3.1. Selecting the Best Carbonizing Temperature for the Preparation of CBC
When preparing CBC, selecting the appropriate carbonizing temperature is the most important factor as it affects the efficacy of CBC in removing fluoride from contaminated water.
Figure 2 shows the Freundlich isotherm for CBC carbonized at 673 K, 873 K, and 1073 K. In the figure, C is fluoride concentration in the solution expressed in a unit of mg/L, while Q is maximum adsorption capacity expressed in a unit of mg/g.
The F− adsorption by CBC carbonized at three temperatures matched well with the Freundlich isotherm. According to the Freundlich isotherm, CBC carbonized at 673 K, 873 K, and 1073 K showed fluoride adsorption capacities of 3.62 mg/g, 5.35 mg/g, and 1.74 mg/g, respectively, at the equilibrium fluoride concentration of 10 mg/L. It was obvious that chicken bones carbonized at 873 K showed the highest adsorption capacity. Further, we observed a yellowish color, undesirable taste, and unpleasant odor in water treated with chicken bones carbonized at 673 K. Therefore, we selected 873 K as the best carbonizing temperature for further experiments.
The results obtained from our study were consistent with the literature. When the carbonizing temperature decreases to less than 773 K, odor, undesirable taste, and yellowish color occur in the treated water due to less (or no) removal of organic matter such as fats, oil, and meat residuals [
3] from chicken bones. Leyva-Ramos and his fellows had reported that carbonizing temperatures below 773 K caused an unpleasant taste and smell and yellowish color in treated water due to the organic matter in bone char [
21], similar to what we found in our study. We detected that the weight-basis C content in chicken bones carbonized at 673 K (14%) was higher than the C content in CBC carbonized at 873 K (9%) and 1073 K (6%), showing that there was considerably more organic matter in CBC carbonized at 673 K. At lower carbonizing temperatures (573 K), the organic matter in bones was insufficiently removed, which made bones unable to provide a large specific surface area and enough pore space for the efficient removal of fluoride [
22].
When the carbonizing temperature increased to more than 873 K, the adsorption capacity of CBC was reduced dramatically. Carbonizing temperatures higher than 873 K can alter the hydroxyapatite structure of bone [
22], leading to a reduction in fluoride removal capacity [
23], as we found in our study. Kawasaki et al. investigated the fact that the adsorption capacity of fluoride onto bone char carbonized at 1073 K for 2 h was higher than that carbonized for 2 h at 1273 K by studying four types of bone char: cow bone char, pig bone char, chicken bone char, and fish bone char [
24]. Mayorga et al. reported that the best fluoride removal performance of bone char was observed at the carbonizing temperature of 973 K for 2 h by studying cow bone char. Furthermore, they mentioned that when the carbonizing temperature increased to more than 1073 K, the fluoride ion adsorption capacity of bone char decreased from 6.0 mg/g to 3.0 mg/g, and when increasing the carbonizing temperature to 1273 K, the fluoride adsorption capacity was reduced to 1.0 mg/g. Finally, using their results, they justified that the carbonizing temperature of bone char plays a major role in water defluoridation [
6]. We selected 873 K as the best carbonizing temperature considering the adsorption capacity and the energy consumption.
3.3. Fluoride Removal Mechanism of CBC
Figure 4 shows change in the fluoride concentration of the stock solution throughout the operation period.
Table 1 gives a detailed description of the amount of F
− in the solution throughout the operation period. A certain amount of F
− was lost from the stock solution due to leakage, and this was considered when calculating the adsorption capacity. The CBC particles with a diameter of 106–212 µm showed an adsorption capacity of 11.2 mg/g at a fluoride concentration of 10 mg/L after the operation period of 148 days. The adsorption capacity was calculated according to the data in
Table 1 (1855 mg F
−/(55*3) g CBC).
It is obvious that the CBC showed an unusually high adsorption capacity, nearly 2 times the adsorption capacity obtained by the Langmuir isotherm, indicating that equilibrium was not established within 24 h. To confirm the unusual fluoride adsorption capacity, the fluoride content in the CBC before and after 148 days of operation was measured using steam distillation.
Table 2 shows the averaged value of the adsorption capacities of F1, F2, and F3 after the adsorption. An adsorption capacity of 11.1 mg/g coincided well with the adsorption capacity obtained from the mass balance calculation for the solution, as shown in
Table 1.
