Michael Mlwilo Samwel

Performance of Coal in Fluoride Attenuation from Fluoride Polluted Water intended for Domestic Use Muhondwa, J.P ∗ and Samwel, M.M 1 School of Environmental Science and Technology, Ardhi University, P.O. Box 35176, Dar es Salaam, Tanzania ABSTRACT: This study analysed and evaluated the performance of coal along with bone char as the control. The moisture content, volatile content, fixed carbon and ash content of coal was 3.26%, 16.84%, 19.23% and 63.93% respectively. The pH and electrical conductivity (EC) of coal and control [bone char (BC)] were 2.91, 7.54 and 65.6 µS/cm and 31.6 µS/cm respectively. The fluoride concentrations and pH of the water samples obtained from Burko river of Arusha National Park and Usa River were 8.8mg/l- F, pH=8.24; and 5.8mg/L- fluoride, pH=7.8 respectively. The fluoride adsorption increased from 77.27% to 98.18% on coal as particle size decreased from (1-1.4) mm to (0.5-1.4) mm respectively after 3hrs contact time per column. The fluoride removal efficiency increases toward neutral and acidic side. Thee coal reduced fluoride concentration from 8.8 mg/L to 1.58mg/L as compared to BC with 0.72mg/L at breakthrough point within 4hrs service time.; This level was within TZS drinking water standards of 4 mg/L-F. The fluoride removal efficiency between control (bone char) and coal in the column were not statically significant when tested at 95%CI (p=0.05. The maximum bed adsorption capacity of coal towards fluoride was 12.938g/kg as compared to BC (79.05g/kg) char. The coal can be applied in large-scale defluriodation units to attenuate excess fluoride from water for domestic use as validated by both Langmuir and Freundlich adsorption isotherms and One-Way ANOVA at 95%CI (p=0.05). Keywords: Coal, Fluorosis, Water pollution , Arumeru, Arusha INTRODUCTION Fluoride is the most electronegative element which has significance to human health once it is taken at a right proportion. About 37 countries worldwide has affected with fluoride greater than 1.5mgF-/L threshold fluoride level in the drinking water; These includes China, India, Tanzania and Sudan. The accessibility of clean water has developed obliquely dominant to the quality of human life. Presently, two-thirds of our earth is covered by water; however, the unfortunate paradox is that in the next decade, bulk of the human population will lack access to safe drinking water. It was projected that in 2035, there would be a one-third reduction in the per capita drinking water accessibility. The rapid urbanization and industrialization cause more xenobiotic substances like fluoride chemicals and associated organic toxicants being diffused into different spheres of the earth resulting in drinkingwater pollution and water scarcity. This situation is more serious in developing countries, as they are grappling with acute issues related to both scarcity and contamination of drinking water. High fluoride concentrations in drinking water and associated fluorosis issues were stated from China, India, African countries such as Tanzania, South Africa, Kenya, Ghana, and Sudan. The WHO has set 1.5 mg/l as the safe limit of fluoride in drinking water (1); Dental and skeletal fluorosis are the main public health problem in some parts of Tanzania especially to regions along the Great Rift Valley (Shinyanga, Singida, Mara, Mwanza, Kilimanjaro and Arusha) due to the excess consumption of ∗ Corresponding author. Tel.: -; fax: -. E-mail address:- 1 Author: E-mail address:- 1 fluoride polluted water. Bone char defluoridation method though existed for about 28 years in Tanzania, yet fetches low adoption owing to existing socio-cultural differences among the community. Minimization of the fluoride related health problems using material with social acceptability was a motivation behind carrying this research. The utilization of efficient, effective, suitable, environmentally friendly, socially and religious acceptable materials to mitigate fluoride level in water should be used to overcome the fluoride associated problems. To reduce fluoride related problems; this research therefore investigates the applicability of coal under adsorption technology to attenuate fluoride from water for domestic use. Adsorption depends on ions (adsorbate) in fluid diffusing to the surface of a solid (adsorbent), where they bond with the solid surface or are held there by weak intermolecular forces(2, 3). Adsorption studies pointed most important characteristics which determined adsorbent appropriateness for practical application: Selectivity for fluoride ions, cost , particle and pore size compatibility, Regenerability, Adsorption