Abstract

The Fe(III)-doped Scoria was prepared to examine its potential use as an efficient sorbent for removal of fluoride and nitrate from water. Structure and morphology of raw scoria (RS) and Fe(III)-doped scoria (FeS) were studied by scanning electron microscopy, X-ray diffraction analysis and Fourier transform infrared spectroscopy. A four-factor central composite design combined with response surface modeling (RSM) was employed for maximizing fluoride and nitrate removal based on 30 different experimental data obtained in a batch system. At optimum condition, the maximum removal of fluoride and nitrate were 78.36% and 81.4%, respectively. The kinetic of fluoride and nitrate adsorption onto RS and FeS were followed the pseudo-first-order with high determination coefficient values (R2 > 0.997). The isotherm data of fluoride was fitted with the Freundlich model, whereas equilibrium data of nitrate are better fitted to the Langmuir isotherm model. The Langmuir maximum adsorption capacities of Fe(III)-doped scoria for fluoride and nitrate were 0.317 and 11.3 mg/g, respectively. In conclusion, Fe(III)-doped scoria is recommended as an economic and efficient sorbent for nitrate and fluoride removal from contaminated water.

INTRODUCTION

Water, one of the most important elements in the Earth, is deteriorating continuously due to geometrical growth of population, civilization, industrialization, domestic, and agricultural activities, and other geological and environmental changes (Ali 2012). Numerous inorganic, organic, and biological pollutants have been reported as contaminants in water resources (Laws 2000). Based on World Health Organization (WHO) and UNICEF report, 748 million people do not have adequate and safe water resource and over 2.5 billion people have access to meagre water supply (Jadhav et al. 2015). Among these contaminants, anionic pollutants such as fluoride and nitrate contamination in groundwater has been recognized as a serious problem worldwide and they are included in major inorganic anions of great concern (Kumar et al. 2014). Excess fluoride intake leads to various types of fluorosis, primarily dental and skeletal fluorosis, pregnancy outcomes, blood pressure and diseases like Alzheimer's, bladder cancer, thyroid disorder, gastrointestinal bleeding and otosclerosis (Amini et al. 2008, 2016; Barbier et al. 2010). It is estimated that more than 200 million people worldwide (specially in countries like Iran, China, Mexico, Argentina and South Africa) rely on drinking water with fluoride concentrations that exceed the WHO guideline of 1.5 mg/L (Bhatnagar et al. 2011). Similarly, the presence of excessive concentrations of nitrate in drinking water causes gastric cancer, goitre, birth malformations, hypertension, intrauterine growth retardation, increased incidence of Sudden Infant Death Syndrome (SIDS), cardiac defects, and methemoglobinemia (Suriyaraj & Selvakumar 2016). The maximum concentration in drinking water stipulated to be 50 mg L−1 by the WHO (Tang et al. 2015).

Various methods have been proposed to remove excessive fluoride and nitrate from drinking water, such as chemical precipitation, electrocoagulation, reverse osmosis and electrodialysis (Cai et al. 2015; Golestanifar et al. 2016). Among these removal methods, adsorption method is considered relatively superior by virtue of its low cost, ease of operation, energy requirements, simplicity of design and regeneration (Nadafi et al. 2014; Jain et al. 2015). However, adsorption has certain limitations, like it could not achieve a good situation at commercial levels. Various types of natural and synthesized adsorbents including activated alumina, activated carbon, zeolites, hematite, schwertmannite, various geomaterials, composites, biosorbents, industrial products, polymer resins/fibers/membranes, organics, different metal oxides, and variety of nanomaterials have been tested for the removal of fluoride and nitrate from aqueous solution (Kumar et al. 2014; Suriyaraj & Selvakumar 2016). Basaltic volcanic scoria is the one of natural abundant adsorbent in many parts of the world such as middle east (Iran, Saudi Arabia), Central America, Southeast Asia (Vietnam, etc.), East Africa (Ethiopia, Kenya, etc.), and Europe (Greece, Italy, Spain, Turkey, etc.) (Moufti et al. 2000). It has been used as adsorbent for several pollutants such as Zn(II) (Kwon et al. 2005) Cd, Cu, Pb, Zn and arsenic (III) (Kwon et al. 2010), chromium (VI) (Moradi et al. 2015) and fluoride (Zhang et al. 2014). Some elements and metal ions with positive charge like Mg-Fe (Wu et al. 2015), Li(III)-Al(III) (Zhang et al. 2012), Fe(III)-Zr(IV) (Hassan et al. 2010; Swain et al. 2013), La(III) (Viswanathan & Meenakshi 2008), Ce(III) (Wang et al. 2013) have been used for modification of adsorbent to improve removal of fluoride and nitrate through enhancement in electrical affinity between adsorbent surface and negative charge pollutants.

In the present study, we have done a comprehensive evaluation for the usefulness of scoria and its modified form as an adsorbent for the removal of and from aqueous solution. First, we report the preparation and characterization of raw and Fe(III)-doped scoria and subsequently investigate the adsorption potential of and in context of response surface methodology (RSM) using central composite design (CCD), with considering operational parameter such as initial pH, adsorbent dose, initial fluoride and nitrate concentration and contact time. The analysis of variance (ANOVA) was conducted based on the proposed models to understand the interaction between the process variables and response. Finally, isotherm and kinetic studies were also an aim of this work which were performed in detail.

EXPERIMENTAL SECTION

Materials

Scoria rock was obtained from Qorveh mine in Kurdistan, west of Iran. All chemicals used in the study were analytical grade, purchased from Merck (Germany) and used without further purification. Stock fluoride and nitrate ion solutions of 1,000 mg/L were prepared from dried sodium fluoride (NaF) and potassium nitrate (KNO3), respectively. The solutions of required concentrations were prepared by diluting the stock solution. All experimentation processes were performed with deionized (DI) water.

Preparation of Fe(III)-doped scoria

The natural scoria was washed with distilled water until a clear washing effluent (<1 NTU) was obtained. To enhance the porosity and remove any dust, they were held in 0.1 N HCl for 24 h. After that, the stones were washed with distilled water to neutral pH, dried in an oven at 105°C, were ground to powder and sieved through 40 (0.42 mm) and 50 (0.3 mm) mesh, respectively. The obtained sample was firstly mixed with 6 N HNO3 for 24 h, dried at 105°C and then mixed with 0.5 M Fe(NO3)3 •9H2O at pH 12 and stirred for 72 h at 25°C (room temperature). NaOH was added dropwise to reach the desired pH. Finally, scoria was dried in an oven at 110°C for 14 h and to remove the undoped Fe, washed with DI and dried at 105°C for 14 h.

Adsorbent characterization

The surface morphology was examined using a JEOL 840A scanning electron microscope (SEM) (JEOL Ltd, Japan). The crystallization of scoria was studied by X-ray diffraction analysis (XRD) (APD 2000, GNR Corporation, Italy) in a 2θ range of 10° to 40° using a Cu-target tube (λ = 1.5406 Å) and a graphite monochromator. Elemental analysis of the natural and Fe(III)-doped scoria was obtained by X-ray fluorescence (XRF) analysis (BELEC-GmbH, Germany). Fourier transform infrared (FT-IR) analyses were conducted on a Vector 22 (Germany) FT-IR Spectrometer between 450 and 4,000 cm−1. The samples were prepared using the KBr pressed-disk technique, with 1% inclusions of the sample to be analyzed.

