POINT OF USE GRANULAR ACTIVATED CARBON FILTERS (POU) EFFICIENCY FOR ATRAZINE REMOVAL

equipment, in order to obtain kinetic parameters, according to Freundlich equation. The parameters were inserted into a design model used in a complementary phase of the study, which evaluated the efficiency of filter prototypes in the removal of the contaminant in dynamic conditions. The results showed that the prototype filter, constructed from 90.0 g of coconut coal, which presented the highest efficiency for the removal of target pollutant in the preliminary trials, were able to reduce average 57.2% of atrazine after 1000.0 liters filtered, under predetermined operating conditions: flow rate 100 L.h-1 in continuous test. The filter prototypes were designed from the following Freundlich isotherms coefficients: qmax = 40.4 ng atrazine / mg carbon (maximum adsorption capacity); 1/n = 0.5138 and KF = 1.10, with correlation coefficient of 0.9960.


Introduction
Household filters have aroused a growing interest in consumers from many countries lately. These devices, containing technology to improve the potable water quality from municipal sources, are usually composed of vegetable-based granular activated carbon (GAC) and enclosed in a plastic sump. Such devices are sized to fit into a certain flow rate range, considered adequate to several home applications and into a certain hydraulic pressure range, both established in the present local technical standards.
The population growth and concentration in Brazilian cities, the rising environment degradation, as well as the poor treatment of domestic and industrial wastewater, contaminating water sources, are the main reasons for people's lack of confidence about the water supplied to their homes. Additionally, the contamination of water sources by traces of chemical agents is important to highlight. These chemical agents are of anthropogenic nature and are present in the formulation of a variety of products and supplies available in the modern market. They are known as "emerging contaminants", many of them considered "endocrine disruptors", like atrazine, for instance; a substance largely employed in the agriculture as herbicide and frequently conveyed by the rain to the water sources, used later for the public supply through the municipal water treatment plants.
Atrazine quantification in potable water has been studied by many authors: Loos et al. (2007) verified its occurrence at concentrations between 3,0 and 16,0 ng/L; Verliefde et al. (2007) quantified 30 ng/L in Netherlands and Garcia-Ac et al., (2009) measured concentrations between 15 e 28 ng/L in potable water samples collected in Canada.
Other authors have reported atrazine in higher levels: Benotti et al., 2009 detected the contaminant at 990 ng/L in potable water in the US; Montagner (2011) detected it at a concentration from 8,3 up to 687 ng/L in Brazil;Rinsky et al., (2012) reported it in the US at 8.900 ng/L.
On the other hand, authors like Yue (2006) reported the contaminant removal in deionized water, at a level of 97 and 84%, by means of carbon in fixed bed with mass values 77 and 136 grams, respectively, at a flow rate 2,4 L/h. Atrazine has been chosen for this study for being a significant component in several agrochemicals, largely employed in crops like sugar cane, soybean, corn and coffee. Being agriculture one of the major pillars of the Brazilian economy, representing approximately 23% of the GDP in 2011, according to Delgado (2013), atrazine application and concentration should be proportional to the areas used for agriculture. The compound and its metabolites reach the water sources through surface runoff. In the potable water, it remains as traces, micro and nano grams per liter, due to the operations and processes used in conventional municipal water treatment plants being inefficient or not suitable. The Brazilian law establishes 2 µg/L as a maximum permitted concentration in the potable water for this substance.

Methods
The study was divided into two sequential steps: first step, static and second step, dynamic. The static step was carried out on a jar-test equipment, with the application of different amounts of granular activated carbon and two types of water: deionized and potable water, according to the Brazilian legislation (Portaria n° 2914/11 -Ministério da Saúde), both intentionally added to standard atrazine, at a concentration 1000 ng/L. The objective of the first step was to compare the adsorption process in both types of water and to adjust experimental data to Freundlich kinetic equation for potable water. They would be later on used to size the carbon beds of the filters evaluated in the dynamic step. Such tests represent the usual method to analyze water and wastewater treatment processes and to obtain many kinetic parameters related to the action of adsorptive agents when removing the contaminants. MOUSSAVI (2010); SURESH (2010); LIMA (2014); PRAMANIK (2015) The dynamic step aimed the evaluation of atrazine removal from potable water with prototype filters, dynamically operated at a flow rate of 100 L/h. This step was started with the quantification of the granular activated carbon mass to be utilized in the prototypes considering the model proposed by Di Bernardo (2005). The model yielded an approximate mass of 30 grams for a complete removal of the contaminant at the flow rate used in this study. Considering this mass as a reference, prototypes were prepared with three different amounts of carbon: 30.0, 60.0 and 90.0 grams; corresponding to the value determined by the study, twice this value and three times, respectively. During the execution of this step, atrazine was quantified in filtered water samples collected in the beginning of the test and after a throughput of 8.0, 250.0, 500.0 and 1000.0. In attempt to be consistent with the previous step, room temperature was kept constant at 24.0±1.0°C by the HVAC equipment.
For atrazine quantification on the samples collected from the filter inlet and outlet, a preconcentration technique was necessary, with a 2,000 times scale, by means of solid phase extraction with cartridges, since the contaminant concentrations were meager, (ng/L). The preparation of samples for the analyzes via HPLC (High Performance Liquid Chromatography) consisted of extraction with methanol, evaporation and elution with analyte. The extraction process was performed with extraction cartridges composed of the phase octadecyl silica (C18) with graphite carbon. Several scientific publications mention this technique like ATHANASIOS (2013)
The GAC adopted at the study had the physicalchemical properties described at Table 1. Atrazine removal efficiency results for coconut carbon in deionized and potable water are shown at Table 2. The initial theoretical concentration was 1000,0 ng/L. The residual concentration presented correspond to an average from two values (i.e., 1 repetition) considering: jar-test equipment at constant room temperature, five types of carbon between 0 (blank) and 100.0mg  Table 2: Kinetic data obtained from atrazine reduction tests in deionized and well water, with free chlorine at 2.0 mg/L at an average temperature of 25.0±0.5°C, pH of 6.6 and contact time 24 hours Figure 1 shows Freundlich equation linearization graphs and Table 3 shows the parameters obtained from the model. Tests were performed with deionized and potable water with chlorine at 2.0mg/L Figure 1: Graphs plotted from experimental data, adjusting the Freundlich equation to atrazine in deionized and potable waters with chlorine at 2.0mg/L, tested with coconut granular activated carbon. A comparative analysis of Freundlich equation parameters obtained for atrazine indicate an increase in the parameter "1/n" and decrease in the parameters "k" and qmax, when deionized water is replaced by potable water. The parameter "k" was reduced from 3.57 for deionized water to 1.10 (ng/mg)(L/ng) 1/n for potable water. This can be directly associated to the presence of interferents, such as chlorine and its byproducts, natural organic matter, salts and metallic elements which, somehow, try to compete with atrazine for carbon active sites, in the adsorption process. A reduction in the parameter qmax was also observed for potable water compared to deionized water, from 61.8 to 40.4 ng of atrazine per mg of carbon.
Other authors have also reported the same reduction in the maximum atrazine adsorption capacity, when the adsorption studies migrate from deionized water to natural waters. Knappe (2009) observed a reduction in the maximum adsorption capacity for atrazine, when the test water was changed from deionized to natural, containing organic matter (water from River Sena -France). In one of his experiments the parameter "k" from the Freundlich equation changed from 2.46 to 0.898 (mg/g). (L/µg) 1/n , same behavior observed in this study. The author assigns the reduction in the adsorption capacity to the competitive coadsorption of the pores by compounds in the organic matter quantified in the water. Li (2003) studied the competition between atrazine and 1,4-dichlorobenzene in the presence of natural organic matter in attempt to evaluate its interference on atrazine adsorption levels, utilizing Freundlich Isotherm for different types of coconut granular activated carbon. A 74% adsorption capacity reduction was noted for atrazine and assigned to the competition of the compounds in the micropores. According to the author, if the isotherms obtained for water with organic matter are parallel to those plotted for distilled water, there is a small interference due to pore blocking by the organic matter in the adsorption process.
Pelekani e Snoeyink (2000) studied the removal of the pesticide Alaclor from water with hulmic acid by the application of granular activated carbon. There was a 31% decrease in the constant "k", when compared to the same test performed with distilled water. Ultimately, Coelho (2012) also associated the reduction in the Freundlich parameter "k" to the competition of organic matter and atrazine molecules for micropores. It changed from 104.71 to 74.13 [(mg.g-1).(L.mg-1) 1/n ] when the test changed from distilled to natural water.
Dynamic step:  The values of atrazine removal efficiency noted (percentage, relative to the initial concentration) at the three amounts of carbon in the prototypes can be considered consistent, since the incremental carbon mass caused an increasing atrazine concentration reduction at the filter outlet samples. Among the three alternatives studied, the most efficient for contaminant removal was that containing 90,0g of carbon for having shown superior to the other prototypes, in all monitored points.
The data herein show that the utilization of point of use household drinking water filters can represent a protection barrier for the municipal water users to the casual presence of the contaminant atrazine, considered an endocrine disruptor.   Chemically activated carbon on a fiberglass substrate for removal of trace atrazine from water Journal of materials chemistry v. 16, p. 3375-3380, 2006