According to studies relating to the fluoride adsorption capacity of bone char, CBC with a particle size >0.075 mm, 0.075–0.30 mm, 0.30–1.18 mm, and 1.18–2.34 mm showed a fluoride adsorption capacity of 0.665 mg/g, 0.661 mg/g, 0.660 mg/g, and 0.643 mg/g, respectively, at the equilibrium fluoride concentration of 10 mg/L [
25]. They also reported the fluoride adsorption capacity of lamb bone char with particle sizes >0.075 mm, 0.075–0.30 mm, 0.30–1.18 mm, and 1.18–2.34 mm for which the fluoride adsorption capacity was 0.482 mg/g, 0.475 mg/g, 0.459 mg/g, and 0.414 mg/g, respectively, at the equilibrium fluoride concentration of 10 mg/L [
25].
In a study using 0.79 mm cattle bone char particles, a fluoride adsorption capacity of 2.71 mg/g was recorded at the equilibrium fluoride concentration of 1 mg/L [
7]. Rojas-Mayorga and his fellow researchers showed a fluoride adsorption capacity of 7.32 mg/g using ~1 mm cow bones at the equilibrium fluoride concentration of 60 mg/L [
6].
No studies have reported such an unusually high fluoride adsorption capacity of bone char.
Figure 5 shows the X-ray diffraction patterns of CBC before and after the fluoride adsorption.
According to
Figure 5, it is clear that the two X-ray diffraction patterns of CBC before and after the fluoride adsorption overlapped, showing that the crystal structure of CBC before and after the fluoride adsorption was similar.
It has been reported that fluoride removal by bone char is a surface reaction process [
22].
Table 3 shows the BET surface area of CBC before and after the fluoride adsorption.
According to
Table 3, the finer particle size of CBC 106–212 µm before and after the fluoride adsorption showed a similar surface area.
SEM images of 106–212 µm CBC used for the study were taken in two different stages to compare the surface, morphology, and size distribution.
Figure 6 and
Figure 7 show SEM images of the CBC before and after fluoride adsorption, respectively.
The SEM images in
Figure 6 and
Figure 7 show similar structures of CBC, as evident from the similar X-ray diffraction patterns of CBC before and after the fluoride adsorption in
Figure 5. This was further confirmed by the almost-equal surface area of CBC before and after the fluoride adsorption, as shown in
Table 3.
It has been reported in the literature that fluoride removal by bone char (CBC) is associated with the two main mechanisms of ion exchange and chemical precipitation. In the presence of fluoride ion, the hydroxyl ion in HAP is replaced by fluoride ion to form insoluble fluorapatite (FAP) [
25] and release the hydroxyl ion into the solution. F
− and OH
− consist of the same charge and a similar size of radius. Therefore, the fluoride ion can replace the hydroxyl ion in mineral structures [
22].
The relevant chemical reaction can be represented as Equation (1) [
26]:
In the presence of an excess fluoride ion, HAP precipitates into calcium fluoride (CaF2), and the phosphate in HAP is released into the solution.
The relevant chemical reaction can be represented as Equation (2) [
22]:
According to the similar XRD patterns, SEM images, and BET surface area of CBC, there was no evidence indicating that the formation of CaF2 took place.
Table 4 shows the anion and cation concentrations of the solution before and after the adsorption. A detailed description of the fluoride concentration in the solution is given in
Table 1.
According to Equation (2), the phosphate in HAP should be released into the solution with the formation of CaF2. On the contrary, there was no evidence of phosphate in the solution.
According to the solubility product constant (Ksp) of CaF2 and the molar concentrations of Ca2+ and F− in the final solution, there was a possibility that CaF2 precipitated in the solution due to the reaction of F− in the solution and released Ca2+ from the CBC to the solution. The Ksp of CaF2 (3.4 × 10−11 mol3/L3) was calculated from the solubility of CaF2 (0.016 g/L in water at 20 °C). The molar concentration of Ca2+ and F− in the final solution was calculated as 2.7 × 10−10 mol3/L3, which exceeded the Ksp value. However, there was no visible CaF2 precipitation in the experimental setup.
Figure 8 shows the XRD patterns of HAP and CBC. Their similar patterns indicate that the major component of CBC was HAP.
