Sawdust of Artocarpus hirsutus impregnated with ferric hydroxide and activated alumina (SFAA) has been studied for defluoridation of water. This paper presents a detailed surface characterization of the adsorbent by studying pore size distribution and surface morphology. By combining the constituents in the right proportion, an adsorbent with a well-developed pore size distribution could be synthesized. The effects of various parameters on fluoride adsorption by SFAA are investigated. The adsorption capacity of SFAA is considerably higher than that of many adsorbents including activated alumina. More importantly, the adsorption capacity remains unchanged for the pH range of 1 to 9, which also makes it attractive for wastewater treatment. Besides a higher efficiency supported by the results of column studies, this adsorbent is economic as the sawdust constitutes 40% by weight of the total adsorbent. Kinetic studies indicate that fluoride adsorption on SFAA follows pseudo second-order model. Breakthrough adsorption capacity of SFAA is 1.21 mg/g, as compared with 0.41 mg/g for activated alumina.

INTRODUCTION

Fluorine occurs in nature in the form of fluorides in a number of minerals and accounts for about 0.3 g/kg of the Earth's crust. Weathering events along with volcanic and fumarolic processes, contribute to the rise in fluoride levels of groundwater. High fluoride concentrations in groundwater occur in large parts of Africa, China, Sri Lanka and India (Meenakshi & Maheshwari 2006). In India, Andhra Pradesh, Tamil Nadu, Karnataka, Kerala, Rajasthan, Gujarat, Uttar Pradesh, Punjab, Orissa and Jammu Kashmir have been reported to contain high fluoride levels in groundwater (Alagumuthu & Rajan 2010a). As prescribed by the World Health Organization (WHO), the maximum permissible limit for fluoride level in drinking water is 1.5 mg/L. Consumption of water with higher fluoride concentrations causes fluorosis, dental fluorosis in particular. Fluoride levels above 10 mg/L in drinking water cause skeletal fluorosis (WHO 2011).

Chemically fluoride ion is a hard base, exhibiting strong affinity towards metal ions, including Al3+ and Fe3+ (Wu et al. 2007; Chen et al. 2011). Consequently chemical additive methods, contact precipitation techniques and adsorption/ion exchange methods are reported to remove fluorides in water. Among these, adsorption is considered to be a fast, efficient and inexpensive method of defluoridation. Numerous adsorbents have been reported for the defluoridation of aqueous solutions (Bhatnagar et al. 2011). Of these, activated alumina based compounds have been extensively studied for the adsorption of fluoride from aqueous solutions (Shimelis et al. 2006; Goswami & Purkait 2012). Use of amorphous Fe/Al mixed hydroxides for fluoride adsorption from aqueous solutions was studied by Sujana et al. (2009).

Many researchers have investigated the use of natural and low-cost materials such as pure chitosan, chitosan with iron (Viswanathan & Meenakshi 2008; Viswanathan et al. 2009), granular ceramic (Chen et al. 2010), zirconium impregnated coconut shell carbon, coconut fiber carbon (Meenakshi & Maheshwari 2007; Sathish et al. 2007; Sathish et al. 2008), cashew nut shell carbon (Alagumuthu & Rajan 2010a), ground nut shell carbon (Alagumuthu & Rajan 2010b), sugar cane charcoal (Mondal et al. 2013), Bermuda grass carbon (Alagumuthu et al. 2011), red mud (Tor et al. 2009) and clays (Gogoi & Baruah 2008) for their fluoride adsorption characteristics. Adsorption capacities for fluoride range from 0.85 to 40.01 mg/g for these materials (Das et al. 2008; Bhatnagar et al. 2011). Natural adsorbents are highly efficient, cost effective and ecofriendly. The generation of chemical sludge is minimum, with the additional advantage of ease of regeneration. In the present study, locally available sawdust impregnated with ferric hydroxide and activated alumina was studied for its efficacy for removal of arsenic from drinking water and the results were reported in our earlier paper (Dhanasekaran et al. 2016). Arsenic removal of greater than 98% was reported for an initial concentration of 2 mg/L. This paper presents the results of investigation for application of the same adsorbent, SFAA for removal of fluoride from drinking water.

MATERIALS AND METHODS

Chemicals

The chemicals used in this study were of analytical grade. Fluoride stock solution of 1,000 mg/L was prepared using sodium fluoride and demineralized (DM) water. This stock solution was diluted with DM water to prepare a fluoride solution of desired concentration.

Pretreatment of sawdust

Different locally available trees sawdust was tested for their suitability as a base material for the adsorbent. Pretreatment of the sawdust was carried out as detailed in our earlier paper (Dhanasekaran et al. 2016). Capacity of this different tree sawdust for adsorption of fluoride was investigated using a solution containing 25.6 mg/L fluoride and results are presented in Table 1. The results show that Artocarpus hirsutus, called ‘Anjili’ locally, exhibited the highest adsorption capacity and therefore it was selected for preparation of the adsorbent. It can be seen that the adsorption capacity for the sawdust is 13.43% at this concentration, which is quite appreciable for a base material. It may be noted that the same sawdust earlier also showed maximum adsorption capacity for arsenic (Dhanasekaran et al. 2016).

Table 1

Selection of sawdust for adsorbent preparation

S. no.Tree nameBinomial nameF̄ adsorbed (mg/g)F̄ adsorption (%)
Coconut Cocos nucifera 0.00 0.08 
Eucalyptus Eucalyptus obliqua 0.007 1.33 
Karuvelam Acacia Arabica 0.011 2.15 
Poovarasam Thespesia populnea 0.028 5.39 
Teak Tectona grandis 0.037 7.30 
Patauk Pterocarpus 0.036 7.34 
Anjili Artocarpus hirsutus 0.069 13.43 
S. no.Tree nameBinomial nameF̄ adsorbed (mg/g)F̄ adsorption (%)
Coconut Cocos nucifera 0.00 0.08 
Eucalyptus Eucalyptus obliqua 0.007 1.33 
Karuvelam Acacia Arabica 0.011 2.15 
Poovarasam Thespesia populnea 0.028 5.39 
Teak Tectona grandis 0.037 7.30 
Patauk Pterocarpus 0.036 7.34 
Anjili Artocarpus hirsutus 0.069 13.43 

C0=25 mg/L, Contact time = 4 h, Temp = 303 K, Volume = 10 mL, Mass = 0.5 g, Speed = 40 rpm.