Adsorption experiments

Two types of batch sorption experiments (kinetic and equilibrium) were carried out in an erlenmeyer flask, 200 mL, which were shaken at 200 rpm at room temperature (22 ± 1°C). In the kinetic experiments, the reaction variables were varied as follows: reaction time from 15 to 75 min, initial fluoride and nitrate concentration 7 and 100 mg/L, respectively, sorbent concentration 4 g/L and pH 7. After completion of the experiments, the solution was centrifuged at 5,000 rpm for 10 min to separate the particle solid. Residual fluoride and nitrate ions were analyzed spectrometrically at λmax = 570 and 220 nm, respectively, in line with standard methods for the examination of water and wastewater.

The kinetic equations used in this work are: pseudo-first-order (Equation (1)); pseudo-second-order (Equation (2)); intra-particle diffusion model (Equation (3)); and Elovich model (Equation (4)).  
formula
(1)
 
formula
(2)
 
formula
(3)
 
formula
(4)
where qt and qe (mg/g) are the adsorption capacity of raw and Fe(III) doped-scoria in any time (t) and equilibrium; K1 (min−1) is the rate constant of the pseudo-first-order model; K2 (g/mg min) is the rate constant of the pseudo-second-order model; kid is intra-particle diffusion rate constant (mg/g min0.5); C is a constant related to the extent of the boundary layer effect; α (mg/g·min) is the initial adsorption rate and β (g/mg) is the sorption constant and is related to the surface coverage and the activation energy for the chemical adsorption.
Equilibrium sorption experiments were conducted under the following conditions: the adsorbent dose 1 to 7 g/L, the initial fluoride and nitrate concentration of 7 and 100 mg/L, respectively, the contact time of 60 min, pH 7 and room temperature. Langmuir (Equation (5)) and Freundlich (Equation (6)) are the equilibrium equations used in this work.  
formula
(5)
 
formula
(6)
where Qmax (mg/g) is maximum adsorption capacity; KL is related to adsorption energy (L/mg); KF (mg/g (L/mg)1/n) and n are Freundlich constants.

Experimental design and optimization by response surface methodology

Response surface methodology (RSM) was used for developing, improving and optimizing processes. A CCD, was employed for the optimization of fluoride and nitrate adsorption onto Fe(III)-doped scoria. To evaluate the effect of operating factors on the adsorption efficiency, four independent variables were chosen: pH, initial adsorbate ( and ) concentration (mg/L), adsorbent dose (g/L) and contact time (min). Table 1 presents the operating ranges and the levels of the independent variables considered in this work. Fluoride and nitrate removal efficiencies were considered as responses in this study. A total number of 30 runs were employed in this study including 24 = 16 cube points, six replications at the center point and eight axial points. Experimental data were analyzed using Minitab 17 software and Design expert 7.0.0 software. In addition, ANOVA was performed to find out the significance interactions between the process variables and the responses considering the F and p-values. The goodness of fit of the obtained regression model was expressed by the coefficient of determination (R2) (always lay between 0 and 1) and adjusted R2 ().

Table 1

Experimental ranges and coded levels of the independent variables

VariableReal values of coded levels
− 2− 10+ 1+ 2
pH (x111 
Adsorbent dose (x20.5 1.5 2.5 3.5 4.5 
Fluoride concentration (x310 13 
Nitrate concentration (x350 75 100 125 150 
Time (x415 40 65 90 115 
VariableReal values of coded levels
− 2− 10+ 1+ 2
pH (x111 
Adsorbent dose (x20.5 1.5 2.5 3.5 4.5 
Fluoride concentration (x310 13 
Nitrate concentration (x350 75 100 125 150 
Time (x415 40 65 90 115 

RESULTS AND DISCUSSION

Characterization of adsorbent

The SEM images of the raw scoria (RS) and Fe(III)-doped Scoria were shown in Figure 1(a) and 1(b), respectively. It was observed from Figure 1(b) that modified scoria has got more porosity due to washing and treating with nitric acid which remove any dust and impurities. It can also be seen from Figure 1(b) that the large amount of iron oxide with irregular shape (big clusters) is coated on the adsorbent surface and changed the surface structure of RS.

Figure 1

SEM images of (a) raw scoria, (b) Fe(III)-doped scoria.

Figure 1

SEM images of (a) raw scoria, (b) Fe(III)-doped scoria.

Figure 2 shows the XRD spectra of Fe(III)-doped scoria (2θ = 4–80°; 2θ steps = 0.02°; step time = 1 s; temp = 25°). The main mineral phases present in scoria are presented in Table 2, which are in accordance with previous studies (Kwon et al. 2005; Mahdizadeh et al. 2015).

Table 2

Mineral phases present in scoria

Mineral phaseMolecular formula
Quartz SiO2 
Orthoclase K(Si3Al)O8 
pyroxenes Ca(Mg, Fe, Al) (Si2Al2)O6 
Hematite Fe2O3 
Albite Na(Si3Al)O8 
Mineral phaseMolecular formula
Quartz SiO2 
Orthoclase K(Si3Al)O8 
pyroxenes Ca(Mg, Fe, Al) (Si2Al2)O6 
Hematite Fe2O3 
Albite Na(Si3Al)O8 
Figure 2

X-ray diffraction patterns of scoria used in the study.

Figure 2

X-ray diffraction patterns of scoria used in the study.

Major elements compositions of the RS showed that aside from the main Si and Al components, Ca is the next high component (Table 3). Based on Table 3, the Fe content increases from 8.9 (in raw sample) to 21.1% in doped scoria, which led to a decrease in the relative amount of Si and Al, and indicated that some Si atoms are substituted by Fe atoms. Si/Al ratio of 2.5 was obtained for Fe(III)-doped scoria which provide the high absolute values of the adsorption energies and meet the ideal ratio (Massoudinejad et al. 2015; Awuah et al. 2016).

Table 3

Elemental analysis of raw and Fe(III)-doped scoria

Major elementsRS (Con. (wt.%))Fe(III)-doped scoria (Con. (wt.%))
SiO2 47.4 39 
Al2O3 21.6 11.2 
Fe2O3 8.9 21.2 
CaO 12.4 9.2 
MgO 3.3 5.5 
Na2– 5.9 
K20.5 3.1 
Loss on ignition 5.9 4.9 
Major elementsRS (Con. (wt.%))Fe(III)-doped scoria (Con. (wt.%))
SiO2 47.4 39 
Al2O3 21.6 11.2 
Fe2O3 8.9 21.2 
CaO 12.4 9.2 
MgO 3.3 5.5 
Na2– 5.9 
K20.5 3.1 
Loss on ignition 5.9 4.9 

Figure 3 shows the FT-IR spectra of raw (a) and doped scoria (b) for identification of different functional groups which could be responsible for uptake of fluoride and nitrate. In general, there is no significant difference between the spectrum of raw and Fe(III)-doped scoria with the exception of sharper picks in doped scoria. Prominent broad bands of O−H stretching and O−H bending were observed in the range of 3,455 and 1,645 cm−1 which related to the presence of water molecules absorbed or entrapped on the surface of adsorbent (Fernández-Jiménez & Palomo 2005). The bands at 736 and 788 cm−1 are related to the stretching vibration of six-fold coordinated Al(VI)−OH and six-fold coordinated Al(VI)−O (Djobo et al. 2014). The bands located around 1,410 and 1,484 cm−1 are attributed to stretching vibrations of O–C–O bond (Fernández-Jiménez & Palomo 2005). The band located at 1,090 can be assigned to asymmetric stretching vibration of T–O–Si, T = Si or Al. The bands at 527–569 cm−1 were attributed to symmetric stretching of Si–O–Si and Al–O–Si (Panias et al. 2007).