capacity It is well recognised that both adsorbent and adsorbate features determine the adsorption mechanism, which can comprise chemical reactions, electrostatic interactions, ion exchange, surface precipitation, or a combination of them. 2. MATERIALS AND METHODS 2.1 Materials The materials used in this study were candidate material (coal) and control (bone char). The bituminous coal (Plate 4) material (50kg) was bought from Ngaka Coal mine Ltd and transferred from Songea to NDRS- Arusha. Bone charcoal is a porous, blackish, granular material prepared from animal bones (Plate 1). Bone char from cow bone prepared from NDRS was used as received as control in this study. The bone char having 1.4 mm size was used. Those adsorbent materials were used in the adsorption process as receaved and no activation chemical or reagents added. The moisture content, volatile content, fixed carbon and ash content of coal was 3.26%, 16.84%, 19.23% and 63.93% respectively. The coal materials were crushed, grinded (Plate 3) and sieved (Plate 2) to (0.5-1) mm and (1-1.4) mm size. The response of candidate materials were evaluated through bench (batch) and pilot (both continuous flow and discontinuous flow) study by using water samples from the main domestic water sources located at Arumeru district in Anusha, Tanzania (Figure 1). Other materials were instruments, equipment, chemicals and reagents; glassware and plasticwares. Plate 4. The coal Plate 1. Coal grinding Plate 2. Coal sieving 2 Plate 3. The bone char (BC) Figure 1. The community water sources 2.2 Methods 2.2.1 Laboratory analysis In order to safeguard the quality assurance and control during sample collection, preparation and analysis, suitable measures and precautions were taken to ensure consistency of the results. Water samples from Usa river and Burko rivers {from Arusha National Park (ANAPA)} were used in the adsorption columns (Figure 2). The precaution and analytical standards methods for the examination of water and wastewater, 20th edition ,1998- APHA, U.S EPA 2003 and ASTM standards for testing materials were used for validation of the analytical procedures. The ASTM analytical standards of D 388; D 3172; D 3173; D 7174 and D 3175 were used to test the proximate analysis of coal. 2.2.1.1 Physicochemical analysis of water samples The pH, TDS, salinity and electrical conductivity of water samples was determined by pH meter. The pH meter was calibrated and the electrode was rinsed using distilled water and then calibrated by the 3 buffer solution to the pH 7.00. Water sample of 50ml was measured and transferred into a 125 ml flask. The calibrated pH meter electrode was dipped into the flask contained water sample and results recorded. Calcium in water samples was determined by titrimetric, EDTA method. The concentration of fluoride was determined by ion-selective electrode (I.S.E). This has been selected for the determination of fluoride in water because of its practical convenience and reliability. Fluoride standard solution of 1.00ppm and 10ppm, deionized water and Total Ionic Strength Adjustment Buffer (TISAB) for maintaining/buffering pH level above 5 to minimize HF complex formation were used. Figure 2. The schematic diagram of an experimental setup 2.2.1.2 Physicochemical analysis of adsorbent materials The fractions of coal and bone char were prepared after grinding and sieving. Bone char as control was used as received from NDRS. Sieves were shaken mechanically for 10 minutes to provide complete separations. The detached fraction of each coal and bone char were washed by deionized water then each dried at (105 ± 5) °C and left overnight; then allowed to cool in the desiccators before characterization. The determination of pH of materials was done after coal and bone char being grinded and passed through 0.5mm sieve separately; then 20g of the resultant sieved material was dissolved in a 20ml of deionised water. Each suspension of material was shaken separately in the mechanical shaker for two hours; then pH was measured. The coal was prepared for adsorption experiment by changing their physical characteristics of mass and particle size along with the BC. 4 2.2.2 Adsorption isotherms The adsorption isotherms experiments for fluoride removal were performed using continuous monitored columns and batch experiments. The operational conditions were temperature (23.35±0.81) ℃, pH (8.27±0.27), initial fluoride level from ANAPA water sources were (8.77±0.67) mg/L range. Water samples allowed to pass through the fixed bed monolayer columns (1-1.4) mm and (0.5-1.4) mm particle size at constant flow rate of 0.137L/min in continuous and discontinuous flow systems. The column (pilot scale) study was then conducted under fixed bed adsorption system were done with replicates samples (n=3) and the discharges from each column were taken after 3hrs retention time per column per experiment for analysis. According to (4)the efficiency and fluoride adsorption capacity of Bone char, zeolite and coal was known after determining the residual fluoride concentration (Ct) using an equation (i-ii) and adsorption isotherms equations (iii-iv) respectively for each material. 