties of the water due to increased dissolved oxygen from the aeration process. Environmental conditions, such as water temperature, may impact the packed tower performance. Although temperatures below freezing can cause operational issues, contact between water and air in PTA will result in a change in the air temperature until it approaches the water temperature. The temperature influences both the Henry's Law constant and the rate of mass transfer coefficient of the contaminant. These parameters affect the size of the equipment and the removal efficiency of the VOCs (Crittenden et al., 2005).

7.1.2.1 Toluene

Full-scale data indicated that a PTA column was capable of reducing an average influent concentration of toluene of 30.9 µg/L in groundwater to an average effluent concentration of 0.94 µg/L. A reduction rate of 96.9% was achieved using an air-to-water ratio of 61, a hydraulic loading rate of 29.8 gpm/ft 2 (72.7 m/h), a tower height of 24.5 ft (7.47 m) and a packed column diameter of 8 ft (2.44 m) (Hand et al., 1986).

A pilot-scale study reported a 98.3% reduction of a toluene concentration in groundwater. An influent concentration of 62.0 µg/L was reduced to 1.1 µg/L using an air-to-water ratio of 25 and a hydraulic loading rate of 30.9 gpm/ft2 (75.4 m/h) (U.S. EPA, 1990). Another pilot-scale study demonstrated that influent concentrations in the range of 3050–13 400 µg/L were reduced to concentrations in the range of 20.9–168 µg/L (95.5–99.3% reduction). The highest removal rate of 99.3% was reported to be achieved with an influent concentration of 3050 µg/L, an air-to-water ratio of 71 and a hydraulic loading rate of 17.8 gpm/ft2 (43.4 m/h) (U.S. EPA, 1990).

Studies by Crittenden et al. (1988) and Adams and Clark (1991) estimated that a 99% removal efficiency and a toluene concentration in finished water of 1 µg/L could be achieved. According to Crittenden et al. (1988), typical air stripping design parameters at full scale, for reduction of toluene, include an A:W ratio of 30, an air column height of 39.04 ft (11.9 m) and a packed column diameter of 8.07 ft (2.4 m). Under these conditions, a 99% reduction of toluene could be achieved in water with an influent concentration of 100 µg/L, thus resulting in an effluent concentration of 1 µg/L.

Diffused aeration generally achieves removal efficiencies in the range of 22–89% for toluene and has higher power requirements than PTA systems (U.S. EPA, 1991a). A diffused aeration system was capable of reducing a toluene concentration of 108 µg/L in spiked groundwater to a finished water concentration of 14.0 µg/L. This removal rate of 87% was achieved using an air-to-water ratio of 15 (U.S. EPA, 1990).

7.1.2.2 Ethylbenzene

A full-scale PTA column reduced an influent concentration of ethylbenzene of 5.1 µg/L in groundwater to below 0.3 µg/L. A reduction rate of greater than 94% was achieved using an air-to-water ratio of 61, a hydraulic loading rate of 29.8 gpm/ft2 (72.7 m/h), a column height of 24.5 ft (7.47 m) and a packed column diameter of 8 ft (2.44 m) (Hand et al., 1986).

Two pilot-scale studies demonstrated that PTA is an effective technology to treat ethylbenzene in drinking water (U.S. EPA, 1990). A pilot-scale column was capable of reducing influent concentrations of ethylbenzene of 202.0 µg/L to 3.9 µg/L and of 59.7 µg/L to below 0.1 µg/L using A:W ratios of 59 and 63, respectively. The column operated with an average hydraulic loading rate of 23 gpm/ft2 (56.1 m/h). Another pilot PTA system reported greater than 92.0% removal of ethylbenzene from groundwater. An influent concentration of 9.0 µg/L was reduced to below 1.0 µg/L using an air-to-water ratio of 25 and a hydraulic loading rate of 30.9 gpm/ft2 (75.4 m/h) (U.S. EPA, 1990).

Crittenden et al. (1988) and Adams and Clark (1991) estimated that a 99% removal efficiency and an effluent ethylbenzene concentration of 1 µg/L could be achieved in the treated water. According to Adams and Clark (1991), estimated full-scale plant air stripping design parameters include an air-to-water ratio of 30 and an air stripper length of 36.1 ft (11.0 m).

Diffused aeration generally achieves removal efficiencies in the range of 24–89% for ethylbenzene, and the process has higher power requirements than PTA systems (U.S. EPA, 1991a). A bench-scale diffused aeration system was capable of reducing an ethylbenzene concentration of 113 µg/L in spiked groundwater to 13.6 µg/L. A removal rate of 88% was achieved using an air-to-water ratio of 15 (U.S. EPA, 1990).