Figure 3

FT-IR spectra of raw pumice (a) and Fe(III)-doped scoria (b).

Figure 3

FT-IR spectra of raw pumice (a) and Fe(III)-doped scoria (b).

Adsorbent screening

Figure 4 shows the experimental adsorption efficiency for each experimental run for fluoride and nitrate sorption in batch mode on RS and Fe(III)-doped scoria. As can be seen, the Fe(III)-doped scoria showed an efficient adsorption of both fluoride and nitrate relative to RS. In general, the behavior of raw and modified scoria was similar for the adsorption of fluoride and nitrate and both pollutants follow similar trend in removal. However, modification of scoria, which includes washing and holding in acid and doping with iron (III) play an important role in enhancement adsorption capacity of scoria. In the following, we have done the analyses of adsorption for Fe(III)-doped scoria.

Figure 4

Fast screening on the adsorption of fluoride and nitrate onto raw and Fe(III)-doped scoria (experimental run number obtained by RSM vs. experimental removal).

Figure 4

Fast screening on the adsorption of fluoride and nitrate onto raw and Fe(III)-doped scoria (experimental run number obtained by RSM vs. experimental removal).

Regression model and ANOVA

The four-factor CCD matrix and experimental and predicted results obtained in the adsorption runs on raw and Fe(III)-doped scoria for both fluoride and nitrate are presented in Tables S1 and S2 (supporting information (SI), available with the online version of this paper). The obtained results were analyzed by ANOVA at 95% confidence level (p < 0.05), which gave the following regression equations (in term of coded factors) (Equations (7) and (8)) for and adsorption onto Fe(III)-doped scoria, respectively.  
formula
(7)
 
formula
(8)
where Y1 and Y2 are response variables of and removal efficiency, respectively. The Design Expert software suggested linear model for adsorption and interactive (2FI) model for adsorption by supporting lack of fit (LOF), Fisher's F-test value (F-value) and model summary statistics (Table 4). Although linear model was also significant and suggested for nitrate adsorption, 2FI model had lower p-value and higher F-value which software calculated the summary statistics for 2FI model. The F-value is a statistically valid measure of how well the factors describe the variation in the data around its mean. The larger F-value indicates more certainty, in that the factors explain adequately the variation in the data around its mean, and the estimated factor effects are real (Liu & Chiou 2005). The ANOVA indicated that the F-values for fluoride and nitrate adsorption on Fe(III)-doped scoria were 23.33 and 6.98, respectively, which is clearly greater than the tabulated F0.05,4,26 (2.74 at 95% confidence level), confirming the adequacy of the model fits. Referring to Table 4, the coefficient of determination (R2) values of 0.79 and 0.83 for and indicated that 79% and 83% of the variability in the response could be explained by the model. Furthermore, the adjusted R2 values are close enough to R2 values which mean insignificant variables are not included in the models (Salarian et al. 2016).
Table 4

Suggested model for fluoride and nitrate adsorption and model summary statistics

Fluoride adsorption on Fe(III)-doped scoria
Sequential model sum of squares
SourceSum of squaresdfMean squareF-valuep-value prob > F
Mean vs total 136,613.3 136,613.3    
Linear vs mean 1,298.2 324.5 23.3 <0.0001 Suggested 
2FI vs linear 52.89 8.8 0.56 0.7508  
Quadratic vs 2FI 103.59 25.8 2.02 0.1417 
Cubic vs quadratic 152.27 10 15.2 1.94 0.2394 Aliased 
Residual 39.11 7.8    
Total 138,259.4 30 4,608.6    
LOF tests
Sum of squaresdfMean squareF-valuep-value prob > F
Linear model 308.7648 20 15.4 1.97 0.2322  
Model summary statistics
Std. dev.R2Adjusted R2Predicted R2PRESS
Linear model 3.73 0.788 0.754 0.722 456.7  
Nitrate adsorption
Sequential model sum of squares
SourceSum of squaresdfMean squareF-valuep-value prob > F
Mean vs total 129,047 129,047    
Linear vs mean 1,361.1 340.2 5.66 0.0022 Suggested 
2FI vs linear 1,033.8 172.3 6.98 0.0005 Suggested 
Quadratic vs 2FI 117.87 29.4 1.25 0.3290  
Cubic vs quadratic 299.27 42.7 6.60 0.0081 Aliased 
Residual 51.8 6.4    
Total 131,910.9 30 4,397.03    
LOF tests
SourceSum of squaresdfMean squareF-valuep-value prob > F
Linear 1,451.009 17 85.35347 13.18002 0.0005  
2FI 417.1563 11 37.9233 5.855998 0.0094  
Model summary statistics
SourceStd. dev.R2Adjusted R2Predicted R2PRESS
Linear 7.753236 0.475259 0.391301 0.150109 2,434  
2FI 4.96813 0.836251 0.750068 −0.51048 4,325.9  
Fluoride adsorption on Fe(III)-doped scoria
Sequential model sum of squares
SourceSum of squaresdfMean squareF-valuep-value prob > F
Mean vs total 136,613.3 136,613.3    
Linear vs mean 1,298.2 324.5 23.3 <0.0001 Suggested 
2FI vs linear 52.89 8.8 0.56 0.7508  
Quadratic vs 2FI 103.59 25.8 2.02 0.1417 
Cubic vs quadratic 152.27 10 15.2 1.94 0.2394 Aliased 
Residual 39.11 7.8    
Total 138,259.4 30 4,608.6    
LOF tests
Sum of squaresdfMean squareF-valuep-value prob > F
Linear model 308.7648 20 15.4 1.97 0.2322  
Model summary statistics
Std. dev.R2Adjusted R2Predicted R2PRESS
Linear model 3.73 0.788 0.754 0.722 456.7  
Nitrate adsorption
Sequential model sum of squares
SourceSum of squaresdfMean squareF-valuep-value prob > F
Mean vs total 129,047 129,047    
Linear vs mean 1,361.1 340.2 5.66 0.0022 Suggested 
2FI vs linear 1,033.8 172.3 6.98 0.0005 Suggested 
Quadratic vs 2FI 117.87 29.4 1.25 0.3290  
Cubic vs quadratic 299.27 42.7 6.60 0.0081 Aliased 
Residual 51.8 6.4    
Total 131,910.9 30 4,397.03    
LOF tests
SourceSum of squaresdfMean squareF-valuep-value prob > F
Linear 1,451.009 17 85.35347 13.18002 0.0005  
2FI 417.1563 11 37.9233 5.855998 0.0094  
Model summary statistics
SourceStd. dev.R2Adjusted R2Predicted R2PRESS
Linear 7.753236 0.475259 0.391301 0.150109 2,434  
2FI 4.96813 0.836251 0.750068 −0.51048 4,325.9  

The above-mentioned models explain excellently the experimental range studied which could be observed from the comparison of the graphical representation of actual versus predicted value (Figure 5(a) and 5(b)). Normal probability plot of the studentized residuals is an important diagnostic tool to assess the suitability of developed mathematical models and also for judging the normality of the residuals. Figure 5(c) and 5(d) show that none of the individual residuals exceeded the residual variance and also suggesting that the errors have a normal distribution with a zero mean and a constant value which mean excellent adequacy of the regression models were utilized. Moreover, plots of the residuals against the run number are depicted in Figure 5(e) and 5(f). This plot was used for lurking variables that might have influenced the response during the experiment. As can be seen, the residuals were scattered randomly around ±3.00, indicated that the residual distribution of the regression models follows normal and independent patterns (Kakavandi et al. 2016). The LOF for fluoride adsorption model is insignificant (p-values = 0.232) indicating the good predictability. However, the LOF for nitrate adsorption is significant and there is a 0.94% chance that a ‘LOF F-value’ this large could occur due to noise. Moreover, the ‘Adequate Precision (AP)’ ratio of developed models for fluoride and nitrate adsorption on Fe(III)-doped scoria were 20.25 and 15.24 respectively, were much greater than 4, confirming the adequacy of the models and high signal to noise ratio. The values of coefficient of variation (CV) for both models were less than 10%.