Adsorbent preparation

Preparation of adsorbent was explained in detail in our previous paper (Dhanasekaran et al. 2016).

Adsorbent characterization studies

Pore characterization by liquid N2 adsorption was used to estimate the Brunauer-Emmett-Teller (BET) surface area, pore size distribution and pore volume of the adsorbents by Sorptomatic (1990) BET analyzer (France). Pore volume in micro- and mesopores is obtained from volume of liquid nitrogen adsorbed up to P/P0 = 0.98. Micropore volume and surface area in mesopores including external surface are determined from t-plots, which is a method of comparing an isotherm of a microporous material with a standard type II isotherm (Lippens & De Boer 1965). Lippens and De Boer proposed the plotting of nitrogen adsorbed volume (Va) at different P/P0 values as a function of the layer thickness (t). The linear Lippens and De Boer equation can be given by 
formula
The Halsey equation for N2 adsorption at 77 K can be expressed as 
formula
From the slope and intercept of V-plot the mesopore surface area, Smeso and micropore volume, Vmicro are calculated (Siminiceanu et al. 2008). 
formula
 
formula

The micropore surface area is calculated by subtracting mesopore surface area from the BET surface area.

Mercury porosimetry is used to determine the macropore volume and macropore surface area by employing Thermo scientific Pascal 140 mercury porosimeter with a maximum test pressure of 400 kPa at 298 K with Hg density 13.59 g/cm3.

The morphology of SFAA before and after fluoride adsorption was investigated using Philips XL-30 electron microscope. As the samples were insulating, samples were gold coated to avoid charging. Fourier transform infrared spectroscopy (FTIR) spectra were recorded between 500 cm−1 and 4,000 cm−1with 16 cm−1 resolution using ABB-MB3000 FTIR spectrometer.

The X-ray photoelectron spectroscopy (XPS) was performed with XPS spectrometer (M/s. Specs, Germany) provided with a monochromatic AlKα radiation and a hemispherical analyzer. The base vacuum was observed to be better than 5.0 × 10−10 mbar during the analysis. The selected area scans were collected with 20 eV pass energy.

Batch sorption experiments

Dry sorbent of 0.1 g was added to 10 mL of synthetic fluoride solution in sample vials of 15 mL and the pH was adjusted to 6.5. These samples were kept in a thermostatic shaker for mixing at 30°C under 40 rpm. Samples removed at predetermined time intervals were centrifuged, filtered using 90 mm ϕ Whatman filter paper and the filtrate was analyzed for residual fluoride concentration using an ion selective electrode. Experiments were repeated to check for reproducibility. The amount of fluoride adsorbed at equilibrium, qe (mg/g), was computed by 
formula
where Ci is initial concentration (mg/L), Ce is equilibrium concentration (mg/L), V is volume of solution (L) and m is the mass of adsorbent (g).

RESULTS AND DISCUSSION

Optimization of sawdust, ferric hydroxide and activated alumina ratio

For optimization of the composition of the adsorbent, mass ratio of the constituents was varied and adsorption capacities were determined for fluoride uptake from water containing 20–100 mg/L fluoride. The compositions of 1:1:1, 1.5:1:1 and 2:1:1 of sawdust: ferric hydroxide: activated alumina have been investigated (Figure 1). For comparison, the adsorption capacities of sawdust and activated alumina were also shown in Figure 1. It can be seen that the composition 1.5:1:1 gives the highest capacity and the value is almost equal to that of activated pure alumina up to 40 mg/L more than activated alumina beyond this. It should be noted that the activated alumina is less than 30% by weight in SFAA. Therefore, the present adsorbent is more economical and ecofriendly than activated alumina. Hence, the composition of 1.5:1:1 is chosen and the experiments related to other operating parameters were carried out using this composition.
Figure 1

Effect of varying composition of SFAA.

Figure 1

Effect of varying composition of SFAA.

Adsorbent characterization studies

Surface area and pore size distribution by nitrogen adsorption

Surface characterization of the adsorbents has been carried out by finding out surface areas and pore volumes in different ranges, viz., micro, meso and macro. In the case of activated alumina and SFAA, the values were determined before and after fluoride adsorption. BET surface area and volume in meso- and micropores (Vmeso+micro) are determined by liquid nitrogen adsorption at cryogenic temperatures. Vmicro and Smeso are evaluated by t-plots. Vmeso and Smicro are calculated by subtracting Vmicro and Smeso from Vmeso+micro and SBET. The macropore volume and macropore surface area are determined by mercury porosimetry. Figure 2(a) shows pore size distribution by liquid nitrogen adsorption, Figure 2(b),t-plots and Figure 2(c) pore size distribution by mercury porosimetry. The results are summarized in Table 2. It can be seen that the sawdust has substantial macropore volume and micropore surface area. In case of activated alumina the surface area is mostly in the mesopore region and the macropore volume is moderate. As a result, the combination adsorbent, SFAA has a well-balanced pore size distribution with the pore volumes and surface areas distributed in macro-, meso- and micropore regions.
Table 2