7.1.2.3 Xylenes

Full-scale data indicated that a PTA column was capable of reducing an average influent concentration of xylenes of 16.6 µg/L in groundwater to an average effluent concentration of 0.6 µg/L. A removal reduction rate of 96.4% was achieved using an air-to-water ratio of 61, a hydraulic loading rate of 29.8 gpm/ft2 (72.7 m/h), a tower height of 24.5 ft (7.47 m) and a packed column diameter of 8 ft (2.44 m) (Hand et al., 1986).

A pilot-scale study indicated that a PTA column is an effective technology for reducing the concentration of xylene isomers in drinking water. Influent m-xylene concentrations in the range of 1020–1900 µg/L were reduced to concentrations ranging from 18.2 to 121 µg/L, achieving reduction rates between 93.1% and 98.2%. o/p-Xylenes concentrations in the range of 1210–1980 µg/L were reduced to concentrations in the range of 62.9–179 µg/L (87.1–94.8% removal). The highest removal rates of 98.2% and 94.8% were achieved for an m-xylene influent concentration of 1020 µg/L and for an o/p-xylene influent concentration of 1210 µg/L, respectively. The column operated with an air-to-water ratio of 71 and a hydraulic loading rate of 17.8 gpm/ft2 (43.4 m/h) (U.S. EPA, 1990).

Another pilot study demonstrated that a PTA column using an air-to-water ratio of 25 and a hydraulic loading rate of 30.9 gpm/ft2 (75.4 m/h) was capable of achieving > 95%, > 83% and > 90% removal for o-, m- and p-xylene, respectively. Reported influent concentrations were 10.0 µg/L, 2.9 µg/L and 6.9 µg/L for o-, m- and p-xylene, respectively (U.S. EPA, 1990).

Studies by Crittenden et al. (1988) and Adams and Clark (1991) estimated that a 99% removal efficiency and an effluent concentration of 1 µg/L of m-xylene could be achieved in the treated water. According to Crittenden et al. (1988), estimated full-scale plant air stripping design parameters include an A:W ratio of 37.3, an air column height of 40.5 ft (12.3 m) and a column diameter of 18.34 ft (5.6 m).

Diffused aeration generally achieves removal efficiencies in the range of 18–89% for xylene isomers and has higher power requirements than PTA systems (U.S. EPA, 1991a). A diffused aeration system was capable of achieving reductions of 82%, 84% and 88% in spiked groundwater with influent concentrations of 120 µg/L, 107 µg/L and 103 µg/L for o-, m- and p-xylene, respectively (U.S. EPA, 1990).

7.1.3 Combination of packed tower aeration and granular activated carbon

In air stripping, the VOCs are removed from the water phase to the air phase, potentially creating an air pollution problem. Generally, the use of GAC after air stripping is intended to treat the off-gas from the air stripper and to reduce the release of VOCs into the atmosphere. However, several articles have indicated that combining aeration technologies and liquid-phase GAC into a two-step treatment train is very effective for achieving low finished water concentrations of VOCs (Robeck and Love, 1983; McKinnon and Dyksen, 1984; Stenzel and Gupta, 1985; U.S. EPA, 1991a). Contaminated groundwater with a high concentration of VOCs was treated with PTA followed by GAC. The application of the aeration process and an adsorption technique reduced the VOC concentrations to below detectable levels in the drinking water (McKinnon and Dyksen, 1984).

In a municipal-scale treatment plant combining these processes, air stripping was used for the bulk reduction of VOC concentrations in water, and activated carbon was used in the second step to further reduce residual VOC concentrations below the detection limit of 0.1 µg/L (Robeck and Love, 1983). The air stripping process preceding liquid-phase GAC adsorption can also extend carbon bed life (Hess et al., 1981; Stenzel and Gupta, 1985; U.S. EPA, 1991a). The common operational problems inherent to PTA systems and GAC adsorption contactors are similar and should be considered when these combined technologies are employed. However, no information specific to the effectiveness of the process for reduction of TEX was available in the literature.