Figure 5

Correlation of actual and predicted adsorption efficiency for fluoride and nitrate (a and b), normal probability plots of residuals for fluoride and nitrate removal (c and d) and plots of studentized residuals versus experimental run number for fluoride and nitrate removal (e and f).

Figure 5

Correlation of actual and predicted adsorption efficiency for fluoride and nitrate (a and b), normal probability plots of residuals for fluoride and nitrate removal (c and d) and plots of studentized residuals versus experimental run number for fluoride and nitrate removal (e and f).

The student's t distribution and the corresponding values, coupled with the parameter estimate for both pollutants, are presented in Table S3 (available with the online version of this paper). The P-values were used for checking the significance of each coefficient, and also to understand the pattern of the mutual interactions between the test variables. The smaller magnitude of the P-value and larger the t-value, the more significant is the corresponding coefficient (Zarei et al. 2010). Referring to Table S3, in this study, linear term of pH (effect ∼ −15) and time (effect ∼ 16) had the highest negative and positive effect on fluoride and nitrate adsorption, respectively.

Graphical presentation of the experimental parameters affecting adsorption process

Three dimensional surfaces plots are graphical representations to predict the adsorption efficiency for different values of the tested variables and provide a simple method to optimize the efficiency of the treatment (Montgomery 2008; Kumar & Bansal 2013). In these plots, two factors or variables are constant and the other two factors will be changed in the experimental ranges. The interactive effect of pH and concentration ( and ) on the removal efficiency is depicted on Figure 6(a) and 6(b). As can be seen, for both pollutants, the removal efficiencies increase with decreasing initial pH. The pH condition is one of the most important variables in sorption studies because it often dramatically affects the amounts of sorbed pollutants through its effect on the chemical species of solute and surface properties of adsorbent (Jenne 1998). Removal of decreases from about 73% to 63% as pH initial increases from 5.0 to 9.0 and for decreases from about 73% to 61% with the same increase in initial pH. The reduction in fluoride and nitrate adsorption in the alkaline pH range may be related to the competition from ions present in the aqueous solution with fluoride and nitrate for the sorption site and also negatively charged adsorbent surface (more than pHpzc) failed to adsorb fluoride and nitrate ions due to electrostatic repulsion (Maliyekkal et al. 2006; Hu et al. 2016). Moreover, the surface of the scoria becomes positively charged in acidic pH which attracts the negatively charged fluoride and nitrate ions by electrostatic attraction resulting in the enhanced removal in acidic pH. At circumneutral pH, fluoride adsorption can be attribute to the ligand-exchange reaction between fluoride and hydroxyl ions (Suriyaraj & Selvakumar 2016). The adsorption behavior of fluoride under various pH values can be summarized as the following reactions (Eskandarpour et al. 2008):  
formula
(9)
 
formula
(10)
 
formula
(11)
 
formula
(12)
where ≡MOH and ≡ are natural and protonated site on Fe(III)-doped scoria and ≡MF is the active site-fluoride complex.
Figure 6

The 3D plots showing effect of (a and b) pH and initial concentration and (c and d) adsorbent dose and initial concentration.

Figure 6

The 3D plots showing effect of (a and b) pH and initial concentration and (c and d) adsorbent dose and initial concentration.

However, adsorption of nitrate is similar to fluoride due to their identical ionic charge. The ligand exchange reactions between the nitrate ions and surface charge of the Fe(III)-doped scoria took a major role in the adsorption process (Suriyaraj & Selvakumar 2016). Figure 6(c) and 6(d) represent the simultaneous effect of fluoride and nitrate concentration and adsorbent dose on the response, with pH and contact time fixed at 7.0 and 45.0 min, respectively. It can be deduced from the graphical representation that the effect of adsorbent dose on the % removal in the studied rang is negligible. This indicated that adsorption of fluoride and nitrate on Fe(III)-doped scoria is not limited by the number of active sites on the surface and there are enough surface area for the adsorption processes (Vildozo et al. 2010). A very low increasing in fluoride adsorption capacity was observed with increase in the initial fluoride concentration which is related to increasing in gradient concentration and the contact between fluoride ions and the scoria surface that provide faster transport with increasing diffusion or mass transfer coefficient (Tuna et al. 2013). However, this trend in nitrate uptake was vice versa and with increasing initial nitrate concentration the removal efficiency was decreased within the studied range.

Optimization of adsorption processes

In to determine the optimum value of variables that yields the highest fluoride and nitrate removal, optimization of operating parameters is necessary. Here, Design Expert 7.0.0 with a multiple-response method was used and the influencing factors were set to values within the studied range, and response (removal of fluoride and nitrate) was set to a maximum value. Under these approaches, the optimum values of the independent variables were obtained as follow: pH 5.0, initial fluoride concentration 5.45 mg/L, adsorbent dose 9.50 mg/L and contact time 60 min, which yielded to 78.36% fluoride adsorption. In the case of nitrate, the maximum removal was 81.4% at an initial pH of 5.05, initial nitrate concentration of 77.21 mg/L, adsorbent dose of 5.03 and contact time of 59.48 min. The values of desirability for fluoride and nitrate were 0.941 and 1.00, respectively. Verification experiments were performed and results showed the removal efficiency of 81.66% and 77.12% for fluoride and nitrate, respectively. This indicates the accuracy and validity of response surface methodology as a reliable statistical technique for optimizing the operational conditions of adsorption process.

Kinetics and isotherm adsorption

The contact time is an important parameter especially in practical application, is the time required for the ions (fluoride and nitrate) to reach an equilibrium state on the adsorbent. According to Figure 7 the fluoride and nitrate were rapidly adsorbed in the first phase (∼30 min) and then reach to equilibrium in second phase about 60 min. The initial fast phase was due to the availability of adsorption surface sites (functional groups) and second slower phase was due to the quick exhaustion of the adsorption sites (Ngah et al. 2011). Therefore, the optimum contact time was limited to 60 min throughout this study. The sorption capacity of RS and Fe(III)-doped scoria for fluoride were 0.165, 0.28 and for nitrate were 2.23 and 4.3 mg/g, respectively.

Figure 7

Effect of contact time on the adsorption of fluoride and nitrate onto RS and FeS.

Figure 7

Effect of contact time on the adsorption of fluoride and nitrate onto RS and FeS.