Surface areas and pore volumes of the adsorbents

AdsorbentBy liquid nitrogen adsorption
From t-plots
Mercury porosimetry
SBET (m2/g)Vmeso+micro (cm3/g)Vmicro (cm3/g)Vmeso (cm3/g)Smeso (m2/g)Smicro (m2/g)Vmacro (cm3/g)Smacro (m2/g)
Sawdust 22.50 0.0133 0.0034 0.00986 7.42 15.08 2.185 0.373 
Activated alumina 118.43 0.172 0.0179 0.1540 113.22 5.21 0.425 0.043 
Activated alumina + F 102.20 0.143 0.0130 0.1300 97.27 4.93 0.416 0.051 
SFAA 61.10 0.0965 0.0313 0.0652 41.20 20.81 0.740 0.102 
SFAA + F 21.90 0.0391 0.0089 0.0301 20.29 1.61 0.737 0.141 
AdsorbentVTOTAL (cm3/g)% Vmicro% Vmeso% VmacroSTOTAL (m2/g)% Smicro% Smeso% Smacro
Sawdust 2.20 0.16 0.45 99.40 22.87 65.93 32.44 1.63 
Activated alumina 0.60 2.99 25.78 71.23 118.44 4.37 95.59 0.04 
Activated alumina + F 0.56 2.32 23.32 74.37 102.25 4.82 95.13 0.05 
SFAA 0.84 3.74 7.80 88.46 62.11 33.50 66.33 0.16 
SFAA + F 0.78 1.16 3.88 94.96 22.04 7.30 92.06 0.64 
AdsorbentBy liquid nitrogen adsorption
From t-plots
Mercury porosimetry
SBET (m2/g)Vmeso+micro (cm3/g)Vmicro (cm3/g)Vmeso (cm3/g)Smeso (m2/g)Smicro (m2/g)Vmacro (cm3/g)Smacro (m2/g)
Sawdust 22.50 0.0133 0.0034 0.00986 7.42 15.08 2.185 0.373 
Activated alumina 118.43 0.172 0.0179 0.1540 113.22 5.21 0.425 0.043 
Activated alumina + F 102.20 0.143 0.0130 0.1300 97.27 4.93 0.416 0.051 
SFAA 61.10 0.0965 0.0313 0.0652 41.20 20.81 0.740 0.102 
SFAA + F 21.90 0.0391 0.0089 0.0301 20.29 1.61 0.737 0.141 
AdsorbentVTOTAL (cm3/g)% Vmicro% Vmeso% VmacroSTOTAL (m2/g)% Smicro% Smeso% Smacro
Sawdust 2.20 0.16 0.45 99.40 22.87 65.93 32.44 1.63 
Activated alumina 0.60 2.99 25.78 71.23 118.44 4.37 95.59 0.04 
Activated alumina + F 0.56 2.32 23.32 74.37 102.25 4.82 95.13 0.05 
SFAA 0.84 3.74 7.80 88.46 62.11 33.50 66.33 0.16 
SFAA + F 0.78 1.16 3.88 94.96 22.04 7.30 92.06 0.64 

SBET – BET surface area (m2/g), Vmeso+micro – Meso- and micropore volume (cm3/g), Vmicro – Micropore volume (cm3/g), Vmeso – Mesopore volume (cm3/g), Smeso – Mesopore surface area (m2/g), Smicro – Micropore surface area (m2/g), Vmacro – Macropore volume (cm3/g), Smacro – Macropore surface area (m2/g).

VTOTAL = Vmicro + Vmeso + Vmacro.

STOTAL = Smicro + Smeso + Smacro.

Figure 2

(a) Pore size distribution by N2 adsorption. (b) t-plot based on N2 adsorption. (c) Pore size distribution by mercury porosimetry.

Figure 2

(a) Pore size distribution by N2 adsorption. (b) t-plot based on N2 adsorption. (c) Pore size distribution by mercury porosimetry.

The table shows surface areas and pore volumes in all the three ranges for activated alumina before and after adsorption of fluoride. For activated alumina, reduction in the volumes or the surface areas after fluoride adsorption is not considerable, implying that fluoride adsorption is not substantial. The distribution of surface areas and pore volumes for SFAA before and after fluoride adsorption shows that the micropore surface area reduced to 7.7% of the original, as against 95% for activated alumina. This indicates that the micropores have been filled very effectively in case of SFAA. Similarly, the Smeso reduced to 49% of the original for SFAA compared with 86% for activated alumina. Thus, even in the mesopore region, adsorption has been more effective by SFAA. There is no difference in macropore volumes before and after adsorption in the case of both activated alumina and SFAA, as the adsorption takes place only in meso- and micropores. The effectiveness of SFAA is reflected in the maximum adsorption capacities of activated alumina (8.60 mg/g) and SFAA (40.28 mg/g) for initial concentration of 1,000 mg/L fluoride as shown in Table 3. A good adsorbent should have a well-developed pore size distribution, as macropores essentially serve as transport pathways and micropores and mesopores contribute to the adsorption sites. Thus, although the BET surface area of activated alumina is more than that of SFAA, the utilization of the micro-/mesopores for adsorption is better in the case of SFAA because of the well-developed approach paths to the micropores in the case of SFAA. This is also reinforced by the column studies presented in ‘Column study’.

Table 3

Fluoride adsorption capacities of various adsorbents

AdsorbentCapacity (mg/g)Experimental conditions
Reference
Temp (°C)pHConc. (mg/L)
SFAA 40.28 30 6.5 20–1,000 Current study 
Activated alumina – current study 8.60 30 6.5 20–1,000 Current study 
Activated alumina (OA-25) 2.00 – 7.0 2.5–14 Ghorai & Pant (2004)  
Acidic alumina 8.40 – 3.6–11.6 5–15 Goswami & Purkait (2012)  
MnO2 coated alumina 2.85 31 7.10 2.5–30 Maliyekkal et al. (2006)  
La(III) impregnated alumina 6.65 Room temp 5.7–8.0 1–35 Puri & Balani (2000)  
Hydrous manganese oxide coated alumina 7.09 25 5.20 10–70 Teng et al. (2009)  
Nano alumina 14.0 25 6.15 1–100 Kumar et al. (2011)  
CaO modified activated alumina 101.01 25 5.5 1–1,000 Camacho et al. (2010)  
Acidic and alkaline alumina 3.0–20.4 30 3.0–12.0 10 Karthikeyan et al. (1997)  
Aluminium hydroxide (THA & UHA) 23.7 & 7.0 24 6.0–7.0 5–30 Shimelis et al. (2006)  
Granular ferric hydroxide 7.0 26 4.0 1–100 Kumar et al. (2009)  
Al(OH)3 coated rice husk 15 27 5.0 5.0–25.0 Ganvir & Das (2011)  
Cellulose @hydroxyapatite 10.0 25 4.0–5.0 10.0 Yu et al. (2013)  
AdsorbentCapacity (mg/g)Experimental conditions
Reference
Temp (°C)pHConc. (mg/L)
SFAA 40.28 30 6.5 20–1,000 Current study 
Activated alumina – current study 8.60 30 6.5 20–1,000 Current study 
Activated alumina (OA-25) 2.00 – 7.0 2.5–14 Ghorai & Pant (2004)  
Acidic alumina 8.40 – 3.6–11.6 5–15 Goswami & Purkait (2012)  
MnO2 coated alumina 2.85 31 7.10 2.5–30 Maliyekkal et al. (2006)  
La(III) impregnated alumina 6.65 Room temp 5.7–8.0 1–35 Puri & Balani (2000)  
Hydrous manganese oxide coated alumina 7.09 25 5.20 10–70 Teng et al. (2009)  
Nano alumina 14.0 25 6.15 1–100 Kumar et al. (2011)  
CaO modified activated alumina 101.01 25 5.5 1–1,000 Camacho et al. (2010)  
Acidic and alkaline alumina 3.0–20.4 30 3.0–12.0 10 Karthikeyan et al. (1997)  
Aluminium hydroxide (THA & UHA) 23.7 & 7.0 24 6.0–7.0 5–30 Shimelis et al. (2006)  
Granular ferric hydroxide 7.0 26 4.0 1–100 Kumar et al. (2009)  
Al(OH)3 coated rice husk 15 27 5.0 5.0–25.0 Ganvir & Das (2011)  
Cellulose @hydroxyapatite 10.0 25 4.0–5.0 10.0 Yu et al. (2013)  

