Biosorption of fluoride ion from water using the seeds of the cabbage tree (Moringa stenopetala)

Conventional water treatment technologies for the removal of fluoride ion may not be feasible for developing countries due to their high investment and operational costs. The aim of this study was therefore, to investigate the fluoride biosorption potential of the seeds of the cabbage tree (Moringa stenopetala). The influence of Moringa dosage, pH, contact time, and initial concentration of fluoride ion was investigated. The maximum fluoride sorption capacity was found to be 1.32 mg.g -1 of dry weight of Moringa seeds at a biomass dosage of 2 g L -1 , pH 7.00, initial fluoride ion concentration of 10 mg.L -1 and a contact time of 60 min. The fluoride level was reduced from 10 to 3.4 mg L -1 . The adsorption of fluoride ion onto Moringa powder was best described by the pseudo-second-order kinetic model (R 2 = 0.99). The adsorption equilibrium data have been fitted well to Langmuir as well as Freundlich adsorption models (R 2 ≥0.97 for both models). The distribution constant (Kd) and maximum adsorption capacity (Bmax) were significantly influenced by the amount of Moringa and equilibrium fluoride ion concentration (p<0.05). The desorption tests indicated that only 20% of the initially bound fluoride ion was regenerated, while the remaining 80% were bounded with the Moringa powder. This suggests that chemisorption was the possible mechanism of fluoride removal.


INTRODUCTION
Fluoride related health hazards are a major environmental problem in many regions of the world. Studies revealed that Ethiopia is among the 25 nations around the globe, where health problem occurs due to the consumption of fluoride contaminated water (Ayoob and Gupta, 2006). Out of 10 million people living in Rift Valley region of Ethiopia, 8.5 million people are exposed for high fluoride contamination. In the Ethiopian Rift Valley, ground and surface water fluoride concentration varies from 0.5 to 264 mg.L -1 (up to 26 mg.L -1 in drinking water sources) (Tekle-Haimanot, 2006). As a result, over 80% of the children in the rift valley areas have developed varying degrees of dental fluorosis . The public health and economic importance of fluorosis is significant in many endemic areas in view of the occurrence of debilitating skeletal fluorosis in humans E-mail: seid.tiku@ju.edu.et. Tel: +251913296056. Fax: +251 471 114484.
Author(s) agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License and more recently, the discovery of pathology in cattle, sheep and other livestock. Crippling skeletal fluorosis is confined to tropical and sub tropical areas (WHO, 1984). Human sufferings due to dental and skeletal fluorosis, medical expenses to treat fluorosis, and untimely retirement of the productive members of the society can be prevented by defluoridating drinking water.
It is evident that there are different techniques available which have been found to be successful in defluoridation of drinking water containing excess fluoride. It is the only practicable option to overcome the problem of excessive fluoride in drinking water, where alternate sources are not available, and extensive research has been done on various methods for its removal. These methods are based on the principle of adsorption (Cengeloglu et al., 2002;Fan et al., 2003), ion-exchange (Wang et al., 2013), precipitation-coagulation (Roy and Dass, 2013), electrolytic defluoridation (Mameri et al., 2001) and electrodialysis (Hichour et al., 2000). However, conventional defluoridation technologies have high operational and maintenance costs, low fluoride removal capacity, a lack of selectivity for fluoride, undesirable effects on water quality, the generation of large amounts of sludge and complicated procedures involved in the treatment. Moreover, most defluoridation methods are unproven and unreliable under field condition in developing countries (Kloos and Tekle-haimanot, 1995;. In view of these serious drawbacks, there is a great need to develop an effective, efficient and eco-friendly adsorbent for the removal of fluoride from water. To this end, a wide range of non-living biomass such as freshwater macrophytes such as Eichhornia crassipes (Sinha et al., 2003;Karmakar et al., 2016), fungi (Amin et al., 2015), algae , yeast  and Moringa olifera (Bazanella et al., 2012;da Conceição et al., 2015) have been investigated as biosorbents. In the present work, seeds from the cabbage tree (Moringa stenopetala) were employed to remove fluoride ions from drinking water. M. stenopetala belongs to the family of the Moringaceae which is represented by a single genus, called Moringa. The water soluble Moringa seed proteins possess coagulating properties similar to those of aluminum sulphate and synthetic cationic polymers. Moringa seeds contain cationic polypeptides with various functional groups, particularly low molecular weight amino acids (Jose et al 1999). These amino acids are deprotonated to carboxylate ligands at pH range of 4 to 8, simultaneously protonating the amino group which facilitates the binding of negatively charged ions with the amino group.
A batch experiment conducted by Sahilu (2010) revealed that a 2 g.L -1 of M. stenopetala seed powder reduced 9 mg/L of fluoride ion from ground water to 2.2 mg/L. In addition studies have shown that the seeds of the cabbage tree (M. stenopetala) removes heavy metals such as hexavalent chromium (Degefu and Dawit, 2013); cadmium (Mataka et al., 2010) and lead (Mataka et al., 2006) from water in batch experiments. The main objective of this study was to determine the fluoride sorption potential of seeds of M. stenopetala from aqueous solution. The influence of biomass dosage, initial fluoride ion concentration, pH and contact time were investigated.