The kinetic parameters of fluoride and nitrate adsorption on raw and Fe(III)-doped scoria are presented in Table 5. The experimental data in Figure 7 were fitted with the linear form using Equations (1) to (4) and results are illustrated in Figure S1 (SI) (available with the online version of this paper). The validity of studied models was compared by the coefficient of determination (R2) and close agreement between the experimental and theoretical values of qe. As can be seen from Table 5, the values of R2 of pseudo-first-order kinetic model for almost all the cases were the highest (>0.977) and followed by those of the intraparticle diffusion model, Elovich model and pseudo-second-order kinetic model, respectively. Also, the qe(theor) values of pseudo-first-order kinetic model agree well with the experimental values. This indicated that pseudo-first-order kinetic model was appropriate to represent the experimental kinetic data and describe well the adsorption behavior. This is means that the rate constant is independent of concentration and mechanism of sorption may not be ion exchange or chemically rate controlling. The data show that the rate constant is more or less independent of initial concentrations of fluoride and nitrate. As was explained previously in the student's t distribution (Table S3), the effect of initial concentration for both fluoride and nitrate was insignificant in the studied range. This can be a good agreement between the proposed model and statistical analyses by RSM and Lagergren pseudo-first-order equation. In most cases in the literature, which comparison has been done between the pseudo-first-order and pseudo-second-order, superiority of K2 over K1 has been found (Ho 2006; Simonin 2016). However, in some studies, the adsorption follows the Lagergren pseudo-first-order equation (Trivedi et al. 1973; Mishra et al. 1996; Mishra & Tiwary 1999). A larger adsorption rate constant k1 for Fe(III)-doped scoria for both fluoride and nitrate indicated quicker adsorption onto adsorbent and also lesser values of K2 represent a faster adsorption. Here, based on Table 5, fluoride and nitrate sorption on raw and Fe(III)-doped scoria may also involve intraparticle diffusion, which could be the rate-limiting step of the sorption process. After pseudo-first-order model, the intraparticle diffusion model had a good fit (high R2 values) with the experimental data. The low values of boundary layer effect (0.001 < C < 0.072) indicated less resistance to mass transfer, thus leading to an increase in the mobility of fluoride and nitrate toward raw and modified scoria (Ahmad et al. 2013). As can be seen from Figure S1(c) (SI) the plots do not pass through the origin (0,0); that may be due to the difference in the mass transfer rate in the initial and final stages of adsorption (Koyuncu & Kul 2014; Omidvar Borna et al. 2016).

Table 5

Kinetic parameters for fluoride and nitrate adsorption onto raw and Fe(III)-doped scoria

Adsorbentqe(exp) (mg/g)Pseudo-first-order model
Pseudo-second-order model
K1 × 10−2qe (theor)R2RMSEK2qe (theor)hR2RMSE
RS (0.155 4.1 0.16 0.980 0.002 0.55 0.17 0.016 0.936 0.002 
FeS (0.27 5.2 0.26 0.977 0.009 0.52 0.29 0.044 0.959 0.004 
RS (2.02 3.6 2.27 0.992 0.018 0.026 2.39 0.152 0.875 0.072 
FeS (3.92 3.9 4.13 0.996 0.076 0.02 4.38 0.398 0.931 0.036 
Intraparticle diffusion model
Elovich model
Adsorbentqe(exp) (mg/g)KidCqe (theor)R2αβR2
RS (0.155 0.02 0.001 0.156 0.987 0.045 27.02 0.981  
FeS (0.27 0.034 0.025 0.288 0.954 0.07 15.38 0.968  
RS (2.02 0.274 0.072 2.19 0.991 0.33 2.00 0.954  
FeS (3.92 0.523 0.012 4.06 0.995 0.75 1.02 0.976  
Adsorbentqe(exp) (mg/g)Pseudo-first-order model
Pseudo-second-order model
K1 × 10−2qe (theor)R2RMSEK2qe (theor)hR2RMSE
RS (0.155 4.1 0.16 0.980 0.002 0.55 0.17 0.016 0.936 0.002 
FeS (0.27 5.2 0.26 0.977 0.009 0.52 0.29 0.044 0.959 0.004 
RS (2.02 3.6 2.27 0.992 0.018 0.026 2.39 0.152 0.875 0.072 
FeS (3.92 3.9 4.13 0.996 0.076 0.02 4.38 0.398 0.931 0.036 
Intraparticle diffusion model
Elovich model
Adsorbentqe(exp) (mg/g)KidCqe (theor)R2αβR2
RS (0.155 0.02 0.001 0.156 0.987 0.045 27.02 0.981  
FeS (0.27 0.034 0.025 0.288 0.954 0.07 15.38 0.968  
RS (2.02 0.274 0.072 2.19 0.991 0.33 2.00 0.954  
FeS (3.92 0.523 0.012 4.06 0.995 0.75 1.02 0.976  
Adsorption isotherm is a beneficial tool for optimizing an adsorption process and designing a desired adsorption system. Here, Langmuir and Freundlich models were used to analyze the adsorption of fluoride and nitrate onto RS and Fe(III)-doped scoria (Table 6). As seen in Table 6, based on determination coefficient (R2), the adsorption of fluoride and nitrate on Fe(III)-doped scoria followed Freundlich and Langmuir isotherm models, respectively. More agreement of fluoride equilibrium data with Freundlich isotherm implies the heterogeneous systems and reversible adsorption process. Freundlich constant ‘n’ greater than 1 denotes the physical adsorption process. When the values of n are equal to unity, the adsorption process is linear and n < 1 signifies the chemical adsorption process (Shariful et al. 2017). The free energy change (ΔG) of fluoride adsorption onto RS and FeS with knowing the KF amounts could be calculated through following equation:  
formula
(13)
where R is the universal gas constant, 8.314 J/K mol and T is the temperature (292 K). The negative free energy values (Table 6) indicate the feasibility of the process and the spontaneous nature of adsorption. The more agreement of nitrate equilibrium data with Langmuir isotherm model implies that the structure of adsorbent is homogeneous and adsorption of nitrate on RS and FeS occurred as a monolayer. The activation energy of adsorption, E (kJ/mol), can be calculated from the β values as follows:  
formula
(14)
when E parameter lies between 8 and 16 kJ/mol, the adsorption process is physical in nature and when it is below 8 kJ/mol, the adsorption process has a chemical character (Sari & Tuzen 2014). Referring to Table 6, all adsorption data were proceeded physically through weak van der Waals forces.
Table 6

Adsorption isotherm parameters for the adsorption of fluoride and nitrate onto raw and Fe(III)-doped scoria

Isotherm modelsRS ()FeS ()RS ()FeS ()
Langmuir 
 Qmax(mg/g) 0.189 0.317 3.27 11.3 
 KL 1.4 8.6 0.026 0.020 
 RL 0.09 0.016 0.27 0.32 
 R2 0.682 0.573 0.897 0.963 
 X2 0.001 0.074 0. 017 0.233 
Freundlich 
 KF 0.12 0.26 0.42 0.68 
 n 5.0 4.3 2.6 1.9 
 ΔG −11.6 −13.5 −14.6 −15.8 
 R2 0.817 0.819 0.852 0.870 
 X2 0.001 0.021 0.020 0.297 
Isotherm modelsRS ()FeS ()RS ()FeS ()
Langmuir 
 Qmax(mg/g) 0.189 0.317 3.27 11.3 
 KL 1.4 8.6 0.026 0.020 
 RL 0.09 0.016 0.27 0.32 
 R2 0.682 0.573 0.897 0.963 
 X2 0.001 0.074 0. 017 0.233 
Freundlich 
 KF 0.12 0.26 0.42 0.68 
 n 5.0 4.3 2.6 1.9 
 ΔG −11.6 −13.5 −14.6 −15.8 
 R2 0.817 0.819 0.852 0.870 
 X2 0.001 0.021 0.020 0.297 

Comparison with other adsorbents

So far, numerous studies have been conducted to investigate the effectiveness of various adsorbents for removal of fluoride and nitrate. Table 7 indicated some of these studies and their experimental conditions and adsorption capacities. According to it, the adsorption capacity of Fe(III)-doped scoria for both fluoride and nitrate is comparable with the other studies especially when the adsorbent is natural. However, some synthesized nanoscale adsorbents have higher adsorption capacities that are due to higher surface area and lesser particle size. Hence, Fe(III)-doped scoria is a promising adsorbent for fluoride and nitrate removal in drinking water.