Besides a higher efficiency, this adsorbent is economic as the sawdust constitutes 40% by weight of the total adsorbent, minimizing the concentration of other metallic hydroxides. This establishes the pivotal role played by Artocarpus hirsutus sawdust in preparation of this efficient and cost effective adsorbent.

Scanning electron microscopy

The scanning electron microscopy (SEM) images of SFAA before and after adsorption of fluoride are recorded under the magnification of 2000x and the images are shown in Figures 3(a) and 3(b). The SEM images show that the granules remain intact after grafting. EDX analysis was performed to find the elemental constituents of pure SFAA, SFAA after fluoride adsorption (Figures 3(c) and 3(d)). The atomic% of Al and Fe on the basis of chemicals presented in pure SFAA was found to be 35.9 and 37.1%, respectively.
Figure 3

SEM image of SFAA before (a) and after (b) fluoride adsorption. (c) EDX spectra of SFAA. (d) EDX spectra of SFAA after fluoride adsorption.

Figure 3

SEM image of SFAA before (a) and after (b) fluoride adsorption. (c) EDX spectra of SFAA. (d) EDX spectra of SFAA after fluoride adsorption.

FTIR

Figure 4 shows the FTIR spectra of SFAA before and after fluoride adsorption. The broad and intense peak at 3,060–3,600 cm−1 is characteristic stretching frequency of hydroxyl groups and the peak which is observed at 1,640 cm−1 is assigned to the bending vibration of adsorbed water. The small peaks at 1,504 and 1,427 cm−1 are assigned to C = C aromatic stretching. FTIR results indicate that mainly hydroxide ions in different functional groups are involved in fluoride ion sorption. After the fluoride adsorption, increases in intensities of the peaks are observed which might possibly be due to the exchange of OH by F (Wu et al. 2010). These results are in conformity with the earlier studies by various researchers. As sawdust is the major component of SFAA, its functional groups (amino, carboxyl, thiol, sulfhydryl, alcohol, phenol, phosphate, polysaccharides and proteins) are reflected in the FTIR (Tuzun et al. 2005).
Figure 4

FTIR result of SFAA before and after fluoride adsorption.

Figure 4

FTIR result of SFAA before and after fluoride adsorption.

XPS

The XPS survey confirmed the distribution of surface elemental composition loaded onto the Artocarpus hirsutus sawdust (Teng et al. 2009). Figure 5 shows the survey of wide scan XPS spectrum, which confirms the presence of Al, Fe, C, and O on the surface of SFAA. Selected area scans were performed for C 1 s, O 1 s, Fe 2p and F 1 s levels (Figure 6).
Figure 5

Full range XPS spectra of SFAA before and after fluoride adsorption.

Figure 5

Full range XPS spectra of SFAA before and after fluoride adsorption.

Figure 6

XPS spectra of (a) C 1 s, (b) Fe 2p, (c) Al 2 s, and (d) F 1 s.

Figure 6

XPS spectra of (a) C 1 s, (b) Fe 2p, (c) Al 2 s, and (d) F 1 s.

Iron and aluminum oxide are present on the SFAA surface which was confirmed by the corresponding peaks. Fe 2p1 and 2p3 peaks have been obtained at 710.8 and 724.8 eV, respectively, similarly Al 2 s peak at has been obtained at 117.9 eV. Samples of SFAA exposed to NaF showed F peak was observed near 685 eV of binding energy (Figure 6(d)), and indicating SFAA adsorbed fluoride on its surface. To confirm the presence of fluoride, the selected area was scanned for F 1 s region and the patterns were recorded for the F loaded specimens (Figure 6(d)). The pattern is fitted with a single Gaussian peak giving rise to a correlation coefficient of 0.98. The position of the peak indicates that fluoride ion is −1 valence state. Thus, XPS studies reveal that fluoride ions are trapped by the matrix.

Effect of pH

The initial pH of solution containing fluoride concentration of 20 mg/L was varied from 1.0 through 12.0 by adding either dilute HNO3 or NaOH drop-wise to achieve the desired pH. Fluoride uptake by SFAA was estimated at different pH values and the results are shown in Figure 7.
Figure 7

Effect of initial pH.

Figure 7

Effect of initial pH.

The adsorption was nearly constant over a wide pH range of 1 to 9, giving percentage removal of about 72% which corresponds to a fluoride uptake of about 2 mg/g. This remarkable feature of the adsorbent over wide pH range which extends even up to the weak alkaline medium makes it attractive for defluoridation of water. However, the adsorption sharply decreases beyond pH 10 implying that the presence of large excess of OH ions in the fluoride solution alters the kinetics of fluoride adsorption. The role of surface groups and their protonation/deprotonation in fluoride adsorption was reported for activated alumina (Leyva-Ramos et al. 2008) and granular ferric hydroxide (Kumar et al. 2009). Functional groups (amino, carboxyl, thiol, sulfhydryl, alcohol, phenol and phosphate groups) present on the surface of SFAA, as reported above in ‘FTIR’ also play a similar role. As the ionic form of fluoride in solution and the electric charge on the sorbent depend on pH, fluoride adsorption is influenced by pH. This dependence on pH is better understood by the determination of point of zero charge (PZC), which is the value of pH that gives zero net surface charge and is an index of the propensity of the system to become either positively or negatively charged. PZC for SFAA was determined by mass titration method to be 5.85, as explained in our previous paper (Dhanasekaran et al. 2016). In the range of pH 1 to 5.85, the surface is protonated which renders the surface positively charged and this positively charged surface attracts the negatively charged fluoride ions from the solution. For the intermediate range, pH range of 5.85 to 9.0, the adsorption continues to be nearly constant as the hydroxyl ion concentration is relatively low and the hydroxyl ion progressively replaced by fluoride ions from the solution. However, for the pH beyond, a dramatic decrease in fluoride adsorption is observed. This is due to high hydroxyl ion concentration and stronger competition of hydroxyl ions to the active sites of SFAA compared with fluoride.