7.1.4 Ozonation and advanced oxidation processes

Ozonation and advanced oxidation processes (AOPs) have been reported to be effective for the reduction of TEX concentrations in drinking water, although full-scale data were not available for these treatment techniques (Fronk, 1987; U.S. EPA, 1990; Lykins and Clark, 1994; Topudurti et al., 1998; Bergendahl et al., 2003; Garoma et al., 2008). AOPs generate highly reactive hydroxyl radicals at ambient temperature and atmospheric pressure. Hydroxyl radicals are strong oxidants and react rapidly and non-selectively with organic contaminants. AOPs that are commercially available are ozone with hydrogen peroxide, ozone with ultraviolet (UV), ozone at elevated pH, ultraviolet with hydrogen peroxide, ultraviolet with titanium dioxide and ozone with ultraviolet and titanium dioxide. The primary advantage of AOPs is their capability to completely convert the organic compounds into carbon dioxide and mineral acids (Crittenden et al., 2005), whereas adsorption processes and air stripping techniques transfer the contaminants from one phase to another, and additional treatment may be required (Glaze et al., 1988; U.S. EPA, 1993; Crittenden et al., 2005). Physical and chemical parameters of the water have a major impact on AOPs. Water quality parameters that may impact the effectiveness of the AOPs are alkalinity, pH, NOM, reduced metal ions (iron and manganese) and turbidity (Fronk, 1987; Crittenden et al., 2005). Consequently, in order to reduce excess energy consumption in AOPs and maximize their effectiveness, pretreatment is commonly applied to remove co-occurring contaminants.

The formation of by-products from the oxidation and/or advanced oxidation of VOCs or other inorganic compounds in the source water should be considered when using these processes. AOPs produce reactive peroxy organic radicals, which undergo radical chain reactions and result in a variety of oxygenated by-products. The typical by-products produced by AOPs are aldehydes, ketones and carboxylic compounds. Another potential problem with the use of ozone or AOPs such as ozone in combination with UV or hydrogen peroxide is the formation of bromate in waters containing significant bromide concentrations (Crittenden et al., 2005).

Krasner et al. (1993) and Siddiqui and Amy (1993) reported that the application of ozone with hydrogen peroxide may increase the formation of bromate at pH in the range 8.0-8.5, when compared with ozone application alone. However, it has been observed that bromate formation can be controlled effectively by applying a higher hydrogen peroxide to ozone ratio (Liang et al., 2001). The formation of by-products may require additional treatment following AOPs and/or process optimization to minimize by-product formation.

Photolytic reactions require a significant amount of electrical energy, and the associated costs can be significant. Consequently, comparison of the process efficiency of different photolytic methods is based on electrical usage per amount of compound destroyed (Crittenden et al., 2005).

7.1.4.1 Toluene

Pilot-scale ozonation tests were conducted for the removal of VOCs from drinking water. The experiments used both distilled water and groundwater, each spiked with selected contaminants at concentrations ranging between 50 and 384 µg/L. The influent concentration of toluene was not reported. The study reported that a transferred (absorbed) ozone dose (defined as the applied ozone minus the off-gas concentration) of 10 mg/L achieved 96% reduction of toluene in groundwater (Fronk, 1987). A bench-scale test demonstrated that ozone effectively oxidized toluene in spiked groundwater with a contact time of 13 minutes. The toluene concentrations of 108.0 µg/L and 7.0 µg/L were reduced by 91% and 100% using ozone doses of 10.1 mg/L and 9.4 mg/L, respectively (U.S. EPA, 1990).

A pilot-scale photocatalytic oxidation system was evaluated for the treatment of groundwater contaminated with VOCs, including toluene. The system utilized a 254 nm UV light source combined with the addition of 70.0 mg/L of H2O2 and 0.4 mg/L of ozone. The system achieved a reduction rate greater than 93% of the influent toluene concentrations, which ranged from 44 to 85 µg/L. To prevent fouling of the photocatalytic reactor, an ion exchange pretreatment system was used to remove iron and manganese from the groundwater. The authors reported the formation of several oxidation by-products, including aldehydes (formaldehyde: 237 µg/L; acetaldehyde: 50 µg/L) and haloacetic acids (monochloroacetic acid: 14 µg/L; dichloroacetic acid: 15 µg/L) (Topudurti et al., 1998).

Bench-scale AOPs using UV and UV/H2O2 have been assessed for the removal of toluene from spiked distilled water. Data demonstrated that UV radiation with an intensity of 85 µW/cm2 was capable of achieving a 69% reduction of a toluene concentration of 51.7 µg/L. The reduction rate was increased to 100% when 10 mg/L of H2O2 has been added to the UV system (U.S. EPA, 1990).