Preparation of biosorbent
M. stenopetala seeds were collected from the Chefe area, South Wello zone, Ethiopia. Seeds were de-shelled by hand and the deshelled seeds were dried in an oven at 105°C for 72 h. The dried seeds were ground in a mortar and sieved through 1 mm mesh. No other chemical or physical treatments were used prior to the biosorption experiments.

Preparation of fluoride solutions
A stock solution of fluoride was prepared by dissolving 2.21 g anhydrous sodium fluoride in one litre of distilled water. The sodium fluoride was previously dried at 105°C to a constant weight and stored in a dessicator. The fluoride concentration in this stock solution was 1000 mg.L -1 . Other concentrations were prepared from this stock solution by dilution and varied between 1 and 40 mg F.L -1 . Fresh dilutions were prepared for each experiment. All the chemicals used were of analytical grade. Total Ionic Strength Adjustment Buffer (TISAB) solution was prepared from 58 g of sodium chloride, 4 g CDTA, and 57 mL of glacial acetic acid for adjusting the pH. The TISAB solution regulates the ionic strength of samples and standard solutions and adjusts the pH and also avoids interferences by polyvalent cations.

Biosorption test
Sorption studies were carried out in batch experiments as a function of biomass dosage (2 to 20 g), contact time (5 to 120 min), pH (2 to 13) and fluoride ion concentration (1 to 40 mg.L -1 ) at a water temperature of 20±1°C. The required amount of fluoride solution was taken in an Erlenmeyer flask, diluted to 250 ml with distilled water and the pH was adjusted to the desired value. Then, a known quantity of Moringa powder was added. These suspensions were placed on a shaker during a certain time. After shaking, the suspension was allowed to settle for 15 minutes. The suspension was filtered by using white band Whatman filter paper. The filtrate was collected and analyzed for Fluoride ion by means of Fluoride Ion Selective Electrode METTLER TOLEDO model. The fluoride ion concentrations before and after sorption were recorded. Percent removal by sorption to the sorbent was computed using the equation: where Co is the initial fluoride ion concentration in the solution and Ce is the final fluoride ion concentration in the solution.

Adsorption kinetics
In order to evaluate the kinetic parameters, pseudo first-order and second-order models were tested to analyze the adsorption kinetics. Kinetic studies are important in determining the optimal contact time required to reach equilibrium (Ghorai and Pant, 2005).
The pseudo-first-order equation is expressed as: The pseudo-second-order equation is expressed as: where qe and qt are the amounts of fluoride adsorbed (mg/g) at equilibrium and at any time t (min), respectively. Where qe and qt are the amounts of fluoride adsorbed (mg/g) at equilibrium and at any time, t (min), respectively. K1 is the pseudo first-order reaction rate constant of adsorption (min -1 ) and K2 the pseudo-second-order rate constant of adsorption (mg/g, min). The value of K1 and qe were calculated from the slope and intercept of the plot of log(qe-qt) versus t, while the value of K2 and qe were calculated from the intercept and slope of a plot of t/qt versus t, respectively (Ghorai and Pant, 2005).

Intra-particle diffusion of fluoride
Weber-Morris model was used to understand the possible contribution of intra-particle diffusion for the removal of fluoride on an adsorbent. The linear form of intra particle diffusion model given by Weber-Morris is: where qt is the amount of fluoride adsorbed per unit mass of adsorbent (mg/g) at a given time t, kp is the rate constant of intraparticle diffusion (min -1/2 ) and t is contact time (min). The uptake is proportional to the square root of contact time during the course of adsorption: where qt is the amount adsorbed at equilibrium (mg/g), kp is the rate constant of intra-particle transport (mg g -1 min -1/2 ). The intra-particle diffusion rate constant (kp) value estimated from the slope of plot of qt versus square root of time (Weber and Morris, 1963).

Adsorption isotherms
The adsorption of fluoride ion was tested with 2 g Moringa and varying fluoride ion concentrations (from 1 to 40 mg. L -1 ) at pH ±7.0. The contact time was 60 min. The volume of test solution was 250 ml. The biosorption equilibrium uptake capacity for each sample was calculated based on a mass balance.
V is the sample volume (L), C0 is the initial fluoride ion concentration (mg.L -1 ), Ce is the final fluoride ion concentration (mg.L -1 ), M is the amount of biomass (g dry matter), S is the amount of fluoride ion sorbed to the Moringa (mg.g -1 ).