Effect of initial concentration

Although fluoride concentration prevalent in groundwater is limited to less than 50 mg/L, industrial effluents contain fluoride levels. Therefore, fluoride solutions of various concentrations from 20 to 1,000 mg/L were used to find the adsorbing capacity of SFAA, so as to assess the suitability of SFAA for industrial wastewater treatment (Figure 8). The maximum adsorption capacity of SFAA for fluoride at room temperature (30 °C) is 40.28 mg/g, which was achieved for an initial fluoride concentration of 954 mg/L. This is considerably higher than the capacity of 8.60 mg/g for activated alumina used in the present study and capacities reported in the literature for various adsorbents. Fluoride adsorption capacities of various adsorbents reported in the literature are compiled in Table 3.
Figure 8

Effect of initial fluoride concentration.

Figure 8

Effect of initial fluoride concentration.

Figure 8 shows the effect of initial concentration of fluoride solution. It can be seen that the adsorbent is very effective at low concentrations, for the given adsorbent dose and contact time. At an initial concentration of 22 mg/L fluoride, removal of more than 99% is achieved. Thus, this adsorbent is highly suitable for treating groundwater polluted with fluoride, to bring down the fluoride levels to the levels prescribed by WHO. In the range of 200 to 1,000 mg/L, removal of 40–50% was achieved with the adsorbent dose used. This establishes that SFAA can be used for industrial wastewater treatment by designing a suitable process scheme. It can be seen from Table 3 that SFAA showed higher fluoride adsorption capacity compared with activated alumina used in the study. Although activated alumina is reported to exhibit a reasonably high fluoride ions adsorption capacity, the capacity is reported to be enhanced by the addition of ferric hydroxide (Sujana et al. 2009). Anjili sawdust serves as a base material, offering several surface functional groups (hydroxide, amino, carboxyl, thiol, sulfhydryl, alcohol, phenol) which have the potential to participate in the fluoride ion exchange reaction. In addition, it also provides a high specific surface area for metal hydroxides, which upon impregnation raise the total functional groups of the adsorbent that take part in fluoride ion exchange. As already explained in ‘Adsorbent characterization studies’, the sawdust contributes significantly in the macropore region and hence serves as a good host material.

Effect of temperature

Although practical application of this adsorbent is defluoridation of water at room temperature, effect of temperature gives an insight into the mechanism of adsorption of fluoride. These sorption experiments were carried out at various temperatures, viz., 303, 308, 313, 323 and 333 K. Figure 9 shows the plot of uptake of fluoride (qe) as a function of concentration of fluoride in solution at equilibrium (Ce). When the temperature was increased from 303 K to 308 K, 313 K, 323 K and subsequently to 333 K, the adsorption increased, indicating that the mechanism is chemisorption.
Figure 9

Effect of temperature.

Figure 9

Effect of temperature.

Adsorption isotherms

Freundlich and Langmuir isotherms are applied for analysis of the adsorption data. Freundlich model which describes non-ideal heterogeneous adsorption is expressed by 
formula
where qe is equilibrium uptake of fluoride per unit mass of adsorbent (mg/g), KF is Freundlich constant, Ce is equilibrium concentration (mg/L) and n is a constant.
The Langmuir isotherm, on the other hand, assumes uniform adsorption on the surface with a monolayer formation and can be represented by the following equation, 
formula
where qe is equilibrium uptake of fluoride per unit mass adsorbent (mg/g), qm is the maximum value of q, KL is Langmuir constant and Ce is equilibrium concentration (mg/L). The isotherms are presented in Figures 10(a) and 10(b) and the isotherm parameters in Table 4. Based on the R2 values, it can be seen that Freundlich model predicted the data well.
Table 4

Isotherm model constants for adsorption of F onto SFAA

Conc. (mg/L)qexp (mg/g)Freundlich isotherm
Langmuir isotherm
Kf (mg/g)(L/mg)1/nnR2KL (L/g)qm (mg/g)R2
20–1,000 40.28 1.96 2.41 0.931 0.003 52.36 0.639 
Conc. (mg/L)qexp (mg/g)Freundlich isotherm
Langmuir isotherm
Kf (mg/g)(L/mg)1/nnR2KL (L/g)qm (mg/g)R2
20–1,000 40.28 1.96 2.41 0.931 0.003 52.36 0.639 

qexp, qm – equilibrium uptake by experiment, Langmuir isotherm.

Figure 10

(a) Freundlich isotherm, (b) Langmuir isotherm. Time = 4 h, Ci = 20–1,000 mg/L, pH = 6.5, Temp = 303 K, speed = 40 rpm, V = 10 mL, mass = 0.1 g.

Figure 10

(a) Freundlich isotherm, (b) Langmuir isotherm. Time = 4 h, Ci = 20–1,000 mg/L, pH = 6.5, Temp = 303 K, speed = 40 rpm, V = 10 mL, mass = 0.1 g.

Thermodynamic study

KF derived from Freundlich isotherm is used to evaluate the thermodynamic parameters ΔG °, ΔS ° and ΔH °. The thermodynamic equilibrium constant Kd is obtained by multiplying KF (in the units of L/g) with water density, 1,000 g/L (Milonjic 2007; Kumar et al. 2009). Van 't Hoff plot of lnKd with reciprocal of temperature (1/T) is shown in Figure 11. The value of ΔH °, and ΔS ° were obtained from the slope and intercept of the plot. The thermodynamic parameters were determined using the following equations (Daifullah et al. 2007): 
formula
 
formula
where ΔG ° is the standard free energy, R universal gas constant (8.314 J/mol K) and T is the absolute temperature in Kelvin (K).
Figure 11

Van 't Hoff plot.