A laboratory study investigated the O3/UV oxidation process for the degradation of benzene, toluene, ethylbenzene and xylenes (BTEX) in contaminated groundwater under different experimental conditions. UV radiation with an intensity of 2.73 W/L (1.4 kWh/m3) combined with the addition of an ozone dose of 28 mg/L achieved a greater than 99% reduction of the influent toluene concentration of 83 µg/L in 30 minutes (Garoma et al., 2008).

7.1.4.2 Ethylbenzene

Pilot-scale ozonation tests were conducted for the removal of VOCs from drinking water. The experiments used both distilled water and groundwater, each spiked with selected contaminants. The concentrations of the contaminants ranged between 50 and 384 µg/L, concentrations typically found in groundwater (Fronk, 1987). The influent concentration of ethylbenzene was not reported. The study reported that a transferred (absorbed) ozone dose (defined as the applied ozone minus the off-gas concentration) of 10 mg/L achieved a 95% reduction of ethylbenzene concentration in groundwater (Fronk, 1987). A bench-scale study demonstrated that ozone effectively oxidized ethylbenzene in spiked groundwater with a contact time of 13 minutes. The study reported a 90% and 100% degradation of the initial ethylbenzene concentrations of 113.0 µg/L and 42.0 µg/L using ozone doses of 10.1 mg/L and 9.4 mg/L, respectively (U.S. EPA, 1990).

7.1.4.3 Xylenes

Pilot-scale ozonation tests were conducted for the removal of VOCs from drinking water. The experiments used both distilled water and groundwater spiked with selected contaminants at concentrations ranging between 50 and 384 µg/L. Although the influent concentrations of xylene isomers were not reported specifically, the tests demonstrated that a transferred (absorbed) ozone dose of 10 mg/L achieved 95%, 97% and 97% reductions of o-, m- and p-xylene in groundwater, respectively (Fronk, 1987). A bench-scale test reported that the ozone effectively oxidized xylene isomers. Spiked groundwater concentrations of 120.0 µg/L for o-xylene, 107.0 µg/L for m-xylene and 103.0 µg/L for p-xylene were reduced to 8.4 µg/L, 7.5 µg/L and 7.2 µg/L, respectively, using an ozone dose of 10.1 mg/L. However, an ozone dose of 9.4 mg/L achieved 100% reduction when the initial concentrations were 60.0 µg/L, 46.0 µg/L and 44.0 µg/L for o-, m- and p-xylene, respectively (U.S. EPA, 1990).

A pilot-scale photocatalytic oxidation system was capable of achieving a reduction rate of greater than 98.0% of total xylene concentrations ranging from 55 to 203 µg/L. The system used a 254 nm UV light source combined with the addition of 70 mg/L of H2O2 and 0.4 mg/L of ozone. To prevent fouling of the photocatalytic reactor, an ion- exchange pretreatment system was used to remove iron and manganese from the groundwater. The authors reported the formation of several by-products, including aldehydes (formaldehyde: 237 µg/L; acetaldehyde: 50 µg/L) and haloacetic acids (monochloroacetic acid: 14 µg/L; dichloroacetic acid: 15 µg/L) (Topudurti et al., 1998).

A laboratory study investigated the O3/UV oxidation process for the degradation of BTEX from contaminated groundwater under different experimental conditions. UV radiation with an intensity of 2.73 W/L (1.4 kWh/m3) combined with an ozone dose of 28 mg/L achieved a greater than 99% reduction of an influent concentration of total xylenes of 71 µg/L in 30 minutes (Garoma et al., 2008).

7.1.5 Emerging technologies

New technologies have demonstrated the potential for removing VOCs, including TEX, but there is insufficient information available at present to fully evaluate them:

  • Membrane pervaporation— Although the use of membranes for the pervaporation extraction of VOCs has been applied primarily in wastewater treatment, this technique has also been studied for the removal of VOCs from groundwater (Jian and Pintauro, 1997; Uragami et al., 2001; Peng et al., 2003). Pervaporation is a process in which a liquid stream containing contaminants is placed in contact with one side of a non-porous polymeric membrane while a vacuum or gas purge is applied to the other side. The components in the liquid stream sorb into the membrane, permeate through it and evaporate into the vapour phase (Lipski and Cote, 1990; Peng et al., 2003).
  • Alternative adsorbents— A study reported that an adsorbent impregnated with platinum and titanium dioxide catalyst reduced an influent concentration of BTEX in the range of 24–201 mg/L, achieving removal efficiencies ranging from 96% to 100% (Crittenden et al., 1997). Fibreglass-supported activated carbon filters demonstrated improved kinetics of adsorption for BTEX when compared with the adsorption kinetics of the GAC filter (Yue et al., 2001).
  • AOPs using Fenton's reagent— A pilot-scale system investigated an AOP using Fenton's reagent for the treatment of organic compounds in groundwater. The study demonstrated that initial concentrations of ethylbenzene (500 µg/L), toluene (1700 µg/L), o-xylene (140.0 µg/L) and m/p-xylene (360.0 µg/L) were reduced to non-detectable levels (detection limits not provided) for each contaminant using a molar ratio of hydrogen peroxide to iron of 75 and an average iron concentration of 10 mg/L. The optimum pH was in the range of 3.5–4. It should be noted that a pH adjustment may be needed after this treatment process (Bergendahl et al., 2003).
  • Bioremediation— A bioremediation technology uses microorganisms under fully controlled conditions to degrade the contaminants to less toxic compounds, such as carbon dioxide, methane, water and inorganic salt (Huck et., 1991; Guerin, 2002; Ohlen et al., 2005; Zein at al., 2006; Farhadian et al., 2008). A field-scale aerobic gravity-flow membrane bioreactor consistently degraded TEX concentrations in groundwater (to below 1 µg/L), achieving removal efficiency of greater than 99.9%. The reactor was operated for 6 months with an influent flow rate of 5 gpm (0.31 L/s). The influent concentrations of 508.24 µg/L for toluene and 235.96 µg/L for ethylbenzene were degraded to 0.11 µg/L and 0.04 µg/L, respectively. A biodegradation of greater than 99.9% was achieved for each xylene isomer. The influent concentrations of o-, m- and p-xylene of 431.61, 638.07 and 716.59 µg/L were reduced to 0.07, 0.14 and 0.1 µg/L, respectively (Zein et al., 2006). Laboratory experiments were conducted with two bioreactor configurations: a submerged fixed film reactor (SFFR) and a fluidized bed reactor (FBR), and both were reported to be effective for the biodegradation of TEX in spiked groundwater. Both bioreactors were aerated throughout the experiments to provide oxygen to the attached aerobic microorganisms on activated carbon support medium. The hydraulic retention time was 32.6 hours for the SFFR and 26.1 hours for the FBR (Guerin, 2002). The initial toluene concentration of 0.19 mg/L was biodegraded in the two reactors, achieving a 99.4% reduction rate. The initial ethylbenzene concentration of 0.015 mg/L was reduced by 97.6% in the SFFR reactor. A 97.6% degradation of ethylbenzene was also achieved with the FBR reactor with an initial concentration of 0.018 mg/L. Each of the reactors achieved a biodegradation of 99% for total influent xylene concentrations ranging from 0.23 to 0.26 mg/L (Guerin, 2002). Laboratory experiments demonstrated that the use of activated carbon as a biomass support in a fluidized bed reactor produces a system in which both adsorption and biodegradation affect the BTEX removal in groundwater. During the start-up period BTEX is removed primarily by adsorption. Once the biofilm is established and steady-state conditions are reached, the removal of BTEX is dominated by biodegradation (Voice et al., 1992).

7.2 Residential scale

Generally, it is not recommended that drinking water treatment devices be used to provide additional treatment to municipally treated water. In cases where an individual household obtains its drinking water from a private well, a private residential drinking water treatment device may be an option for decreasing TEX concentrations in drinking water.

Health Canada does not recommend specific brands of drinking water treatment devices, but it strongly recommends that consumers use devices that have been certified by an accredited certification body as meeting the appropriate NSF International (NSF)/American National Standards Institute (ANSI) drinking water treatment unit standards. These standards have been designed to safeguard drinking water by helping to ensure the material safety and performance of products that come into contact with drinking water. Certification organizations provide assurance that a product conforms to applicable standards and must be accredited by the Standards Council of Canada (SCC). In Canada, the following organizations have been accredited by the SCC to certify drinking water devices and materials as meeting NSF/ANSI standards (SCC, 2014):

An up-to-date list of accredited certification organizations can be obtained from the SCC (2014).

A number of certified residential treatment devices from various manufacturers are available that can remove TEX from drinking water. The certified devices for removal of TEX from drinking water rely on adsorption (activated carbon) and reverse osmosis (RO) technologies.

Filtration systems may be installed at the faucet (point of use) or at the location where water enters the home (point of entry). Point-of-entry systems are preferred for the reduction of VOCs such as TEX, because they provide treated water for bathing and laundry as well as for cooking and drinking. This will reduce the potential for VOC exposure through inhalation.

