Fluoride Contamination in Groundwater and Its Removal By Adsorption and Electrocoagulation Processes

2.1. Study area

The location map of studied area including the situation, location and coordinates of the Sarım and Karataş villages in Sanliurfa is given in Figure 1. The drinking water of the villages within the study area was generally supplied from the wells at 100-150 meters depth which were drilled in the last 10-15 years by the government.

2.2. Methods

2.2.1. Electrocoagulation The experimental set-up includes electrodes, reactor, magnetic stirrer and power supply. Seven Al plate electrodes (purity 99.5%) were used in the study. The distance between both electrodes was 10 millimeters. They were connected to a direct current power unit (Rigol DP832A model, 30V, 3A). Groundwater samples were taken from Sarım village. The characteristics of the groundwater and EC reactor are given in Table 1 and 2, respectively. Fluoride removal was performed by electrocoagulation and adsorption method on the water sample. 1252

Figure 1. The location map of the studied area Table 1. Physicochemical features of the groundwater in the studied area Parameter pH Electrical conductivity (µS/cm) Temperature (ºC) F- (mg L-1) Cl- (mg L-1) NO3- (mg L-1) SO42- (mg L-1)

In each experiment, the electrolytic reactor contained 1500 mL water. A magnetic stirrer was used at 250 rpm (IKA RH basic 2) in the EC reactor. EC tests were carried out by batch method. The oxide lamina on the Al electrode surface areas was removed by the solution of the HCl (37%) and the hexamethylenetetramine aqueous solution (3%) for 5-minute immersion before each trial. They were then assembled the reactor. The voltage and current was set to a desired value (5V-1A) through adjustment of the power supply and the experimental study was started at room temperature (25 ºC). The duration of the experiment was 270 min and samples were taken for 30 min intervals. The 0.45 µm syringe filters were used to filter the samples. Initial and final pH and conductivity for each sample was measured with Portable Hach-Lange HQ40d multimeasurement device. At the end of the experiment, they were cleaned with distilled water to eliminate the residuals on the electrode faces. The electrochemical reactor assembly is shown in Figure 2. EC involves in situ generation of coagulants by electrocoagulation of aluminum electrodes. Aluminum cations are generated the anode (1) and hydrogen gas is evolved at the cathode (2). [39]. Anode: Al(s) → Al(aq)3+ + 3e-

During the dissolution of Al at the anode various aqueous aluminum species are produced. The aluminum cations are transformed to polymeric species and form Al(OH)3 precipitate: 1253

The fluoride removal by EC was carried out by means of a chemical substitution in which Freplaced OH- group from Al(OH)3 flocs [10].

Figure 2. The set-up used in the study Table 2. Typical features of electrocoagulation process Electrodes Material Shape Size (millimeters) Thickness (millimeters) Number Connection mode Effective electrode face area (cm2)

Aluminum Rectangular plate 90 × 200 3 7 Monopolar parallel 648

Reactor Type Mode Dimensions (mm) Volume (L) Used water volume (L) Stirring mechanism and rpm

Glass Batch mode 100 × 100 × 250 2.5 1.5 Magnetic bar and 250 rpm

2.2.2. Adsorption Adsorbent used in the adsorption study was clay mineral as a local material procured from Yarımtepe village, Sanliurfa. The tests were executed by Batch method. The 0.45 µm syringe filters were used to filter the samples. Fluoride analyses were carried out by Hach-Lange HQ40d 1254

multimeter device using TISAB method [40]. Physicochemical features of the clay minerals used in adsorption tests shown in Table 3. Table 3. Physicochemical features of the clay minerals used in adsorption tests Clay-Yarımtepe Content SiO2 Al2O3 Fe2O3 MgO CaO Na2O K2O

Concentration (%) 30.506 7.641 5.603 6.285 23.654 0.069 2.019

The compounds of Al2O3 and Fe2O3 located in the clay minerals are thought to have increased adsorption effect. The certain quantity of fluoride adsorbed, and percent of fluoride adsorption was calculated by means of the equation below. 𝐶𝑠 = (𝐶𝑜 − 𝐶𝑒) × 𝑉 ⁄𝑊 (4) And % 𝐴𝑑𝑠𝑜𝑟𝑝𝑡𝑖𝑜𝑛 = (𝐶𝑜 − 𝐶𝑒) × 100⁄𝐶𝑒