𝑅% = 𝐶𝑜 −𝐶𝑡 𝐶𝑜 × 100% ………………………………………………………………...(i) Whereby: Co =initial fluoride concentration (mg/L), Ct =Final or residual fluoride concentration (mg/L) and R= percentage fluoride removal efficiency Adsorption isotherms show how a solute distribute between liquid and solid phases at the time of equilibrium. They deliver some insight into the adsorption mechanism as well as the surface properties and affinities of the adsorbent. Also, the isotherm data is used to predict the adsorption capacity of an adsorbent while designing an adsorption system. The adsorption capacities determined by using the equation: 𝑄𝑡 = 𝐶𝑜 −𝐶𝑡 𝑝 ……………………………………………………………………………(ii) whereby: 𝐶𝑜 (mg-F/L) is the original concentration 𝐶𝑡 is an equilibrium concentration and 𝑝 (g/L) is the ratio of sorbent mass (mg/L) to sorbent free solution volume (L). The fluoride removal efficiency: Langmuir adsorption isotherm Langmuir isotherm is based on the assumption that there is a finite number of binding sites which are homogeneously distributed over the adsorbent surface. These binding sites bear the same affinity for adsorption of a single molecular layer and there is no interaction between adsorbed molecules. The equation of Langmuir isotherm is characterized as: 𝑄𝑒 = 𝑄𝑚𝑎𝑥 𝐾𝐿 𝐶𝑒 ………………………………………………………………………(iii) 1+𝐾𝐿 𝐶𝑒 Where 𝐶𝑒 (mg/L) and 𝑄𝑒 (mg/g) are the liquid phase concentration and solid phase concentration of adsorbate at equilibrium, respectively. The adsorption capacity, 𝑄𝑚𝑎𝑥 (mg/g) is the amount of adsorbate at complete monolayer coverage and 𝐾𝐿 (L/mg) is the Langmuir isotherm constant that relates to the energy of adsorption. To assess the adsorption capacity for a particular range of adsorbate concentration, the linearized equation was used: The Plot of 𝐶𝑒 𝑄𝑒 𝑉𝑠 𝐶𝑒 ∶ 𝐶𝑒 𝑄𝑒 = 1 𝐾𝐿 𝑄𝑚𝑎𝑥 + 𝐶𝑒 𝑄𝑚𝑎𝑥 Freundlich adsorption isotherm The Freundlich isotherm model is based on the multilayer adsorption of an adsorbate onto the heterogeneous surfaces of an adsorbent. The famous expression for the Freundlich model: 5 1 𝑄𝑒 = 𝐾𝐹 𝐶𝑒 𝑛 …………………………………………………………………………(iv) Where, KF is the Freundlich constant [(mg/g) (L/mg)1/n] related to the bonding energy, and n is the heterogeneity factor. The value of n varies with the heterogeneity of the adsorbent and for favourable adsorption process, n value should be less than unity. The linear form of the Freundlich equation was used to describe adsorption isotherm data. The Freundlich isotherm constants KF and n were calculated from the slope and the intercept of the plot of 𝐿𝑜𝑔10 𝑄𝑒 vs. 𝐿𝑜𝑔10 𝐶𝑒 which is a straight line. The favourable equilibrium constant RL, (𝑅𝐿 = 1 𝐾𝐿 𝐶𝑜 ) was used to show the favourability of the reaction from the Langmuir model. The adsorption state may be linear (RL=1), irreversible (RL=0), favourable (0 1). Freundlich isotherm model; also provided good correlations with the laboratory data. In the Freundlich model the value of 1/n is the adsorption intensity or reactivity at active sites and 1/n > 1 the adsorption is a favourable physical adsorption, when 1/n<1 it is favourable chemisorption, 1/n=1 is constant and 1/n=0 shows no adsorption(5). 3. RESULTS AND DISCUSSIONS 3.1 Physical and chemical characterization Some physical and chemical composition of coal was also declared by manufacturer. The chemical composition of bituminous coal and bone char (BC) as a control; were shown in the table 1 as stated by the manufactures. The physiochemical characteristics (Table 1 and Table 2) of coal and as received bone char analysed at NDRS laboratory. Table 1. The physical characterization Physical characteristics Candidate material Control Coal Bone char EC (µS/cm) 65.6 87.9 pH 2.91 7.54 Particle size 4 (mm) (2-30) (1-1.4) Table 2. The chemical characterization Chemical characteristics 2 Chemical composition as per manufacturer XRD analytical results Analysed in the NDRS Laboratory 4 Coal and Zeolite were sieved to attain the particle size of bone char 3 6 Candidate material 2 Control Coal Bone char SiO2 52.31% 0.89% Al2O3 31.20% 0.44% Fe2O3 2.31% NA CaO 7.52% 49.8% Na2O 0.51% 0.96% TiO2 1.63% NA P2O5 NA 32.90% Calcium 3, Ca2+ (mg/L) 3.1% 24.2% The fluoride concentrations and pH of the water samples obtained from ANAPA and Usa River were 8.8mg/l- F, pH=8.24; and 5.8mg/L- fluoride, pH=7.8 respectively. Water samples from ANAPA streams varied its physical and chemical characteristics (table ) due to variability of weather conditions from the sources; while water samples from Usa river was sampled (50 buckets with 20L capacity) and stored in 1000L tank. The background fluoride concentrations analysed exceeded the WHO, US EPA and TBS permissible limits (Table 3 ). Hence, there is the need for attenuation of fluoride from the water to achieve water quality standards for domestic uses particularly drinking. Table 3 Physical and chemical characteristics water samples for domestic uses Average Initial conditions Drinking Water Standards Water sample Parameters TZS 789:2008 WHO 2017 Min US EPA 2018 ANAPA River (A) Usa River (U) Maji ya chai river (M) Min Max Max. Fluoride (mg/L) Ca2+ (mg/L) 8.8±0.67 5.8 21 - 4 1.5 4.23±0.1 4 5.67 75 300 75 pH 8.25±0.27 7.8 8.5 6.5 9.2 Colour (TCU) 8 13 - 1.5 50 EC (µS/cm) 431±49.10 322 988 - TDS (mg/L) 215.5±24.6 161 494 - 600 Salinity (ppt) 0.22±002 0.16 0.49 - - 6.5 Min Max 2.0 4 - 8.5 6.5 8.5 15 15 250 500 - - - Table 4. ANAPA Water quality characteristics per numbers of experiments in the column No. run F(mg/L) pH Ca (mg/L) Temp (℃) EC (µS/cm) TDS (mg/L) Salinity (ppt) 1 7.7 8.4 4.2 23.6 422 211.0 0.21 2 8.7 8.2 4.1 24.7 345 172.5 0.17 3 8.8 8.6 4.4 23.2 487 243.5 0.24 4 8.9 7.9 4.3 22.2 438 219.0 0.22 5 9.8 8.4 4.2 23.2 468 234.0 0.23 6 8.7 8 4.2 23.2 426 213.0 0.21 Average Std. deviation 8.77 0.67 8.25 0.27 4.23 0.10 - - - 0.22 0.02 7 3.3 The fluoride removal efficiency of candidate materials from water along with bone char as control 3.3.1 Bench (Batch) adsorption system 3.3.1.1 The fluoride removal efficiency of coal The fluoride removal efficiency increased with contact time and adsorbent doses from 15g to 75g per 0.5L and decreased when the dose is above 75g/0.5L until the equilibrium reached where Ct=Co and Qt=Qe and declines; the fluoride attenuation results expressed in the table . The coal has showed a sufficient fluoride removal efficiency. The batch adsorption system took place with (0.5-1) mm particle size, Co=5.8mg/L of initial Fluoride level in water samples Table 5 Fluoride removal efficiency on the continuous monitored fixed bed adsorbent columns. Initial fluoride, Co=8.8mg/L Residual fluoride Removal concentration, efficiency, R% Fluoride removed (Co-Ct), mg/L Ct (mg/L) Breakthrough Ct/Qt t (hrs.) BC Coal BC Coal BC Coal BC Coal 1 0.12 0.27 97.45 94.26 4.58 4.43 0.005 0.012 2 0.25 0.74 94.68 84.26 4.45 3.96 0.007 0.025 3 0.27 1.23 94.26 73.83 4.43 3.47 0.006 0.035 4 0.38 1.58 91.91 66.38 4.32 3.12 0.007 0.041 5 0.57 1.65 87.87 64.89 4.13 3.05 0.009 0.036 6 0.72 1.87 84.68 60.21 3.98 2.83 0.010 0.038 7 1.2 2.61 74.47 44.47 3.5 2.09 0.017 0.062 8 1.57 3.43 66.60 27.02 3.13 1.27 0.022 0.120 Average 86.49 64.41 4.07 3.03 0.01 0.05 Std. deviation. 10.84 21.35 0.51 1.00 0.01 0.03 ratio, 3.3.1.2 Effect of adsorbent dose The optimal dose for coal was observed as 75g; since beyond this no significant removal efficiency observed (figure); since the number of accessible adsorption sites and the surface area increase by increasing the adsorbent dose and therefore they result in increased in the amount of adsorbed fluoride to 85% in coal (table ) as dose increased from 15g to 75g but decreased beyond 75g (Figure 3 ); on further increase the dose, the adsorption remains almost the same. Even though percent adsorption 8 increases with increase in adsorbent dose, amount adsorbed per unit mass decreases due to overlapping of adsorbent sites as presented by (6). Figure 3. Fluoride uptake by coal under different dosage 3.3.1.3 Adsorption Isotherms The linear regression coefficients for adsorption data fitted well on Langmuir isotherm model with determination coefficient, R2=0.9717 and 0.9862 for coal and zeolite respectively than Freundlich isotherm. Figure 4 show that 95% of data lie (fit) within prediction band as expected. The adsorption reaction of coal in batch system was reversible (RL<0) and unfavourable (1/n<0) under Freundlich and Langmuir isotherm respectively (table 5). The adsorption can be favourable by using activated and modified coal. Bone char has the specific ability to take up fluoride from water. This is believed to be due to its chemical composition, mainly as hydroxyapatite (Ca10(PO4)6(OH)2); in which the calcium , phosphate and hydroxyl group(s) can be chemically bonded and exchanged with fluoride from the 𝑦𝑖𝑒𝑙𝑑𝑠 media; Ca10(PO4)6(OH)2 + 2F– �⎯⎯⎯� Ca10(PO4)6 F2 + 2OH– 95%CI, the fluoride removal best fitted data increased with contact time and dose from 15g to 50g as for coal is 50g (r= 0.9861, R2= 0.9723); 15g (r= 0.9845, R2=0.9693). This indicates that, coal is effective at 50g; and the efficient decreased at the dose higher than optimal dose (figure) as time contact time increased with increasing dose. Table 5. The behaviour of coal coal under batch adsorption isotherms Freundlich Adsorption Isotherm (F.A.I) Langmuir Adsorption Isotherm (L.A.I) Parameters Coal Parameters Coal 𝑲𝑭 (L/mg) 𝑹𝟐 (Linear) 1/n - -0.957 𝑄𝑚𝑎𝑥 (g/kg) 𝑅 2 (Linear) 𝐾𝐿 (L/mg) RL - -811.3 -0.00014 9 Figure 4. Langmuir and Freundlich adsorption isotherms test results for 15g of coal in the batch system 3.3.2 Fixed bed adsorption columns The laboratory results obtained after running an experimental setup by using water samples from Burko river of Arusha National Park (ANAPA). The discontinuous (fixed bed) monitoring results per The adsorption isotherm were carried out using columns beds (media) with (1-1.4) mm particle sizes. The fluoride removed declined over time as the adsorbent bed saturated with fluoride due to the competition of co-cations and fluoride on the absorbent sites. Overall means of residual fluoride was statistically significant at (p, 0.0013<0.05) and F (2, 21) = 9.330. The means between bone char and coal are not statistically significant (p, 0.0006<0.05); at this point coal shared common characteristics with bone char on fluoride adsorption characteristics. The fluoride removal Figure 5. Coal column performance on fluoride attenuation efficiency of coal in continuous flow columns (90L each) were (64.41±21.35)% with BC as control (86.4±10.84)%; while in the discontinued flow columns (240L each) were (64.48±22.30)% coal as compared to BC (97.03±2.17)%; when both particle size were (1-1.4) mm. In the continuous operated columns; the positive relationship between coal and bone char was high (r= 0.9663, R2 = 0.9337) on the fluoride removal efficiencies. This can be attributed by high ANC of coal in the uncontrolled pH conditions. The main characteristics of these materials is their contents of metal lattice and hydroxyl groups (7), which can be exchanged with fluoride. The ion exchange of a metal compound M can be illustrated as: -M-OH(s) + F– M-F(s) + OH– 10 Figure 6.An equilibrium fluoride concentration and adsorption capacities of coal along with BC Figure 7.The behaviour of coal along with BC under Adsorption isotherms 11 The overall performance of the adsorbents in the columns was efficient at start-up and start to further decline in performance after two weeks of operation. Keeping bone char constant, the bituminous coal show higher fluoride removal efficiency (figure ) as removed 8.61mg/L of fluoride at 97.84% removal efficiency of fluoride from initial (8.8mg/L); adsorption system with m=5kg, 1-1.4mm particle size, Q=0.137L/min, and Co=8.8mg/L of initial fluoride concentration of ANAPA water samples 3.3.2.1 The optimization of fluoride removal attributes i. Effect of pH Figure 8. The impact of pH on the fluoride attenuation in water by a discontinuous flow monitored of the fixed bed columns The hydrogen potential for water sample were analysed before and after adsorption process. This is the key factor which affects the adsorption system. Both adsorbents work best within neutral range of pH; the adsorption capacities of the adsorbents declined at the higher pH level. pH changes can cause changes of surface charge of an adsorbent. Figure 8 shows the effect of pH by using 5kg and (1-1.4) mm particle size bone car and coal in the columns; the percentage of fluoride attenuation increased toward neutral and acidic side. This is because surface of an adsorbent is highly protonated in extremely acidic media, but it is neutralised and tends to have negative charge in alkaline media (8).Therefore, the high efficiency in acidic media can be attributed to a gradual increase in attractive forces, and the low efficiency in alkaline media can be explained by the repulsion between the negatively charged surface and fluoride ions. The R2 on bone char was 0.8695 and coal was 0.8444; The group means were statistically significant (p, 0.0013<0.05) and F (2, 21) = 9.330. Also, the fluoride removal was enhanced at acidic pH due to the formation of positively charged surface sites is highest at acidic pH, which attracts more negatively charged fluoride ions by electrostatic attraction. Different to this, fluoride adsorption declines at high solution pH because of strong competition between fluoride ions and hydroxide ions for active adsorption sites. At alkaline pH, fluoride is adsorbed by an ion-exchange mechanism (6). The Pearson correlation analysis (r) show that at 95%CI; coal and bone char have best correlation on the fluoride removal efficiency (r= 0.9663, R2= 0.9337). ii. Effect of contact time The effect of contact time was noted by changing it at constant adsorbent dose. The discoveries are shown in the Figure which show the fluoride adsorption increases with the initial increase in contact time and after a few hours the process of adsorption reaches its equilibrium after 4 hours and the rate of reactions become almost constant. This is because, initially all the adsorbent sites are vacant and the intake of fluoride is rapid, which is seen from the steep rise in the adsorption curve. Thereafter, the fluoride adsorption decreases due the saturated active sites on the adsorbent surface. When the equilibrium is reached, the rate of fluoride adsorption and desorption became the same. 12 𝑦𝑖𝑒𝑙𝑑𝑠 ( 𝑏 ≈ 1 �⎯⎯⎯� 𝐾𝑎 = 𝐾𝑑 ). iii. Breakthrough curves Breakthrough curves were obtained using both three columns containing 5kg each with 1-1.4 mm particle size and Break-through curves were obtained by plotting the ratio of residual fluoride concentration (Ct) to initial concentration (Co) with contact time. Since the service time (τ) is the time to breakthrough point in fluoride; For defluoridation of drinking water, breakthrough is taken as the point where the effluent fluoride concentration reaches the drinking water limits. The equivalent ratio of residual fluoride concentration to initial concentration (Ct/Co) at breakthrough point for the standards for the initial 8.8 mg /L of fluoride is found as 0.72mg/L(τ=6hr) and 1.58mg/L (τ=4hr) from the columns with bone char and coal respectively; those fluoride levels were within TZS standards (Figure 9). The Dunnett multiple comparison of means in unpaired one-Way ANOVA show that, the means of fluoride removal efficiency were not statistically Figure 9. The breakthrough curves of bone char significant at (p, 0.8181>0.05) between Bone char and and coal columns Coal. F (2, 21) = 26.55; There is some similarity in adsorption behaviour between coal and bone char characteristics. iv. Effect of residual calcium ions and fluoride adsorption mechanisms Generally, the results show that; the quantity of fluoride removed increased as calcium ions (Ca2+) decreased from the aqueous solution per unit increase in contact time (days). The calcium released by bone char used to uptake fluoride from water thereby reduced the free calcium from the aqueous solution and increased the fluoride removed as contact time increased from 2nd to 7th day of column operation (Figure 10). Bone char reacts with fluoride by electrostatic attraction between fluoride in the water and bone char components (−Ca−OH2+). The fluoride removal reaction can be represented by the following equation (9): 𝑦𝑖𝑒𝑙𝑑𝑠 Ca10(PO4)6(OH)2 + 2F- �⎯⎯⎯� Ca10(PO4)6F2 + 2OH- The fluoride removed slightly by coal under the influence of calcium ions. This is due to the presence of other components in the coal particularly OH- and -COOH which dissociates and attract fluoride ions from an aqueous solution by these mechanisms: 𝑦𝑖𝑒𝑙𝑑𝑠 Generally: >MOH+F-+H+ �⎯⎯⎯� >MF+H2O Reactions: >C=O….H-F, >C-O-H…...F-H 13 and >C-O-H…...H-F- Figure 10. The behaviour of candidate materials and control in fluoride removal under the effect of calcium ions v. Optimization of Particle size The 5kg of coal and control (BC) with particle size of (0.5-1.4) mm was used in the fixed bed adsorption column. The water samples from ANAPA (8.8mg/L fluoride) was allowed to flow though the column at 0.137L/min rate per 3hrs contact time intervals per experiment and the results from laboratory analysis showed that; the adsorption increased from 77.27%, 97.05 to 87.27%, 98.18% on coal and BC as particle size decreased from (1-1.4) mm to (0.5-1.4) mm respectively per unit increase in contact time per column (Figure 11). This is due to the increase of adsorbents surface areas for adsorption process, hence more fluoride (adsorbant) come into contact with the adsorption sites. Figure 11. The effect of particle size on the performance of the fixed bed adsorption columns 14 3.3.2.2 Adsorption Isotherms on the fixed bed-column study The adsorption data fitted well the Freundlich isotherm by coal with R2=0.7050 as compared to bone char (R2=0.8566); with an Freundlich adsorption equilibrium constant, KF for coal (43.45L/kg). The results show that adsorption on coal and BC was both favourable physiosorption (1/n<0). The Langmuir qualification results showed the correlation coefficient of determination R2= 0.8473 on coal and 0.9909 on bone char; The maximum bed adsorption capacity, Qmax (g/kg) of coal was 12.938g/kg as compared to bone char (79.05g/kg) in the fixed bed columns (Table 6). The adsorption can be favourable by using activated and modified coal. The experimental fluoride adsorption capacities of coal in continuous flow columns were (36.40±9.32)g/kg as compared to BC (53.33±18.19) mg/kg; while in the discontinued flow were (31.97±6.92)g/kg coal as compared to BC (52.92±17.78)g/kg. The table show that there is the High positive correlation between coal and bone char ; in the removal of fluoride under both continuous and discontinuous flow when the particle size used in the columns were (1-1.4)mm. Table 6. The equilibrium isotherm models result of fixed bed adsorption columns Empirical model: Freundlich Adsorption Isotherm (F.A.I) Theoretical model: Langmuir Adsorption Isotherm (L.A.I) Parameters Coal BC Parameters Coal BC 𝐾𝐹 (L/kg) 31.99 43.45 12.938 79.05 0.7050 0.8566 𝑄𝑚𝑎𝑥 (g/kg) 0. 8473 0.9909 1/n -0.4185 -0.1438 𝐾𝐿 (L/kg) -0. 0773 -0.0273 RL 1.571 𝑅 2 (Linear) Linear equation 𝑦 = −0.4185𝑥 + 1.505 𝑅 2 (Linear) Linear equation 𝑦 = 0.1438𝑥 + 1.638 𝑦 = 0.7729𝑥 − 0.0551 𝑦 = 0.0273𝑥 −- 1.147 Figure 12. Multiple correlation between control and candidate materials in the fixed bed columns Discontinuous flow BC versus Coal BC versus Coal Pearson r Pearson r r Continuous flow 0.7576 r 0.9663 95% confidence 0.2446 to 0.9392 interval 95% confidence 0.8199 to 0.9941 interval R squared R squared 0.5739 15 0.9337 4. Practical application of findings from adsorption experiments Thomas’s model was used for the explanation of the adsorption process ; to achieve the determination of adsorption capacity of an adsorbent which is required in column design and applicability of the coal as predicted by adsorption isotherms. This model was applied in the continuous fluidized beds with constant initial fluoride of 7.7mg/L, filtration rate of 0.137L/min, mass of 5kg which generated bed depth with 50cm in both columns. This model shown in the following equation (a): 𝐶 𝑙𝑛 �� 𝐶𝑜 � − 1� = 𝑡 𝑚𝑞𝑜 𝐾𝑇ℎ 𝑄 − 𝐾𝑇ℎ 𝐶𝑜 𝑡……………………………………………………………………………………………. (a) Whereby; 𝑄 is the filtration velocity (mL/min) ; 𝑞𝑜 is the equilibrium adsorbate uptake per gram of adsorbent; 𝑚 is the mass of adsorbent (kg); 𝑡 is the projected filtration time (min); 𝐶𝑜 is the inlet adsorbate concentration (mg/L), 𝐶𝑡 is the outlet or residual concentration at time t (mg/L) and 𝐾𝑇ℎ is the Thomas rate constant (mL/min-mg). The columns data were fitted to the Thomas model to determine the Thomas’s rate constant 𝐾𝑇ℎ and maximum solid phase concentration 𝑞𝑜 by plotting 𝐶 𝑙𝑛 �� 𝐶𝑜 � − 1� against 𝑡; the results were shown in the table 4.16, figure 𝑡 The amount of fluoride adsorbed (𝑞𝑜 ) in the columns still increased with decreasing particle size from (1-1.4) mm to (0.5-1.4) mm in the Coal (0.4668g/kg) as compared with bone char (0.530g/kg); to 0.7548g/kg as compared with bone char (1.5359g/kg) respectively. At this point, the Thomas’s model showed the good adsorption behaviour with r=96.7% and r=82.3% on coal versus BC by using the columns with adsorbents beds with particle size of (1-1.4) mm and (0.5-1.4) mm respectively. There was no significant difference between adsorption behaviour of coal and BC (p<0.05, R2=93.6%). The summary performance of the results as evaluated by various adsorption models in this continuous flow study were presented in the table 7 Table 7. The performance of coal as evaluated by Thomas’s adsorption model in the column study under the continuous flow Parameters (1-1.4) mm (0.5-1.4) mm Control Candidate materials Control Candidate materials BC Coal BC Coal R2 0.9814 0.9277 0.8245 0.8811 𝐾𝑇ℎ (mL/min-mg) 0.0517 0.0587 0.0178 0.0363 0.530 0.4668 1.5359 0.7548 r 1 0.9676 1 0.8237 𝑞𝑜 (g/kg) 16 Figure 13. Thomas’s model regression results for continuous fluidize (flow) columns According to the determined adsorption capacity and performances in both continuous and discontinuous fluidized beds columns; column adsorption using coal as adsorbent is applicable as the best to be used on get rid of fluoride in both drinking water and wastewater treatment. According to the ministry of water on the report: strategies for scalping up defluoridation activities and piloting research findings, 2013; the empirical formula used to estimate the number of days a given bone char that can last in the columns is: 𝐶= 𝐻𝐼 103 ………………………………………………………………………………(b) 𝐵𝐷𝐺 Whereby D= number of users (persons), C=Period of operation in days, I=Mass of bone char required (kg), B= Water demand per person per day (Lperson-1day-1), G=Amount of fluoride