Where certified point-of-entry treatment devices are not available for purchase, systems can be designed and constructed from certified materials. Periodic testing by an accredited laboratory should be conducted on both the water entering the treatment device and the water it produces to verify that the treatment device is effective. Devices can lose their removal capacity through usage and time and need to be maintained and/or replaced. Consumers should verify the expected longevity of the components in their treatment device as per the manufacturer's recommendations.

RO systems should only be installed at the point of use as the water they have treated may be corrosive to internal plumbing components. This system requires larger quantities of influent (incoming) water to obtain the required volume of drinking water, as the systems reject (waste) part of the influent water. A consumer may need to pretreat the influent water to reduce the fouling and extend the service life of the membrane.

7.2.1 Toluene

The certified devices for the removal of toluene from drinking water that rely on an adsorption (activated carbon) technology can be certified either specifically for toluene removal or for the removal of VOCs as a group, using surrogate testing.

For a drinking water treatment device to be certified to NSF/ANSI Standard 53 for the reduction of toluene only, the device must be capable of reducing an average influent concentration of 3.0 mg/L to a maximum effluent concentration of 1.0 mg/L (NSF/ANSI, 2013a). For a drinking water treatment device to be certified to NSF/ANSI Standard 53 by surrogate testing, the device must be capable of reducing an influent toluene concentration of 0.078 mg/L to a maximum product water concentration of 0.001 mg/L (NSF/ANSI, 2013a).

Although no certified residential treatment devices using reverse osmosis are currently available for the reduction of toluene in drinking water, toluene is included in NSF/ANSI Standard 58 (RO). For a drinking water treatment device to be certified to NSF/ANSI Standard 58 by surrogate testing, the device must be capable of reducing toluene from an influent concentration of 0.078 mg/L to a maximum concentration of 0.001 mg/L (NSF/ANSI, 2013b).

7.2.2 Ethylbenzene

The certified devices for the removal of ethylbenzene from drinking water that rely on an adsorption (activated carbon) technology can be certified either specifically for ethylbenzene removal or for the removal of VOCs as a group, using surrogate testing.

For a drinking water treatment device to be certified to NSF/ANSI Standard 53 for the reduction of ethylbenzene only, the device must be capable of reducing an average influent concentration of 2.1 mg/L to a maximum effluent concentration of 0.7 mg/L (NSF/ANSI, 2013a). For a drinking water treatment device to be certified to NSF/ANSI Standard 53 by surrogate testing, the device must be capable of reducing an influent ethylbenzene concentration of 0.088 mg/L to a maximum product water concentration of 0.001 mg/L (NSF/ANSI, 2013a).

Although no certified residential treatment devices using reverse osmosis are currently available for the reduction of ethylbenzene in drinking water, ethylbenzene is included in NSF/ANSI Standard 58 (RO). For a drinking water treatment device to be certified to NSF/ANSI Standard 58 by surrogate testing, the device must be capable of reducing ethylbenzene from an influent concentration of 0.088 mg/L to a maximum concentration of 0.001 mg/L in the treated water (NSF/ANSI, 2013b).

7.2.3 Xylenes

The certified devices for the removal of total xylenes from drinking water that rely on an adsorption (activated carbon) technology can be certified either specifically for total xylene removal or for the removal of VOCs as a group, using surrogate testing.

For a drinking water treatment device to be certified to NSF/ANSI Standard 53 for the reduction of total xylenes, the device must be capable of reducing an average influent concentration of 30.0 mg/L to a maximum effluent concentration of 10.0 mg/L (NSF/ANSI, 2013a). For a drinking water treatment device to be certified to NSF/ANSI Standard 53 by surrogate testing, the device must be capable of reducing an influent concentration of total xylenes of 0.07 mg/L to a maximum product water concentration of 0.001 mg/L (NSF/ANSI, 2013a).

Although no certified residential treatment devices using reverse osmosis are currently available for the reduction of xylenes in drinking water, xylenes are included in NSF/ANSI Standard 58 (RO). For a drinking water treatment device to be certified to NSF/ANSI Standard 58 by surrogate testing, the device must be capable of reducing total xylenes from an influent concentration of 0.07 mg/L to a maximum concentration of 0.001 mg/L (NSF/ANSI, 2013b).

Report a problem or mistake on this page
Please select all that apply:

Thank you for your help!

You will not receive a reply. For enquiries, contact us.

Date modified: