Approximately 600 - 700 DBPs have been reported in the literature as the major disinfectants used (chlorine, ozone, chlorine dioxide, chloramines) as well as their combinations. DBPs are the result from the reactions between organic and inorganic matter in water with chemical treatment agents during the water disinfection process. The THMs are halogen-substituted single carbon compounds with the general formula CHX₃, where X may be fluorine, chlorine, bromine or iodine or combination thereof. There are four of these combinations: chlorodibromomethane (CHClBr₂), bromoform (CHBr₃), chloroform (CHCl₃) and bromodichloromethane (CHBrCl₂) (Hua et al. 2010, Karim et al. 2011). Chemically, the formation of THMs (specifically compounds based on chlorine) occurs in the chlorination process when the chlorine (Cl₂) is added to water and is dissolved in the medium of interest. The Cl₂ molecule (with zero oxidation state) rapidly undergoes a disproportionation reaction in which one of the chlorine atoms is reduced to chloride Clˉ and the other is oxidized to Cl⁺, a highly reactive species that rapidly displaces a proton from water to form hypochlorous acid (HOCl). The overall result is the hydrolysis of the Cl₂ molecule.

Cl₂ + H₂O ↔ HCl + HClO

K_a (20˚C) = 3.3 × 10⁴

Hypochlorous acid is a weak acid that undergoes partial dissociation in an alkaline pH to form the hypochlorite ion (OCl⁻) [2] .

HOCl + H₂O ↔ H⁺ + OCl⁻

K_a = 2.7 × 10⁻⁸

The addition of hypochlorite ion produces a dissociation to form OH⁻ [2] .

OCl⁻ + H₂O ↔ HClO + OH⁻

K_hb = 3.7 × 10⁻⁷

The total concentration of Cl₂, HOCl and OCl⁻ is referred as free available chlorine.

DBPs are the result of a reaction between a disinfecting agent (chemical or physical) and a precursor chemical in the source water. The factors affecting formation of DBPs are: chlorine residual, temperature, contact time, pH, and NOM. Chlorine dose from 2 to 16 mg/L will lead to an increase of reactions, therefore causing an increase of HAAs and THMs formation [3] . The residual chlorine is define as the sum of the concentration of Cl₂, HOCl and OCl⁻ and is referred as free available chlorine. Seasonal variations are related to variations of raw water quality, water temperature, and treatment plant operations. In fact, changes in operational parameters during water treatment, for example, chlorine dose and pH adjustment, are generally a response to changes in both water temperature and raw water quality [4] . The contact time to form THMs or other DBPs should be less than 5 h (90%) and levels off at 96 hours [5] . The probability of THMs formation increases in pH range from 7.0 to 9.0, no positive effect at pH > 9.5 [5] . NOM is composed of an heterogeneous mixture of humic substances (HS), carboxylic acids, proteins, amino acids, polysaccharides, hydroxyl acids and other common components of surface waters and are primarily of concern because it is a precursor to the formation of DBPs in drinking water [6] [7] .

From 2004 to 2011, THMs concentrations were found in CWTP and at different sampling sites across the water distribution system (Piñas Abajo and Río Hondo schools and Comerio Family Health Center). The present research has been developed with the aim to provide new insights on this situation and to suggest possible solutions. The main objectives of this research can be summarized as: 1) to quantify the THMs concentration in CWTP, Piñas Abajo and Río Hondo schools, Comerio Family Health Center and the residence of Vázquez Ortiz family; 2) to evaluate different nanomaterials to be used as potential adsorbents; and 3) to suggest the use of new nanomaterials to produce drinking water with sufficient quality. According to data of 2010, the municipality of Comerio has an estimated population of 20,778. The CWTP owned and operated by PRASA, supplies drinking water to an estimated 80% of the Comerio township population. The sampling sites were CWTP, Piñas Abajo and Río Hondo schools, and the Comerio Family Health Center. In addition, a house was selected for comparison purposes (Vázquez Ortiz residence). The water samplings were collected every two weeks for a period of one year. The USEPA 502.2 method was the procedure used to evaluate the THMs concentrations [8] . The water sampling collection began from February 2011 through February 2012.

The average pH value was 7.84. The lowest pH value (7.06) was found in CDT on January 13, 2012, while the highest pH value (8.69) was found in Río Hondo school on 07/01/2011. The average temperature value was 25.7˚C. The minimum temperature value was 20.2˚C and it was found in Piñas Abajo school on February 25, 2011. The highest temperature value was 30.5˚C and it was reported in the same sampling site on August 12, 2011. The average residual chlorine value was 1.06 mg/L. The lowest value was 0.06 mg/L on 04/08/2011 in CWTP and the highest value was 2.20 mg/L found in CDT on 3/25/2011. The average value of TOC was 1.75 mg/L. The lowest value was 0.95 mg/L in CWTP on 08/26/2011 and the highest value was 14.26 mg/L in Vázquez Ortiz residence on April 8, 2011.

The average THMs value found in this research was 0.0545 mg/L. The highest value was 0.1101 mg/L on 02/25/2011 in CDT sampling site. The lowest value was 0.0854 mg/L on 08/12/2011 in the same sampling site (CDT). THMs formations were found in all sampling sites, moreover, in some sampling sites the THMs formation was detected only one time specifically in (Vázquez Ortiz residence, CWTP, and Río Hondo and Piñas Abajo schools). The major THMs detection in water samples was found in CDT sampling sites for ten times, (Table 1).

Different physical approaches can be evaluated to reduce the formation of THMs and to remove them from drinking water. Three different nanostructured materials were tested: multiwalled carbon nanotubes (MWCNTs), zeolites (mordenite), and graphene flakes with the purpose to identify the appropriate nanostructured material to adsorb THMs on their surface. CNTs are being explored as replace of charcoal (activated carbon) because this nanomaterial effectively removes metal and organic contaminants from water [9] . MWCNTs contain between 2 and more than 30 concentric cylinders placed around a common central hollow with spacing between the layers close to that of the interlayer distance in graphite (0.34 nm). Other characteristics are: Electrical conductivity, optical activity, and mechanical strength.

Zeolites are open-framework structures composed of tetrahedral SiO₄ and AlO₄, and they are widely used as solid acid/base catalysts. Mordenite (MOR) occupies a special place among zeolites used in adsorption and catalysis thanks to its acid resistance, thermal stability, and acid properties. MOR is a zeolite that consists of two types of channels; one is composed by 12-member rings (0.67 × 0.7 nm²) and the other, which comprises the side pockets to the main channels, is composed of 8-member rings (0.26 × 0.57 nm²) [10] . Graphene is an allotrope of carbon whose structure consists of a single layer of carbon atoms bonded in a hexagonal lattice [11] . It can be viewed as less than 10 layers of graphite. This material has attracted intense scientific interest due to its unique two dimensional structure and remarkable electronic properties [12] . It has a large theoretical specific surface area (2630 m²・g⁻¹) ( [13] , 2011), high intrinsic mobility (200,000 cm²・v⁻¹・s⁻¹), thermal conductivity (~5000 Wm⁻¹・K⁻¹), optical transmittance (~97.7%), good electrical conductivity and high mechanical strength [14] .

With the aim to test the different nanomaterials as potential adsorbers of THMs, four solutions of each THM were prepared. Solutions of chloroform, bromoform, diclorobromomethane and dibrochloromethane with a concentration of 90 ppb were prepared using CO₂-free water and subsequently mixed with the different THMs in an Erlenmeyer flask of 1000 mL. Next, 25 mg of each nanomaterial (MWCNTs, MOR and graphene) were added to each solution. In this step different contact times were measured (30, 60, 120, 180 and 240 minutes, respectively). After completed each contact time, approximately 90 mL of each sample was transferred to a centrifuge tube and the mixture was centrifuged for 15 minutes at 3000 revolutions per minute (rpm). With the aim to remove all possible residues of nanomaterial, the reaction mixture was also filtered by using disk filters of 0.45 μm. Samples were analyzed with a TOC analyzer Phoenix 8000 Kekmar Dohrmann and GS-MS Shimadzu GC-MS-QP5050A purge and trap system. Each sample was run in duplicates and the mean value of each sample was reported.

Figure 1 shows HRTEM (high resolution Transmission Electron Microscopy) images of the graphene used in the present research. As can be seen there, disaggregation apparently was not fully accomplished during the ultrasound treatment in ethanol before observation by HRTEM. In these images, each horizontal line represents a graphene sheet of one atom of thickness.

Mordenite was analyzed by SEM. Figure 2(a) and Figure 2(b) show the crystalline structure with different orientations and aggregations of MOR obtained at low magnification. Figure 2(c), obtained at 50000× corresponds to a single crystal of MOR.

MWCNTs were also characterized by SEM. As can be seen in Figure 3(a), at low magnification agglomerations of MWCNTs with different shapes and particle size can be observed. At higher magnification these small aggregates can be observed as mixed fibers (see Figure 3(b)). Each of these fibers corresponds to a multiwalled carbon nanotube.

Graphene flakes used in this research showed the highest specific surface area BET with a value of ca. 1850 m²/g. This BET result means that graphene has potentially more adsorption capacity than MOR (420 m²/g) and MWCNTs (585 m²/g), respectively.

Although MWCNTs has more BET surface area than MOR, carbon nanotubes have lesser capacity of adsorption (with some contaminants such as THMs) than MOR. MOR was the nanomaterial with the lowest BET surface area found in this research, with a value of 420 m²/g. Although MOR has a lower surface area in comparison with MWCNTS and graphene, MOR is composed by monodirectional channels that can increase the adsorption into the inner pore structure. On the other hand, the high Si/Al atomic ratio of this material indicates the presence of acid sites that can increase the efficiency of interaction with other compounds (i.e. THMs).

The adsorption of each nanomaterial was measured in different contact times of each THMs. Graphene was found to be the nanomaterial with the highest adsorption capacity for THMs. In the first 30 minutes graphene adsorbed 61.2 ppb of CHCl₃, 66.5 ppb of CHBr₃, 59.1 ppb of Br₂CHCl and 51.2 ppb of BrCHCl₂. After 1 hour, graphene adsorbed 78.2 ppb, 80.2 ppb, 75.7 ppb, and 66.6 ppb of CHCl₃, CHBr₃ Br₂CHCl, and BrCHCl₂ respectively.

The adsorption process in graphene is very fast during the first 120 minutes, where graphene adsorbed the majority of THMs. At contact time of 120 min graphene was able to absorb the following concentrations of THMs: 84.2 ppb of CHCl₃, 85.8 ppb of CHBr₃, 88.3 ppb of Br₂CHCl and 80 ppb of BrCHCl₂. At this contact time is when graphene adsorbed 80 ppb or higher, being this concentration the MCL of THMs in drinking water. The reason for the extremely high adsorption capacity of graphene is possibly due to the high specific surface area in comparison with MOR and MWCNTs. Graphene is formed by flakes of single layers or associations of no more than 6 - 7 sheets (as was observed by HRTEM), being this fact the reason for the high surface area and the high adsorption capacity.

In the first 30 minutes MOR adsorbed 45.2 ppb of CHCl₃, 54.7 ppb of CHBr₃, 44 ppb of Br₂CHCl and 36.1 ppb of BrCHCl₂. The adsorption process of THMs by MOR is very fast during the first 60 minutes, when 70 ppb of CHCl₃, 75.8 ppb of CHBr₃, 66.8 ppb of Br₂CHCl and 57.2 ppb of BrCHCl₂ were adsorbed. In 120 minutes MOR adsorbed 81 ppb of CHCl₃, 85.6 ppb of CHBr₃, 78.2 ppb of Br₂CHCl and 74 ppb of BrCHCl₂. As occurred for graphene, at this contact time (120 min) MOR adsorbed 80 ppb or higher concentrations of CHCl₃ and CHBr₃, being this concentration the MCL of THMs in drinking water. In the case of Br₂CHCl and BrCHCl₂, MOR adsorbed over 80 ppb after 180 minutes of contact time (Br₂CHCl (81.7 ppb), BrCHCl₂ (81.3 ppb)). After 240 minutes the process is completed with a final adsorption of 89.4 ppb (CHCl₃), 92.4 ppb of CHBr₃, 87.5 ppb of Br₂CHCl and 84.9 ppb of BrCHCl₂. MOR is a mono directional zeolite with channels organized in 12 member rings. Due to this particular structure, MOR is able to incorporate pollutants inside their internal channels and also to absorb them on the surface.

In this research MWCNTs was to be found the nanomaterial with the lowest adsorption capacity in comparison with graphene and MOR. During the first 30 minutes, MWCNTs adsorbed 32 ppb, 30.2 ppb, 24 ppb and 20.8 ppb of CHCl₃, CHBr₃, Br₂CHCl and BrCHCl₂, respectively. The adsorption process experienced by this nanomaterial is lower in comparison with graphene and MOR, for example after 60 minutes MWCNTs adsorbed 54 ppb of CHCl₃, 47.7 ppb of CHBr₃, 55.6 ppb of Br₂CHCl and 46.9 ppb of BrCHCl₂. After 120 minutes, MWCNTs does not complete the adsorption of 80 ppb (MCL) in each solution; only 71 ppb of CHCl₃, 69.2 ppb of CHBr₃, 55.6 ppb of Br₂CHCl and 66 ppb of BrCHCl₂ were adsorbed. MWCNTs adsorbed 80 ppb or higher concentrations after 180 minutes of contact time (80 ppb of CHCl₃ and 82.5 ppb of BrCHCl₂). At 240 minutes 84.7 ppb of CHBr₃, and 85.1 ppb of Br₂CHCl were adsorbed by MWCNTs. The results observed using MWCNTs are currently being analysed and further experiments are performed to determine the significance of this unexpected behavior.

The adsorption behavior of THMs by graphene, MOR and MWCNTs is summarized in Figures 4-7, corresponding to the adsorption results of each pollutant on the three nanostructured materials.

In order to complement the adsorption results previously obtained, total organic carbon (TOC) analyses were measured, after different contact times with the nanomaterials, using the same solutions that those used to study

the THMs adsorption. Figures 8-11, show the C/Co ratio, where C is the TOC concentration measured at different contact times, and Co is the TOC concentration measured before addition of the nanomaterial (zero time). At zero time, C/Co is always 1.0 in each THMs solution. In the case of graphene, during the first 30 minutes C/Co was reduced to c.a. 0.9, in presence of each THMs solution. In all cases, after 60 minutes C/Co decreased to 0.7. After 2 hours, C/Co was found to be in the range of 0.4 to 0.5 (Figures 8-11). After 240 minutes, graphene was able to adsorb the majority of the organic matter present in the solution as THMs (C/Co was reduced to ca. 0.2).