Where Cs is the adsorption capacity (mg g ) in the solid phase at equilibrium; Co, Ce are initial and equilibrium concentrations of fluoride (mg L-1), respectively. V is the volume of the aqueous phase and W is the mass (g) of adsorbent used in the experiments. A preparatory work kinetic experiment was performed to set up the equilibration time for fluoride adsorption. Tests were made with double sampling in 50 mL polypropylene test tubes. Clay mineral samples (0.2 g) were shaken with 40 mL of fluoride solutions having 5 mg L-1 of initial concentration on a reciprocating shaker with 150 rpm for 1-2-3-4-5-11-24 hours to acquire the equilibration time at room temperature and then supernatant which was filtered from syringe filtrate was poured carefully and analyzed for the residual fluoride. Adsorption data were analyzed using linearized form of the Langmuir (Equation 6) and Freundlich isotherms (Equation 7): (𝐶𝑒⁄𝐶𝑠) = (1⁄𝑄𝑜𝑏) + (𝐶𝑒⁄𝑄𝑜) (6) 𝑙𝑜𝑔 𝐶𝑠 = 𝑙𝑜𝑔 𝐾𝑓 + 1⁄𝑛𝑓 ∗ 𝑙𝑜𝑔 𝐶𝑒

where b is the Langmuir constant and Qo is the maximum amount of adsorbent that can be adsorbed. Plotting Ce/Cs versus Ce, the slope is 1/Qo and the intercept is 1/Qob. Kf and 1/nf are the empirical Freundlich constants. By plotting log Cs (amount adsorbed to adsorbent, mg g-1) versus log Ce (amount of fluoride in solution, mg L-1), 1/nf is calculated as the slope of the curve and the intercept is equal to log Kf [41].

3.1. Dental fluorosis cases in southeastern of Turkey

Atasoy and Yesilnacar [4] determined the high F in Sarım and Karataş groundwater and detected the dental fluorosis cases in the villages. It was reported that the dental screening was carried out by a dentist from Sanliurfa Provincial Health Directorate in a project carried out in 1255

2010 (Figure 3) [17]. The amount of fluoride in ground water was between 1 and 4 mg L -1. Sarım and Karataş samples were higher than the values reported by WHO (0.5-1.5 mg L-1). High fluoride concentration in groundwater caused the dental fluorosis in the region. As a continuation of this study in order to investigate whether the problem persists, ground water samples were taken from the same wells again in 2018. According to the results of the groundwater analysis, the highest F concentration was found 2.35 mg L-1 and this value was still above the permissible limit. Additionally, a female student who underwent a dental screening 10 years ago was reached and dental screening was repeated. It was found that the dental fluorosis problem was still present for the same student and the others (Figure 4) [42].

Figure 3. Some images from the dental examinations (year 2010)

The girl student was 11 years old The same one is 18 years old in in 2010 2018

3.2. Fluoride removal by electrocoagulation

The performance of EC process was estimated for the removal of fluoride from natural groundwater sample gathered from Sarım and Karataş villages in Sanliurfa. pH and conductivity of groundwater were 8.29 and 486 µS/cm, respectively. Other parameters were mentioned in Table 1. Electrocoagulation process successfully decreased the fluoride concentration below the maximum contaminant level of 1.5 mg L-1 (99.4% removal). F removal rate increased with the operation time due to the arising Al+3 ions by aluminum anode during the process time. Table 4 showed the EC operating conditions and energy consumption. According to the results, the energy consumption increased with the decreased conductivity [43,44]. The lowest energy consumption value (0.13 kWh/m3) corresponded to the highest conductivity (469 µS/cm). In this work, the maximum energy consumption was found lower than the other F removal studies of EC [44-47]. The pH of a solution is one of the most significant parameters in EC process [48,49]. Especially in this study pH effected the removal of fluoride and performance of electrochemical process. Controlling of water pH was very difficult during the EC process due to its instability throughout 1256

the experiment. Both of pH and conductivity were measured in certain time intervals as shown in Figure 5. Both decreased while the F removal rate was increasing. At the end of the operation time, pH and conductivity were measured as 7.7 and 291 µS/cm, respectively. Table 4. The EC operating conditions Parameter Electrode material Current density (mA cm-2) Temperature (ºC) Initial pH Final pH Initial electrical conductivity (µS cm-1) Final electrical conductivity (µS cm-1) Operating time (min) Energy consumption (kWh m-3) Energy consumption per kg fluoride (kWh kg F-1)