Table 7

Comparison of Fe(III)-doped scoria adsorbent performance with some recently reported adsorbents for removal of fluoride and nitarte

IonAdsorbentExperimental conditionsqe (mg/g)Reference
Fluoride  − modified bentonite Solution pH range = 2–10, pHZPC = 8.5. 2.91 Gitari et al. (2015)  
Montmorillonite (ANC) Concentration range = 1–6 mg/L, solution pH = 4, contact time = 30 min, adsorbent dose = 0.3 g, pHZPC = 5.5–6.5, Surface area = 110 m2/g 1.324 Ramdani et al. (2010)  
Activated kaolinite Dosage = 1 g, Contact time = 30 min, pHZPC = 4, Solution pH = 3, concentration range = 2–10 mg/L, Surface area = 32.43 m2/g 0.134 Meenakshi et al. (2008)  
Activated alumina (ɣAL2O3pHPZC = 8, F concentration range = 50–100 mg/L, contact time = 16–24 h, pH = 5–6, T = 30°C 0.86 mmol/g Ku & Chiou (2002)  
La(III) impregnated on alumina Concentration range = 2 mM/L, contact time = 20 h, pH = 5.7–8, T = room temperature 0.35 mM/g Puri & Balani (2000)  
Fe(III)-doped scoria Adsorbent dose = 4 g/L concentration range = 7 mg/L, pH = 7, contact time = 60 min, room temperature 0.27 Present study 
Nitrate Wheat straw charcoal Concentration range = 0–25 mg/L, contact time = 10 min, T = 15°C 1.10 Mishra & Patel (2009)  
Original and activated red mud Concentration range = 5–250 mg/L, contact time = 60 min, pH = 6.0, room temperature 1.85 and 5.85 mmol/g Cengeloglu et al. (2006)  
Cross-linked and quaternized chinese Reed Concentration range = 10–40 mg/dm3, contact time = 10 min, pH = 5.8, T = 25°C 7.55 mg/g Namasivayam & Höll (2005
Iron-modified pumice Concentration range = 50–300 mg/L, pH = 5, contact time = 50 min, room temperature 21.09 Golestanifar et al. (2016
Fe(III)-doped scoria Adsorbent dose = 4 g/L concentration range = 100 mg/L, pH = 7, contact time = 60 min, room temperature 3.92 Present study 
IonAdsorbentExperimental conditionsqe (mg/g)Reference
Fluoride  − modified bentonite Solution pH range = 2–10, pHZPC = 8.5. 2.91 Gitari et al. (2015)  
Montmorillonite (ANC) Concentration range = 1–6 mg/L, solution pH = 4, contact time = 30 min, adsorbent dose = 0.3 g, pHZPC = 5.5–6.5, Surface area = 110 m2/g 1.324 Ramdani et al. (2010)  
Activated kaolinite Dosage = 1 g, Contact time = 30 min, pHZPC = 4, Solution pH = 3, concentration range = 2–10 mg/L, Surface area = 32.43 m2/g 0.134 Meenakshi et al. (2008)  
Activated alumina (ɣAL2O3pHPZC = 8, F concentration range = 50–100 mg/L, contact time = 16–24 h, pH = 5–6, T = 30°C 0.86 mmol/g Ku & Chiou (2002)  
La(III) impregnated on alumina Concentration range = 2 mM/L, contact time = 20 h, pH = 5.7–8, T = room temperature 0.35 mM/g Puri & Balani (2000)  
Fe(III)-doped scoria Adsorbent dose = 4 g/L concentration range = 7 mg/L, pH = 7, contact time = 60 min, room temperature 0.27 Present study 
Nitrate Wheat straw charcoal Concentration range = 0–25 mg/L, contact time = 10 min, T = 15°C 1.10 Mishra & Patel (2009)  
Original and activated red mud Concentration range = 5–250 mg/L, contact time = 60 min, pH = 6.0, room temperature 1.85 and 5.85 mmol/g Cengeloglu et al. (2006)  
Cross-linked and quaternized chinese Reed Concentration range = 10–40 mg/dm3, contact time = 10 min, pH = 5.8, T = 25°C 7.55 mg/g Namasivayam & Höll (2005
Iron-modified pumice Concentration range = 50–300 mg/L, pH = 5, contact time = 50 min, room temperature 21.09 Golestanifar et al. (2016
Fe(III)-doped scoria Adsorbent dose = 4 g/L concentration range = 100 mg/L, pH = 7, contact time = 60 min, room temperature 3.92 Present study 

CONCLUSION

This work deal with the adsorption of fluoride and nitrate onto raw and Fe(III)-doped scoria from aqueous solutions using batch mode experiments. The scoria in doped form with iron(III) could be an effective and low cost adsorbent for the removal of anionic pollutants especially fluoride and nitrate. The following conclusions were drawn from the results of this study:

  • 1

    The RS was doped with Fe(III) with facile precipitation method which enhanced the adsorption capacity of scoria.

  • 2

    The raw and modified scoria was characterized using SEM, FT-IR and XRD. It was revealed that the Fe(III)-doped scoria contain Si/Al ratio of 2.5 which provide the high absolute values of the adsorption energies.

  • 3

    Effect of operational parameters on the adsorption was evaluated by the RSM and the optimum values for fluoride removal were; pH 5.0, initial fluoride concentration 5.45 mg/L, adsorbent dose 9.50 mg/L and contact time 60 min, which yielded to 78.36% fluoride removal and for nitrate were; pH 5.0, initial nitrate concentration of 77.21 mg/L, adsorbent dose of 5.03 and contact time of 59.48 min with the maximum removal of 81.4%. Adsorption of fluoride and nitrate onto Fe(III)-doped scoria followed linear and interactive 2FI models, respectively.

  • 4

    The examination of kinetic models indicated that the pseudo-first-order kinetic model is more suitable for the adsorption process.

  • 5

    The Freundlich model better described the isotherm data of fluoride and, for nitrate, the equilibrium data were fitted better with the Langmuir model.

Conclusively, Research results demonstrate that Fe(III)-doped scoria is a promising adsorbent for removing fluoride and nitrate from drinking water, because it has high adsorption capacity, easy availability, and very low cost.

ACKNOWLEDGEMENTS

The authors gratefully acknowledge the Research Council of Kermanshah University of Medical Sciences (Grant Number: 93320) for the financial assistance. This work was performed in partial fulfillment of the requirement for MS of Environmental Health Engineering of Heshmat Mohammadi, in School of Public Health, Kermanshah University of Medical Sciences, Kermanshah, Iran.