Freundlich isotherm
The Freundlich isotherm can be described as follows: Ce is the equilibrium fluoride ion concentration (mg.L -1 ), S is the amount of fluoride ion sorbed to the Moringa (mg.g -1 ), Kf is the Freundlich constant denoting the adsorption capacity (mg.g -1 ), n is the empirical constant, indicating adsorption intensity (L.g -1 ). Kf and n denote, respectively the adsorption capacity (mg.g -1 ) and adsorption intensity (L.g -1 ) of the Moringa.
The logarithmic form of the equation is given as follows: log S = log Kf + 1/n log Ce

The Langmuir model
The Langmuir model can be described as follows: where Ce is the equilibrium fluoride ion concentration (mg.L -1 ), S is the amount of fluoride ion sorbed to the Moringa (mg.g -1 ), K is the adsorption maximum (mg.g -1 ), b is the constant (L.g -1 ). The constant b represents the binding strength of the adsorption place for the ion.

Ligand binding model
The adsorption of fluoride ion to the biomass was fitted to a ligand binding model with one site saturation. This can be described as follows: where S is the amount of fluoride sorbed to the Moringa (mg.kg -1 ), Bmax is the maximum binding capacity (mg.g -1 ), Ce is the equilibrium fluoride concentration (mg.L -1 ), and Kd is the equilibrium distribution constant.

Desorption and reusability studies
In order to determine the reusability of the biosorbent, adsorptiondesorption experiment were performed. A 250 ml test solution containing 20 mg.L -1 of fluoride ion was transferred into 250 ml Erlenmeyer flask. The pH of the solution was adjusted at around 7 using 0.2 M HCl or 0.2 M NaOH. After addition of 10 g of Moringa powder the mixture was shaken on a mechanically for 2 h. The suspension was filtered using Whatman filter paper and the residue left after filtration was subjected to desorption. Fluoride ion concentrations present in the filtrate were determined. Desorption The pH of M. stenopetala in 1M KCl solution and water was determined using pH electrode.

Data analysis
The obtained data were fitted by means of Sigma Plot version 13 onto a ligand binding model with one site saturation. The binding rate coefficient of Fluoride ion on half of the binding sites of Moringa biomass (Kd) and the maximum adsorption capacity (Bmax) was obtained from the curve. Significance was considered at p<0.05.

Effect of contact time on fluoride biosorption
The amount fluoride ion sorbed increases with time and reached its steady state in 60 min at which maximum sorption capacity (1.26 mg.g -1 ) were achieved (Figure 1). This is due to the fact that initially a large number of vacant surface sites are available for adsorption. With increasing time, the remaining vacant surface sites may be difficult to occupy due to repulsive forces between the molecules of the solid and bulk phase (Saravanane et al., 2002;. However, increase in contact time beyond 60 min did not increase the sorption efficiency, which might be due to the presence of fewer adsorption sites.

Effect of pH on fluoride biosorption
The fluoride sorption capacity progressively increased as the pH of the solution increased from 2 to 7 (Figure 2). At pH below 7, the decrease in fluoride adsorption efficiency and fluoride adsorption capacity might be due to the formation of hydrofluoric acid, which would reduce the coulombic attraction between adsorbent surface and the fluoride ion (Kagne et al., 2009). Moringa seeds contain cationic polypeptides with various functional groups, particularly low molecular weight amino acids (Jose et al 1999). The carboxylic group of the amino acids would progressively be deprotonated as carboxylate ligands at pH range of 4 to 8, simultaneously protonating the amino group. Such positively charged NH 3 + groups facilitate the Moringa-fluoride binding. As the pH rises above 7, sorption capacity dramatically decreased. This is due to the stronger competition for active sites between fluoride and hydroxide ions (Tembhurkar and Dongre, 2006). As hydroxyl ion concentration increases the overall charge on the biomass surface becomes negative. This causes a hindrance to the biosorption of the negatively charged Fluoride ion, resulting in a decrease of biosorption of fluoride at higher pH levels .

Effect of initial fluoride ion concentrations
As illustrated in Figure 3, by changing the initial concentration of fluoride ion concentration from 2 to 40 mg.L -1 , removal efficiency was reduced from 54 to 26%. A reduction in percent removal at higher fluoride ion concentrations may be due to the increase in the number of fluoride ions competing for the available binding sites on the biomass and the lack of binding sites available for binding of fluoride ions at higher concentration. Similarly, it has been found that fluoride removal using iron ore was smaller at higher fluoride ion concentration . In contrast, sorption capacity increases from 0.07 to 1.32 mg.g -1 as the initial fluoride ion concentration increased from 2 to 40 mg.L -1 . This finding is consistent with the results of , who found that the activity of fluoride ion increase as its concentration increases.