Table 5 shows the number of moles of PO
43−, Ca
2+, F
−, and OH
− in 100 g of CBC before and after the fluoride adsorption based on chemical analysis. The Ca
2+/PO
43− molar ratio of 1.86 for CBC (before the fluoride adsorption) was similar to that of 1.67 for hydroxyapatite: [Ca
10(PO
4)
6(OH)
2] (HAP).
The number of moles of OH− in CBC before the fluoride adsorption was calculated based on the molar ratio of Ca2+:OH− (10:2) before the fluoride adsorption, assuming that the major component of CBC is hydroxyapatite. The number of moles of OH− in CBC after the fluoride adsorption was calculated based on the molar ratio of Ca2+:OH− (10:2) and by reducing the F− moles.
According to the chemical composition, CBC contained 65.3% HAP and 9% C on a weight basis. The percentage of HAP in the CBC was calculated by the sum of the percentages of Ca
2+ (27.7%) and PO
43− (35.2%) in the CBC digested with nitric acid and OH
− (2.4%), which was calculated from the molar ratio of Ca
2+:OH
− (10:2). The result obtained in our study was consistent with the literature. Brunson and Sabatini and Abe et al. have mentioned that bone char contains approximately 75% hydroxyapatite [Ca
10(PO
4)
6(OH)
2], 9–11% calcite (CaCO
3) [
22], and 8–10% C [
27]. Further, we could detect 0.6% Mg
2+, 0.5% Na
+, 0.1% K
+, 0.1% Cl
−, 1.3% N, and 0% CO
32− on a weight basis as trace components in the CBC. Ooi et al. also reported that Ca and P are the major components in bone char and that Na, Mg, O, and C are minor components in bone char based on their study of bovine bone char [
28].
The increase of Na
+ in the final solution (
Table 4) was mainly due to the addition of NaF to the solution to maintain the fluoride concentration. Cl
−, K
+, Mg
2+, and Ca
2+ ions, which were not present in the initial solution, were detected in the final solution (
Table 4) after 148 days of operation. This was due to the dissolution of those ions in the final solution from CBC as we detected them as components in CBC. A certain amount of Na
+ may also have been released into the solution by the dissolution from CBC as we also detected Na
+ as a trace component in CBC.
When the reaction of Equation (1) is taken into consideration, a certain amount of HAP was converted to FAP. Considering the number of OH
− moles in CBC before and after the experiment in
Table 5, 45.6% of HAP could be converted to FAP. According to Equation (1), the same molar of OH
− should be released into the solution; however, a significant change in pH value was not observed.
Table 6 shows the pH, number of OH
− moles, and alkalinity of the solutions.
The total amount of F
− removal was 97.6 mmol. This was calculated according to the data in
Table 1 (1855 mg/19 g/mol). The released OH
− could be partly neutralized by CO
2 dissolved from the atmosphere to produce alkalinity as we detected 762 μeq/L of alkalinity in the final solution.
In relation to the high adsorption by bone char, Mwaniki reported that Cl
− ions increased the rate of fluoride adsorption onto bone charcoal [
29]. Abe et al. also reported that fluoride adsorption by bone char increased in the presence of Cl
− ions in the solution. They discussed the “salting out” effect of NaCl that is relevant to the excess fluoride adsorption by bone char. NaCl dissociates in water by giving Na
+ and Cl
− ions to the solution. Na
+ and Cl
− ions in the solution are hydrated with water molecules by reducing the water molecules for the dissolution of fluoride. Therefore, the fluoride ion in the solution is enhanced to be adsorbed onto bone char [
27].
In contrast, our experiment showed that higher Cl
− concentrations decreased the fluoride adsorption capacity of CBC.
Figure 9 shows the Freundlich isotherm for CBC in the presence of chloride. According to the Freundlich isotherm, the adsorption capacities of CBC in the presence of Cl
− concentrations of 0 mol/L, 0.01 mol/L, 0.1mol/L, and 1.0 mol/L were 5.1 mg/g, 4.4 mg/g, 4.3 mg/g, and 3.6 mg/g, respectively, at a fluoride concentration of 10 mg/L. According to
Table 4, Cl
− ions were slightly released from CBC to the solution as 0.002 mol/L was detected in the final solution. The release of Cl
− from CBC to the solution caused a decrease in the adsorption of fluoride onto the CBC. Consequently, “salting out” was not the reason for the excess adsorption of fluoride.