REFERENCES

REFERENCES
Ahmad
,
M.
,
Lee
,
S. S.
,
Oh
,
S.-E.
,
Mohan
,
D.
,
Moon
,
D. H.
,
Lee
,
Y. H.
&
Ok
,
Y. S.
2013
Modeling adsorption kinetics of trichloroethylene onto biochars derived from soybean stover and peanut shell wastes
.
Environmental Science and Pollution Research
20
(
12
),
8364
8373
.
Ali
,
I.
2012
New generation adsorbents for water treatment
.
Chemical Reviews
112
(
10
),
5073
5091
.
Amini
,
M.
,
Mueller
,
K.
,
Abbaspour
,
K. C.
,
Rosenberg
,
T.
,
Afyuni
,
M.
,
Møller
,
K. N.
,
Sarr
,
M.
&
Johnson
,
C. A.
2008
Statistical modeling of global geogenic fluoride contamination in groundwaters
.
Environmental Science & Technology
42
(
10
),
3662
3668
.
Amini
,
H.
,
Haghighat
,
G. A.
,
Yunesian
,
M.
,
Nabizadeh
,
R.
,
Mahvi
,
A. H.
,
Dehghani
,
M. H.
,
Davani
,
R.
,
Aminian
,
A.-R.
,
Shamsipour
,
M.
,
Hassanzadeh
,
N.
,
Faramarzi
,
H.
&
Mesdaghinia
,
A.
2016
Spatial and temporal variability of fluoride concentrations in groundwater resources of Larestan and Gerash regions in Iran from 2003 to 2010
.
Environmental Geochemistry and Health
38
(
1
),
25
37
.
Awuah
,
J. B.
,
Dzade
,
N. Y.
,
Tia
,
R.
,
Adei
,
E.
,
Kwakye-Awuah
,
B.
,
Richard
,
A.
,
Catlow
,
C.
&
de Leeuw
,
N. H.
2016
A density functional theory study of arsenic immobilization by the Al(iii)-modified zeolite clinoptilolite
.
Physical Chemistry Chemical Physics
18
(
16
),
11297
11305
.
Barbier
,
O.
,
Arreola-Mendoza
,
L.
&
Del Razo
,
L. M.
2010
Molecular mechanisms of fluoride toxicity
.
Chemico-Biological Interactions
188
(
2
),
319
333
.
Bhatnagar
,
A.
,
Kumar
,
E.
&
Sillanpää
,
M.
2011
Fluoride removal from water by adsorption – a review
.
Chemical Engineering Journal
171
(
3
),
811
840
.
Cai
,
H.-m.
,
Chen
,
G.-j.
,
Peng
,
C.-y.
,
Zhang
,
Z.-z.
,
Dong
,
Y.-y.
,
Shang
,
G.-z.
,
Zhu
,
X.-h.
,
Gao
,
H.-j.
&
Wan
,
X.-c.
2015
Removal of fluoride from drinking water using tea waste loaded with Al/Fe oxides: a novel, safe and efficient biosorbent
.
Applied Surface Science
328
,
34
44
.
Cengeloglu
,
Y.
,
Tor
,
A.
,
Ersoz
,
M.
&
Arslan
,
G.
2006
Removal of nitrate from aqueous solution by using red mud
.
Separation and Purification Technology
51
(
3
),
374
378
.
Djobo
,
J. N. Y.
,
Tchadjié
,
L. N.
,
Tchakoute
,
H. K.
,
Kenne
,
B. B. D.
,
Elimbi
,
A.
&
Njopwouo
,
D.
2014
Synthesis of geopolymer composites from a mixture of volcanic scoria and metakaolin
.
Journal of Asian Ceramic Societies
2
(
4
),
387
398
.
Eskandarpour
,
A.
,
Onyango
,
M. S.
,
Ochieng
,
A.
&
Asai
,
S.
2008
Removal of fluoride ions from aqueous solution at low pH using schwertmannite
.
Journal of Hazardous Materials
152
(
2
),
571
579
.
Fernández-Jiménez
,
A.
&
Palomo
,
A.
2005
Composition and microstructure of alkali activated fly ash binder: effect of the activator
.
Cement and Concrete Research
35
(
10
),
1984
1992
.
Gitari
,
W.
,
Ngulube
,
T.
,
Masindi
,
V.
&
Gumbo
,
J.
2015
Defluoridation of groundwater using Fe3+-modified bentonite clay: optimization of adsorption conditions
.
Desalination and Water Treatment
53
(
6
),
1578
1590
.
Golestanifar
,
H.
,
Asadi
,
A.
,
Alinezhad
,
A.
,
Haybati
,
B.
&
Vosoughi
,
M.
2016
Isotherm and kinetic studies on the adsorption of nitrate onto nanoalumina and iron-modified pumice
.
Desalination and Water Treatment
57
(
12
),
5480
5487
.
Hassan
,
M. L.
,
Kassem
,
N. F.
&
Abd El-Kader
,
A. H.
2010
Novel Zr(IV)/sugar beet pulp composite for removal of sulfate and nitrate anions
.
Journal of Applied Polymer Science
117
(
4
),
2205
2212
.
Ho
,
Y.-S.
2006
Review of second-order models for adsorption systems
.
Journal of Hazardous Materials
136
(
3
),
681
689
.
Jadhav
,
S. V.
,
Bringas
,
E.
,
Yadav
,
G. D.
,
Rathod
,
V. K.
,
Ortiz
,
I.
&
Marathe
,
K. V.
2015
Arsenic and fluoride contaminated groundwaters: a review of current technologies for contaminants removal
.
Journal of Environmental Management
162
,
306
325
.
Jain
,
S.
,
Bansiwal
,
A.
,
Biniwale
,
R. B.
,
Milmille
,
S.
,
Das
,
S.
,
Tiwari
,
S.
&
Siluvai Antony
,
P.
2015
Enhancing adsorption of nitrate using metal impregnated alumina
.
Journal of Environmental Chemical Engineering
3
(
4, Part A
),
2342
2349
.
Jenne
,
E. A.
1998
Adsorption of Metals by Geomedia: Data Analysis, Modeling, Controlling Factors, and Related Issues
.
Academic Press
,
San Diego, CA, USA
.
Kakavandi
,
B.
,
Jahangiri-rad
,
M.
,
Rafiee
,
M.
,
Esfahani
,
A. R.
&
Babaei
,
A. A.
2016
Development of response surface methodology for optimization of phenol and p-chlorophenol adsorption on magnetic recoverable carbon
.
Microporous and Mesoporous Materials
231
,
192
206
.
Koyuncu
,
H.
&
Kul
,
A. R.
2014
An investigation of Cu(II) adsorption by native and activated bentonite: kinetic, equilibrium and thermodynamic study
.
Journal of Environmental Chemical Engineering
2
(
3
),
1722
1730
.
Ku
,
Y.
&
Chiou
,
H.-M.
2002
The adsorption of fluoride ion from aqueous solution by activated alumina
.
Water, Air, and Soil Pollution
133
(
1–4
),
349
361
.
Kumar
,
E.
,
Bhatnagar
,
A.
,
Hogland
,
W.
,
Marques
,
M.
&
Sillanpää
,
M.
2014
Interaction of inorganic anions with iron-mineral adsorbents in aqueous media – a review
.
Advances in Colloid and Interface Science
203
,
11
21
.
Kwon
,
J.-S.
,
Yun
,
S.-T.
,
Kim
,
S.-O.
,
Mayer
,
B.
&
Hutcheon
,
I.
2005
Sorption of Zn(II) in aqueous solutions by scoria
.
Chemosphere
60
(
10
),
1416
1426
.
Kwon
,
J.-S.
,
Yun
,
S.-T.
,
Lee
,
J.-H.
,
Kim
,
S.-O.
&
Jo
,
H. Y.
2010