Simultaneous fluoride and ammonia from semiconductor wastewater [44], arsenic from groundwater [24] , chromium (VI) from pretreated acidic semiconductor effluents [50], iron from contaminated drinking water [51] and hydrated silica from groundwater [52] also removed with EC processes successfully. López-Guzmán et al. [53] investigated the removal efficiencies in simultaneous presence were 85.68% for fluoride and approximately 100% for arsenic. Un et al. [49] determined fluoride removal from water and wastewater and the removal efficiency was 97.6%.

Figure 5. (a) Change in pH and F removal rate during reactor operation time, (b) Change in conductivity and F removal rate during reactor operation time. Alcacio et al. [54] investigated arsenic removal from groundwater collected at a depth of 200 m in Central Mexico by a continuous filter press EC reactor. Arsenic removal efficiency and energy consumption were 89.6% and 0.89 kWh m-3, respectively. Can et al. [55] investigated effect of some operational parameters on the arsenic removal by the batch EC reactor using iron plate electrodes. Arsenic removal efficiency and the energy consumption were 99.98% and 0.52 kWh m-3. Rozales et al. [52] determined arsenic removal from groundwater and the energy 1257

consumption of during the EC tests was 3.6 kWh m-3. Uzun and İpek [56] determined the total energy consumption in the nitrate treatment of the groundwater was 0.4 kWh m-3. Xu et al. [57] investigated Simultaneous removal of cadmium, zinc and manganese using electrocoagulation and the corresponding energy consumption was 2.73 kWh m-3.

3.3. Fluoride adsorption on the clay mineral

The fluoride adsorption on the clay mineral depending on the contact time is shown in Figure 6. At the beginning, the amount of adsorbed fluoride raised with enhancement contact time. However, after some time, it approached a nearly constant value and reached an equilibrium within 2 hours. 36% of the initial fluoride concentration was adsorbed at the equilibrium time. It is thought that the equilibrium period was prolonged due to the thrust between the negatively charged clay particles and F ions. There was no significant change in adsorption after the 2nd hour. Adsorption on the clay mineral was fast initially then slow until a certain equilibrium time. Initial reactions are associated with diffusion of contaminants to the surface of the adsorbent. 40 35 % Adsorption

Figure 6. Time-dependent fluoride adsorption percentages on the clay mineral In the description of adsorption data, Freundlich and Langmuir equations were used. The values of Freundlich adsorption constants of clay mineral are given in Table 5. The values of 1/n f for clay mineral was <1.0 indicating the L type isotherm. Langmuir equation also displayed a good fit to experimental data for fluoride as denoted L type isotherm. Mandal and Mayadevi [58] and Sundaram et al. [59] also reached that the fluoride adsorption in clay mineral fitted well with both Langmuir and Freundlich equations. Table 6 shows the Langmuir adsorption constants for the F adsorption on the clay mineral. A small value of 1/n f (<1.0) indicated a weaker physical bond between adsorbate and adsorbent representing the heterogeneous adsorbent surface. Low value for Kf (0.14) notified the low F adsorption on the clay mineral. It can be accounted by negatively charged places on the clay which made the material infertile adsorbent for anionic F. This means that negatively charged pollutants generally exhibited a few sorption on the clay materials. Table 5. Freundlich constants for the fluoride adsorption on the clay mineral Isotherm Freundlich Isotherm

Table 6. Langmuir constants for the fluoride adsorption on clay and bauxite minerals Adsorbent Clay mineral

As a result, the high amount of F in groundwater in Sarım and Karataş revealed the dental fluorosis problem in the region. The removal of fluoride by electrocoagulation and adsorption methods were investigated to produce a short-path solution. Low F adsorption was found on the clay mineral. Clay material is thought to be an inefficient adsorbent for the negatively charged fluoride ions. Despite the low F removal by adsorption, about 99.4% removal efficiency was achieved in EC process at the end of 270 min. Although the electrocoagulation seems to be a high cost method, solar energy can be a good alternative to minimize the energy costs. Therefore, the electricity requirement can be supplied by photovoltaic solar panels during the long period of sunny days in the region in a year.