Figure 11

Van 't Hoff plot.

From Figure 11, ΔH ° and ΔS ° were calculated to be 17.86 KJ/mol and 0.1 KJ/mol.K, respectively. The endothermic nature of adsorption is indicated by the positive value of ΔH ° for the chemisorption process whereas the negative value of ΔG ° confirms that the reaction is spontaneous at all the temperatures. The adsorption mechanism of defluoridation is mainly by F ̄ and OH̄ exchange. Desorption of OH̄ attached to active sites on the surface needs some activation energy to leave the surface, which might explain the endothermic nature of the adsorption process.

Kinetic study

Kinetic performance of an adsorbent is of great importance for plant application. From kinetic analysis, the solution uptake rate, the residence time required for completion of adsorption reaction and the scale of adsorption apparatus required can be determined. The kinetics of fluoride adsorption on SFAA was analyzed using pseudo first-order, pseudo second-order and intraparticle diffusion kinetic models and the model parameters are presented in Table 5.

Table 5

Kinetic model parameters

Experimental
Pseudo first-order
Pseudo second-order
Intraparticle diffusion
Ciqexpq1K1R2q2K2R2KidR2
0.324 0.384 0.039 0.998 0.329 1.298 0.999 0.009 0.871 
10 0.514 0.474 0.038 0.996 0.521 0.514 0.999 0.015 0.864 
15 0.588 0.464 0.029 0.982 0.593 0.398 0.999 0.015 0.814 
20 0.677 0.448 0.026 0.984 0.682 0.301 0.999 0.016 0.822 
Experimental
Pseudo first-order
Pseudo second-order
Intraparticle diffusion
Ciqexpq1K1R2q2K2R2KidR2
0.324 0.384 0.039 0.998 0.329 1.298 0.999 0.009 0.871 
10 0.514 0.474 0.038 0.996 0.521 0.514 0.999 0.015 0.864 
15 0.588 0.464 0.029 0.982 0.593 0.398 0.999 0.015 0.814 
20 0.677 0.448 0.026 0.984 0.682 0.301 0.999 0.016 0.822 

Temp = 303 K, pH = 6.5, Speed = 40 rpm, Contact time = 4 h, Volume = 10 mL, Mass = 0.1 g.

Pseudo first-order Lagergren model

The most widely used Lagergren rate equation for sorption of a solute from a liquid solution (Lagergren 1898) is a pseudo first-order equation and is presented in terms of adsorption capacity as follows: 
formula
The equation can be rewritten as 
formula
where qe and qt are the adsorption capacities (mg/g) at equilibrium and time t, respectively; and K1(min–1) is the rate constant of the pseudo first-order adsorption.
The plot of log(qe–qt) versus t as shown in Figure 12(a) should give a linear relation, from which K1 and qe can be determined from the slope and intercept, respectively. It can be seen that the Lagergren equation does not fit well for the whole range, particularly the initial times.
Figure 12

Pseudo (a) first-order, and (b) second-order sorption kinetics of F̄ on SFAA.

Figure 12

Pseudo (a) first-order, and (b) second-order sorption kinetics of F̄ on SFAA.

Pseudo second-order Ho model

Ho proposed this model based on the assumptions that the adsorption is second order with respect to solid concentration and the adsorption follows Langmuir equation (Ho & McKay 2000). The typical second-order rate equation in solution systems is as follows: 
formula
The linear form of pseudo second-order kinetic rate equation may be expressed as 
formula
where qe and qt are the adsorption capacity (mg/g) at equilibrium and time t; K2 is pseudo second-order rate constant (g/mg.min–1). From the slope and intercept of the straight line obtained from the plot of t/qt versus t, the values of qe and K2 are calculated and presented in Table 5. It can be seen from R2 values (Table 5) that Ho's model gives a much better fit of the experimental data than Lagergren equation, for the entire range of concentrations studied. This shows that fluoride adsorption by SFAA follows pseudo second-order kinetics and is chemical adsorption. Equilibrium adsorption capacities (qe) predicted by this model are close to the experimental values.

Intraparticle diffusion model (pore diffusion)

The intraparticle diffusion model (Weber & Morris 1963) is applied to describe competitive adsorption. The sorption rate is controlled by several factors including diffusion of the solute: (a) from the solution to the film surrounding the particle (bulk diffusion); (b) from the film to the particle surface (external or film diffusion); and (c) from the surface to the internal sites (surface or pore diffusion). This intraparticle diffusion occurs in macro-, meso- and micropores present within the adsorbent.

The initial rate of intraparticle diffusion is obtained by linearization of the curve qt = f (t0.5). The plot of qt against t0.5 may present multi-linearity (Panday et al. 1985; Allen et al. 1989). 
formula
where qt is adsorption capacity (mg/g) at time t; Kid is the intra-particle (pore) diffusion rate constant (mg/g min−0.5) and C is the intercept that gives an idea about the thickness of the boundary layer. Larger C value implies a greater boundary layer effect.
The plot of qt against t0.5 indicates that three steps occur in the adsorption processes as shown in Figure 13. The first linear portion (0–4 min) is external surface adsorption (film diffusion). This represents binding of fluoride ions by the active sites, distributed on the outer surface of SFAA. The second linear portion (4–30 min) represents intraparticle diffusion (pore diffusion). The third linear portion (30–160 min) indicates the final equilibrium stage, where the intra-particle diffusion starts to slow down, due to the extremely low solute concentration in solution. The linear portion of the plot for contact time of >4 min does not pass through the origin and the intercept is significant. This implies that both external mass transfer and intraparticle diffusion contribute to the overall adsorption. The intraparticle diffusion model parameters are given in Table 5.
Figure 13

Intraparticle diffusion model of fluoride adsorption on SFAA.

Figure 13

Intraparticle diffusion model of fluoride adsorption on SFAA.