Effects of Moringa dosage on fluoride removal
As shown in Figure 4, fluoride sorption capacity decreased from 5 to 0.05 mg.g -1 when the biomass dosage increased from 2 to 20 g.L -1 . The reason might be attributed to the high biomass concentration which could make a "screen" effect on the dense outer layer shielding the binding sites from fluoride ion in the solution and thereby lowering the specific uptake at higher biomass loadings .

Adsorption kinetics studies
The derived parameters such as, qe, k 2 , and Kp, for the kinetic models are presented in Table 1. The coefficient    (Table 1). Therefore, the adsorption of fluoride ion onto Moringa powder is best described by the pseudo-second-order kinetic model suggest that fluoride should be adsorbed by chemisorption, which involves the sharing of electrons between fluoride and the adsorbent (Bhaumik and Mondal, 2015). However, the adsorption of fluoride did not follow the pseudo-first order equation (results not shown).
The plot of qt versus t 1/2 for intra-particle diffusion in the adsorption of fluoride ion onto Moringa powder was used to obtain the diffusion rate parameters. As presented in Figure 5b, the intra-particle diffusion rate constant (kp) value estimated from the slope of plot of qt versus square root of time was found to be 0.163 min 1/2 for the initial fluoride concentration of 20 mg/L. If intra-particle diffusion is a rate-controlling step, then the plots should be linear and pass through the origin (Weber and Morris, 1963). However, in this study, the plot does not pass through the origin. This suggested that fluoride removal is a complex process and the intra-particle diffusion was not the only rate controlling step which is similar with the findings of .

Adsorption isotherm of fluoride
A graphical representation of Freundlich and Langmuir adsorption isotherm is presented in Figures 6 and 7 , indicate good adsorption capacities. Kf is defined as the adsorbate adsorbed per unit weight of adsorbent (Chen et al., 2010). The Langmuir isotherms show a k value of 0.92 mg.g -1 and a value for constant b of 0.16 L.mg   , which is greater than the theoretical maximum adsorption capacity (K = 0.92 mg.g -1 ) obtained from Langmuir isotherm. Both the distribution constant (K d ) and maximum adsorption capacity (Bmax) were significantly influenced by the amount of adsorbent and equilibrium fluoride concentration (P<0.05).

Desorption test
In this study, desorption tests were employed to elucidate the nature of adsorption processes. A 10 mg.L -1 fluoride solution was allowed to adsorb onto 2 g.L -1 of Moringa seed powder. After 2 h of this sorption experiment, the concentration of fluoride in the filtrate was found to be 3.4 mg.L -1 , which means 7.6 mg.L -1 of fluoride ion was adsorbed on Moringa seeds (Table 2). Attempts were made to desorb fluoride ion from these fluoride ion loaded Moringa seeds using a 0.02 M KCl solution (pH 6.03).  After 2 h of desorption, only 20% of the initially bound fluoride ion was regenerated, while the remaining 80% remained bound with the Moringa powder, which indicates that most of the fluoride ions are able to form strong bonds with the positively charged functional group (NH 3 + ) of Moringa powder. The negligible desorption of fluoride ion with 0.02 M KCl indicates the predominance of chemical bonding between Moringa powder and fluoride ion. This implies that physical adsorption is not playing a significant role in fluoride removal by Moringa powder.

Conclusions
The results of this study revealed that application of M. stenopetala seeds as a biosorbent introduces a less expensive and environmentally friendly method for removal of fluoride ion from aqueous media. The removal of this pollutant was found to be depending on biomass dosage, pH, initial concentration and contact time. The adsorption capacity of Moringa for fluoride was 1.32 mg.g -1 of dry weight of Moringa seeds. The adsorption equilibrium data has been fitted very well to Langmuir as well as Freundlich adsorption models (r 2 ≥0.97). The desorption tests indicated that only 20% of the initially bound fluoride ion was regenerated, while the remaining 80% were bounded with the Moringa powder. This suggests that chemisorption was the proposed mechanism for fluoride removal. Moringa seeds powder could be applicable for the removal of fluoride ion from water, but could not bring fluoride concentration to permissible level. Hence, chemical activation or impregnation would increase the efficiency of this biosorbent. Further studies are required to determine functional groups of Moringa seeds which are responsible for fluoride fixation. Moreover, the surface morphology of the biosorbent and mechanism of fluoride-biosorbent interaction should be determined using scanning electron microscope (SEM) and/or Fourier transform infrared (FTIR).