Removal of divalent heavy metals (Cd, Cu, Pb, and Zn) and arsenic(III) from aqueous solutions using scoria: kinetics and equilibria of sorption
.
Journal of Hazardous Materials
174
(
1–3
),
307
313
.
Laws
,
E. A.
2000
Aquatic Pollution: an Introductory Text
.
John Wiley & Sons
,
Hoboken, NJ, USA
.
Maliyekkal
,
S. M.
,
Sharma
,
A. K.
&
Philip
,
L.
2006
Manganese-oxide-coated alumina: a promising sorbent for defluoridation of water
.
Water Research
40
(
19
),
3497
3506
.
Massoudinejad
,
M.
,
Asadi
,
A.
,
Vosoughi
,
M.
,
Gholami
,
M.
,
Kakavandi
,
B.
&
Karami
,
M. A.
2015
A comprehensive study (kinetic, thermodynamic and equilibrium) of arsenic (V) adsorption using KMnO4 modified clinoptilolite
.
Korean Journal of Chemical Engineering
32
(
10
),
2078
2086
.
Meenakshi
,
S.
,
Sundaram
,
C. S.
&
Sukumar
,
R.
2008
Enhanced fluoride sorption by mechanochemically activated kaolinites
.
Journal of Hazardous Materials
153
(
1–2
),
164
172
.
Mishra
,
P. C.
&
Patel
,
R. K.
2009
Use of agricultural waste for the removal of nitrate-nitrogen from aqueous medium
.
Journal of Environmental Management
90
(
1
),
519
522
.
Montgomery
,
D. C.
2008
Design and Analysis of Experiments
.
John Wiley & Sons
,
Hoboken, NJ, USA
.
Moradi
,
M.
,
Hemati
,
L.
,
Pirsaheb
,
M.
&
Sharafi
,
K.
2015
Removal of hexavalent chromium from aqueous solution by powdered scoria-equilibrium isotherms and kinetic studies
.
World Applied Sciences Journal
33
(
3
),
393
400
.
Moufti
,
M. R.
,
Sabtan
,
A. A.
,
El-Mahdy
,
O. R.
&
Shehata
,
W. M.
2000
Assessment of the industrial utilization of scoria materials in central Harrat Rahat, Saudi Arabia
.
Engineering Geology
57
(
3–4
),
155
162
.
Nadafi
,
K.
,
Vosoughi
,
M.
,
Asadi
,
A.
,
Borna
,
M. O.
&
Shirmardi
,
M.
2014
Reactive Red 120 dye removal from aqueous solution by adsorption on nano-alumina
.
Journal of Water Chemistry and Technology
36
(
3
),
125
133
.
Namasivayam
,
C.
&
Höll
,
W. H.
2005
Quaternized biomass as an anion exchanger for the removal of nitrate and other anions from water
.
Journal of Chemical Technology and Biotechnology
80
(
2
),
164
168
.
Ngah
,
W. S. W.
,
Fatinathan
,
S.
&
Yosop
,
N. A.
2011
Isotherm and kinetic studies on the adsorption of humic acid onto chitosan-H2SO4 beads
.
Desalination
272
(
1–3
),
293
300
.
Omidvar Borna
,
M.
,
Pirsaheb
,
M.
,
Vosoughi Niri
,
M.
,
Khosravi Mashizie
,
R.
,
Kakavandi
,
B.
,
Zare
,
M. R.
&
Asadi
,
A.
2016
Batch and column studies for the adsorption of chromium(VI) on low-cost Hibiscus Cannabinus kenaf, a green adsorbent
.
Journal of the Taiwan Institute of Chemical Engineers
68
,
80
89
.
Panias
,
D.
,
Giannopoulou
,
I. P.
&
Perraki
,
T.
2007
Effect of synthesis parameters on the mechanical properties of fly ash-based geopolymers
.
Colloids and Surfaces A: Physicochemical and Engineering Aspects
301
(
1–3
),
246
254
.
Puri
,
B. K.
&
Balani
,
S.
2000
Trace determination of fluoride using lanthanum hydroxide supported on alumina
.
Journal of Environmental Science & Health Part A
35
(
1
),
109
121
.
Ramdani
,
A.
,
Taleb
,
S.
,
Benghalem
,
A.
&
Ghaffour
,
N.
2010
Removal of excess fluoride ions from Saharan brackish water by adsorption on natural materials
.
Desalination
250
(
1
),
408
413
.
Salarian
,
A.-A.
,
Hami
,
Z.
,
Mirzaei
,
N.
,
Mohseni
,
S. M.
,
Asadi
,
A.
,
Bahrami
,
H.
,
Vosoughi
,
M.
,
Alinejad
,
A.
&
Zare
,
M.-R.
2016
N-doped TiO2 nanosheets for photocatalytic degradation and mineralization of diazinon under simulated solar irradiation: optimization and modeling using a response surface methodology
.
Journal of Molecular Liquids
220
,
183
191
.
Sari
,
A.
&
Tuzen
,
M.
2014
Cd(II) adsorption from aqueous solution by raw and modified kaolinite
.
Applied Clay Science
88–89
,
63
72
.
Shariful
,
M. I.
,
Sharif
,
S. B.
,
Lee
,
J. J. L.
,
Habiba
,
U.
,
Ang
,
B. C.
&
Amalina
,
M. A.
2017
Adsorption of divalent heavy metal ion by mesoporous-high surface area chitosan/poly (ethylene oxide) nanofibrous membrane
.
Carbohydrate Polymers
157
,
57
64
.
Swain
,
S. K.
,
Patnaik
,
T.
,
Patnaik
,
P. C.
,
Jha
,
U.
&
Dey
,
R. K.
2013
Development of new alginate entrapped Fe(III)–Zr(IV) binary mixed oxide for removal of fluoride from water bodies
.
Chemical Engineering Journal
215–216
,
763
771
.
Trivedi
,
H. C.
,
Patel
,
V. M.
&
Patel
,
R. D.
1973
Adsorption of cellulose triacetate on calcium silicate
.
European Polymer Journal
9
(
6
),
525
531
.
Vildozo
,
D.
,
Ferronato
,
C.
,
Sleiman
,
M.
&
Chovelon
,
J.-M.
2010
Photocatalytic treatment of indoor air: optimization of 2-propanol removal using a response surface methodology (RSM)
.
Applied Catalysis B: Environmental
94
(
3–4
),
303
310
.
Viswanathan
,
N.
&
Meenakshi
,
S.
2008
Enhanced fluoride sorption using La(III) incorporated carboxylated chitosan beads
.
Journal of Colloid and Interface Science
322
(
2
),
375
383
.
Wang
,
X.-H.
,
Song
,
R.-H.
,
Yang
,
H.-C.
,
Shi
,
Y.-J.
,
Dang
,
G.-B.
,
Yang
,
S.
,
Zhao
,
Y.
,
Sun
,
X.-F.
&
Wang
,
S.-G.
2013
Fluoride adsorption on carboxylated aerobic granules containing Ce(III)
.
Bioresource Technology
127
,
106
111
.
Zhang
,
T.
,
Li
,
Q.
,
Xiao
,
H.
,
Lu
,
H.
&
Zhou
,
Y.
2012
Synthesis of Li–Al layered double hydroxides (LDHs) for efficient fluoride removal
.
Industrial & Engineering Chemistry Research
51
(
35
),
11490
8
.
Zhang
,
S.
,
Lu
,
Y.
,
Lin
,
X.
,
Su
,
X.
&
Zhang
,
Y.
2014
Removal of fluoride from groundwater by adsorption onto La(III)- Al(III) loaded scoria adsorbent
.
Applied Surface Science
303
,
1
5
.

Supplementary data