removed (g/L) and H=Bone char sorption capacity of F- (g/kg). Here, D=4 persons, I=4kg, B=30 Lperson-1day-1; the recommended flow rate was 20L/family/day only for drinking and cooking; and for economic reasons the defluoridation unit should use raw water with fluoride concentration below 10mg/L; and replacement of media between three to four months. Since coal performed well; the maximum fluoride adsorption capacity and fluoride removed by coal under Langmuir’s isothermal conditions were 12.938g/kg and 4.43mg/L; since bone char was a control, the same empirical model (equation (b)) for determination of maximum operation period (days) the candidate’s materials can last before saturation was used: 𝐶= 𝐻𝑥4 103 30𝑥4𝑥4𝑥𝐺 The period of operation, C ≅23 days for coal column. This time can be extended by using chemically and physically modified coal for the adsorption; Hence it can be applicable at the household and be extended to community level in the fluoride mitigation from the water for domestic use. The laboratory scale columns experiments can be used to generate parameters for improving the design of full scale adsorption based water treatment systems; as their performances were feasible and agreed with (10-13) to uptake fluoride from water at applicable efficiency and effectiveness of 64.41% and 36.4g/kg capacity determined under continuous fluidized beds experiments. 17 5. CONCLUSION Generally; as bone char act as control, the coal demonstrated higher performance under the several experimental conditions as validated by both Langmuir and Freundlich adsorption isotherms and OneWay ANOVA at 95%CI (p=0.05). Hence; It can be applied in large-scale defluoridation units to attenuate excess fluoride from water for domestic use. The saturated coal can be regenerated and used as the building materials such as concrete. Also the final saturated coal should be disposed into the incinerator. The performance characteristics of coal in the fluoride adsorption should be improved by chemical modification. REFERENCES 1. Al-Akwa AA, Al-Maweri SA. Dental caries prevalence and its association with fluoride level in drinking water in Sana'a, Yemen. Eur J Dent. 2018;12:15-20. 2. Wang X, Zhou Y, Xu C, Song H, Pang X, Liu X. A dual-responsive fluorescent probe for detection of fluoride ion and hydrazine based on test strips. Spectrochim Acta A Mol Biomol Spectrosc. 2019;211:125-31. 3. Li Y, Wang S, Nan Z, Zang F, Sun H, Zhang Q, et al. Accumulation, fractionation and health risk assessment of fluoride and heavy metals in soil-crop systems in northwest China. Sci Total Environ. 2019;663:307-14. 4. Maurice Onyango YK, Ochieng Aoyi , Eileen Bernardo, and Hitoki Matsuda. Adsorption equilibrium modeling and solution chemistry dependence of fluoride removal from water by trivalentcation-exchanged zeolite F-9. Elsevier. 2004. 5. El-Said GF, El-Sadaawy MM, Aly-Eldeen MA. Adsorption isotherms and kinetic studies for the defluoridation from aqueous solution using eco-friendly raw marine green algae, Ulva lactuca. Environ Monit Assess. 2017;190:14. 6. Bharudin I, Abdul Rahim SN, Abu Bakar MF, Ibrahim SN, Kamaruddin S, Latif MT, et al. De novo transcriptome resources of the lichens, Dirinaria sp. UKM-J1 and UKM-K1 collected from Jerantut and Klang, Malaysia. Data Brief. 2018;19:2416-9. 7. Mezzenga R, Zhang Q, Bolisetty S, Cao Y, Handschin S, Adamcik J, et al. Selective and efficient removal of fluoride from water by in-situ engineered amyloid fibrils-ZrO2 hybrid membranes. Angew Chem Int Ed Engl. 2019. 8. Sivasankar V. Surface Modied Carbons as Scavengers for Fluoride from Water. 2016. 9. Herath H, Kawakami T, Tafu M. The Extremely High Adsorption Capacity of Fluoride by Chicken Bone Char (CBC) in Defluoridation of Drinking Water in Relation to Its Finer Particle Size for Better Human Health. Healthcare (Basel). 2018;6. 10. Velazquez-Jimenez LH, Hurt RH, Matos J, Rangel-Mendez JR. Zirconium-carbon hybrid sorbent for removal of fluoride from water: oxalic acid mediated Zr(IV) assembly and adsorption mechanism. Environ Sci Technol. 2014;48:- 11. Shahid MK, Kim JY, Choi YG. Synthesis of bone char from cattle bones and its application for fluoride removal from the contaminated water. Groundwater for Sustainable Development. 2019;8:-. Saikia J, Sarmah S, Saikia P, Goswamee RL. Harmful weed to prospective adsorbent: lowtemperature-carbonized Ipomoea carnea stem carbon coated with aluminum oxyhydroxide nanoparticles for defluoridation. Environ Sci Pollut Res Int. 2019;26:-. Waghmare SS, Arfin T. Fluoride Removal by Clays , Geomaterials , Minerals , Low Cost Materials and Zeolites by Adsorption : A Review. 2015;4:3663-76. 19