Interference of other ions

The effect of anions such as chloride, sulfate, nitrate, bicarbonate and cations such as calcium, magnesium, and potassium, on the adsorption of fluoride by SFAA was examined experimentally and the results are given in Figures 14(a) and 14(b). Hydroxide, carbonate and bicarbonate interfere with fluoride adsorption by SFAA in decreasing order (Figure 14(a)). Chloride, sulfate and nitrate did not show significant interference. Hydroxide ion, as already explained in the FTIR section, is a competitor of fluoride for ion exchange on the surface of the adsorbent. Moreover, increase of OH increases pH, which results in decrease in fluoride adsorption. Similarly, increase in carbonates and bicarbonates also increased the pH beyond 9, as seen from the experimental data, and hence there is a reduction in fluoride removal. Kumar et al. 2011 also reported similar findings that presence of chloride and nitrate ions had negligible effect on fluoride removal by activated alumina, as these ions are outer-spherically sorbing anions.
Figure 14

Interference of (a) anions and (b) cations.

Figure 14

Interference of (a) anions and (b) cations.

Similarly interference of cations was studied and reported in Figure 14(b). It is seen that increasing sodium or potassium reduces fluoride adsorption, possibly due to the high solubility of sodium or potassium fluoride. Calcium enhances (77 to 87%) the fluoride adsorption by forming calcium fluoride, which is insoluble in water.

Column study

For continuous application on industrial scale, the adsorbent should be amenable for column operation. Sawdust (S.D), activated alumina (A.A), and SFAA were compared for their fluoride adsorption in column operation. Water containing fluoride concentration 20 mg/L was passed through 1 cm inner diameter (di) column with 2 g of adsorbent at a constant flow rate of 2 mL/min. The breakthrough curves for different adsorbents are shown in Figure 15. The breakthrough adsorption capacity of SFAA was found to be 1.12 mg/g as compared to 0.41 mg/g for activated alumina.
Figure 15

Column studies of different adsorbents for fluoride adsorption.

Figure 15

Column studies of different adsorbents for fluoride adsorption.

Fluoride removal from groundwater

SFAA was assessed for fluoride removal from groundwater from Avarangattur, Tamil Nadu and the results are given in Table 6. It can be seen that SFAA was effective despite the presence of other competing ions. SFAA reduced fluoride concentration from 8.2 to 1.12 mg/L.

Table 6

Groundwater sample parameters

Water parameterRaw water
Treated water
ABCABC
pH 8.41 6.05 5.89 6.91 7.01 7.01 
Temp (K) 303 303 303 303 303 303 
Anions 
 F 8.2 4.4 5.2 1.12 0.55 0.42 
 Cl 119.2 156.5 37.9 75.8 13.9 2.98 
 NO3 95.9 84.6 11.8 15.7 7.61 0.0 
 SO42– 73.9 49.9 1.82 0.0 0.0 0.0 
Water parameterRaw water
Treated water
ABCABC
pH 8.41 6.05 5.89 6.91 7.01 7.01 
Temp (K) 303 303 303 303 303 303 
Anions 
 F 8.2 4.4 5.2 1.12 0.55 0.42 
 Cl 119.2 156.5 37.9 75.8 13.9 2.98 
 NO3 95.9 84.6 11.8 15.7 7.61 0.0 
 SO42– 73.9 49.9 1.82 0.0 0.0 0.0 

Temp = 303 K, Speed = 40 rpm, Time = 4 h, Volume = 10 mL, Mass = 0.1 g.

A = Avarangattur (Tamil Nadu), B & C = Hot spring, Chandana village [Godda district, Jharkhand].

CONCLUSION

The performance of a novel adsorbent, SFAA, which was already reported to be effective for arsenic removal, is demonstrated for defluoridation of water to meet the drinking water standards laid down by WHO. Compared with the fluoride adsorption capacity of 8.6 mg/g for activated alumina, capacity for SFAA is 40.28 mg/g for batch adsorption from initial fluoride concentration of 1,000 mg/L. Superiority of SFAA over activated alumina is explained by detailed surface characterization. SFAA has an optimum pore size distribution, the sawdust contributing to the maximum in the macropore range. Adsorption capacity is unchanged over a wide range of pH, 1 to 9, making it useful for defluoridation of groundwater as well as industrial effluents. Fluoride adsorption by SFAA follows Freundlich adsorption isotherm and follows pseudo second-order kinetic model. Hydroxides, carbonates, bicarbonates, sodium and potassium decrease the adsorption. Breakthrough adsorption capacity of SFAA is 1.21 mg/g, which is four times higher than that of activated alumina employed in the study.

ACKNOWLEDGEMENTS

The authors are grateful to Dr V. Jayaraman, Mr A. Sree Rama Murthy, MCD, Chemistry Group, IGCAR, Dr P. Chandra Mohan, WSCD, BARCF for their help in BET studies. The authors are thankful to Mrs Clinsha, UGC-DAE for the help in XPS studies.

REFERENCES

REFERENCES
Alagumuthu
G.
Veeraputhiran
V.
Venkataraman
R.
2011
Fluoride sorption using Cynodon dactylon based activated carbon
.
Hemijska industrija
65
(
1
),
23
35
.
Allen
S.
McKay
G.
Khader
K.
1989
Intraparticle diffusion of a basic dye during adsorption onto sphagnum peat
.
Environmental Pollution
56
(
1
),
39
50
.
Bhatnagar
A.
Kumar
E.
Sillanpaa
M.
2011
Fluoride removal from water by adsorption-a review
.
Chemical Engineering Journal
171
(
3
),
811
840
.
Camacho
L. M.
Torres
A.
Saha
D.
Deng
S.
2010
Adsorption equilibrium and kinetics of fluoride on sol–gel-derived activated alumina adsorbents
.
Journal of Colloid and Interface Science
349
(
1
),
307
313
.
Chen
N.
Zhang
Z.
Feng
C.
Sugiura
N.
Li
M.
Chen
R.
2010
Fluoride removal from water by granular ceramic adsorption
.
Journal of Colloid and Interface Science
348
(
2
),
579
584
.
Das
N.
Vimala
R.
Karthika
P.
2008
Biosorption of heavy metals-an overview
.
Indian Journal of Chemical Technology
7
,
159
169
.
Dhanasekaran
P.
Sai
P. S.
Babu
C. A.
Prabhu
R. K.
Rajan
K.
2016
Arsenic removal from groundwater by Anjili tree sawdust impregnated with ferric hydroxide and activated alumina
.
Water Science and Technology: Water Supply
16
(
1
),
115
127
.
Ganvir
V.
Das
K.
2011
Removal of fluoride from drinking water using aluminum hydroxide coated rice husk ash
.
Journal of Hazardous Materials
185
(
2–3
),
1287
1294
.
Gogoi
P. K.
Baruah
R.
2008
Fluoride removal from water by adsorption on acid activated kaolinite clay
.
Indian Journal of Chemical Technology
15
,
500
503
.
Goswami
A.
Purkait
M. K.
2012
The defluoridation of water by acidic alumina
.
Chemical Engineering Research and Design
90
(
12
),
2316
2324
.
Karthikeyan
G.
Apparao
B. V.
Meenakshi
S.
1997
Defluoridation properties of Activated Alumina
. In:
2nd International Workshop on Fluorosis Prevention and Defluoridation of Water
.
The International Society for Fluoride Research
, pp.
78
82
.
Kumar
E.
Bhatnagar
A.
Ji
M.
Jung
W.
Lee
S. H.
Kim
S. J.
Lee
G.
Song
H.
Choi
J. Y.
Yang
J. S.
2009
Defluoridation from aqueous solutions by granular ferric hydroxide (GFH)
.
Water Research
43
(
2
),
490
498
.
Kumar
E.
Bhatnagar
A.
Kumar
U.
Sillanpaa
M.
2011
Defluoridation from aqueous solutions by nano-alumina: characterization and sorption studies
.
Journal of Hazardous Materials
186
(
2
),
1042
1049
.
Lagergren
S.
1898
About the theory of so-called adsorption of soluble substances
.
Kungliga Svenska Vetenskapsakademiens Handlingar
24
(
4
),
1
39
.
Leyva-Ramos
R.
Medellin-Castillo
N. A.
Jacobo-Azuara
A.
Mendoza-Barron
J.
Landin-Rodriguez
L. E.
Martinez-Rosales
J. M.
Aragon-Pina
A.
2008
Fluoride removal from water solution by adsorption on activated alumina prepared from pseudo-boehmite
.
Journal of Environmental Engineering Management
18
(
5
),
301
309
.
Lippens
B.
De Boer
J.
1965
Studies on pore systems in catalysts: V. The t method
.
Journal of Catalysis
4
(
3
),
319
323
.
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
.
Meenakshi
A.
Maheshwari
R. C.
2006
Fluoride in drinking water and its removal
.
Journal of Hazardous Materials
137
(
1
),
456
463
.
Meenakshi
A.
Maheshwari
R. C.
2007
Investigations on the batch performance of fluoride adsorption by coconut shell carbon
.
Asian Journal of Water, Environment and Pollution
4
(
2
),
119
122
.
Milonjic
S.
2007
A consideration of the correct calculation of thermodynamic parameters of adsorption
.
Journal of the Serbian Chemical Society
72
(
12
),
1363
1367
.
Mondal
N.
Bhaumik
R.
Roy
P.
Das
B.
Datta
J.
2013
Investigation on fixed bed column performance of fluoride adsorption by sugarcane charcoal
.
Journal of Environmental Biology
34
,
1059
1064
.
Panday
K.
Prasad
G.
Singh
V.
1985
Copper (II) removal from aqueous solutions by Fly ash
.
Water Research
19
(
7
),
869
873
.
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
.
Sathish
R. S.
Raju
N. S. R.
Raju
G. S.
Nageswara Rao
G.
Kumar
K. A.
Janadhana
C.
2007
Equilibrium and kinetic studies for fluoride adsorption from water on Zirconium impregnated coconut shell carbon
.
Separation Science and Technology
42
(
4
),
769
788
.
Sathish
R. S.
Sairam
S.
Raja
V. G.
Rao
G. N.
Janardhana
C.
2008
Defluoridation of water using zirconium impregnated coconut fiber carbon
.
Separation Science and Technology
43
(
14
),
3676
3694
.
Shimelis
B.
Zewge
F.
Chandravanshi
B. S.
2006
Removal of excess fluoride from water by aluminum hydroxide
.
Bulletin of the Chemical Society of Ethiopia
20
(
1
),
17
34
.
Siminiceanu
I.
Lazau
I.
Ecsedi
Z.
Lupa
L.
Burciag
C.
2008
Textural characterization of a new iron-based ammonia synthesis catalyst
.
Chemical Buletin
53
,
1
2
.
Sujana
M. G.
Soma
G.
Vasumathi
N.
Anand
S.
2009
Studies on fluoride adsorption capacities of amorphous Fe/Al mixed hydroxides from aqueous solutions
.
Journal of Fluorine Chemistry
130
(
8
),
749
754
.
Teng
S. X.
Wang
S. G.
Gong
W. X.
Liu
X. W.
Gao
B. Y.
2009
Removal of fluoride by hydrous manganese oxide-coated alumina: performance and mechanism
.
Journal of Hazardous Materials
168
(
2–3
),
1004
1011
.
Tor
A.
Danaoglu
N.
Arslan
G.
Cengeloglu
Y.
2009
Removal of fluoride from water by using granular red mud: batch and column studies
.
Journal of Hazardous Materials
164
(
1
),
271
278
.
Tuzun
I.
Bayramoglu
G.
Yalcin
E.
Basaran
G.
Celik
G.
Arica
M. Y.
2005
Equilibrium and kinetic studies on biosorption of Hg(II), Cd(II) and Pb(II) ions onto microalgae Chlamydomonas reinhardtii
.
Journal of Environmental Management
77
(
2
),
85
92
.
Viswanathan
N.
Meenakshi
S.
2008
Selective sorption of fluoride using Fe(III) loaded carboxylated chitosan beads
.
Journal of Fluorine Chemistry
129
(
6
),
503
509
.
Viswanathan
N.
Sundaram
C. S.
Meenakshi
S.
2009
Removal of fluoride from aqueous solution using protonated chitosan beads
.
Journal of Hazardous Materials
161
(
1
),
423
430
.
Weber
W.
Morris
J.
1963
Kinetics of adsorption on carbon from solution
.
Journal of the Sanitary Engineering Division American Society of Civil Engineers
89
(
17
),
31
60
.
WHO
2011
Guidelines for Drinking-Water Quality
.
World Health Organization
,
Geneva
, pp.
375
377
.
Wu
H.
Chen
L.
Gao
G.
Zhang
Y.
Wang
T.
Guo
S.
2010
Treatment effect on the adsorption capacity of alumina for removal fluoride
.
Nano Biomedicine and Engineering
2
(
4
),
231
235
.