Preparation of porous hybrid adsorbents based on fluor(calcium silicate)/activated carbon and its application in the removal of iron (III) from ammonium phosphate solutions

2.1 Materials

Diatomaceous earth (CP grade) was obtained from Chengdu Jinshan Chemical Co. Ltd., China. Sodium fluoride (GR grade) was produced by Tianjin Guangfu Fine Chemical Research Institute, China. Hydrated lime (AR grade) was purchased from Chengdu Kelong Chemical Co. Ltd., China. Deionized water was produced by Aquapro making water machine (ABZ1-1001-P) in laboratory. The sundried corn cob was impregnated with 60 wt% H₃PO₄ for 10 h, then calcined within a tubular calcinator (MTI corporation, USA, model:OTF-1200X) at 550 °C for 1 h, and activated carbon was obtained after washing by deionized water 3 times, filtering and drying at 105 °C for 6 h.

2.2 Preparation of FCS/AC

The raw materials including DE, AC and hydrated lime were mixed together with different mass mixing ratios (see Table 1), then added the additive sodium fluoride about 10% the mass of DE. The mixture was stirred and treated by wet-milling, forming suspension with a water-solid mass ratio of 15. Then transferred to a stirred autoclave heated at 140, 160 and 180 °C for 8, 10 and 12 h, respectively. The adsorbents FCS/AC were prepared after filtering, washing by deionized water 3 times, and drying at 105 °C for 6 h.

2.3 Characterization of FCS/AC

The samples need to be dried at 105 °C for 2 h, then ground and sieved for SEM, FTIR, XRD and BET measurements. The morphological structure of adsorbents was examined by using a S-4800 scanning electron microscope (SEM, Hitachi, Japan). The elemental composition of adsorbents before and after Fe³⁺ adsorption was analyzed by energy dispersive spectrometer (EDS, Hitachi, Japan). The functional groups of adsorbents were observed by fourier transform infrared spectroscopy (FTIR, Nexus 670, Nicolet iS50, USA). The XRD patterns of adsorbents were recorded with a X-ray diffractometer (D/Max-rA, Japan) using Cu Kα radiation. The BET specific surface area was measured from N₂ adsorption-desorption isotherms using a KUBO-X1000 adsorption analyzer (Builder, China).

2.4 Batch adsorption experiments

All the adsorption experiments were performed three times reflecting the standard deviation of their replicates. The relative uncertainty (u_r, %) of parameters (q_e, q_max,exp) was calculated in the experiments.

Stock solutions of Fe³⁺ ions (5000 mg/L) were prepared by dissolving ammonium ferric sulfate in deionized water. Ammonium phosphate solutions were prepared using ammonium dihydrogen phosphate (MAP) as raw material with percent content of 15% (The values of percent content calculated with P₂O₅, $W_{{\text{P}}_2{\text{O}}_5}(\%)=\frac{\frac{1}{2}\times \frac{m_{\text{MAP}}}{M_{\text{MAP}}}\times M_{{\text{P}}_2{\text{O}}_5}}{m_{\text{MAP}}+m_{{\text{H}}_2\text{O}}}\times 100\%$ ), and working solutions of Fe³⁺ ions ranging from 15 to 75 mg/L were prepared by diluting the stock solutions with MAP solutions. The batch adsorption experiments were conducted in 250 mL beakers and agitated on a temperature controlled agitator at 130 rpm.

In the adsorption process, the effect of pH values on the adsorption of Fe³⁺ onto FCS/AC in MAP solutions was investigated within the range of pH 2.0–5.0. The initial pH of Fe-MAP solutions (200 mL, 55 mg/L) were adjusted to the pre-determined pH values using NH₃·H₂O (1 + 5) and HCl (1 + 5). FCS/AC powder of 0.3 g was added to Fe-MAP solutions and the adsorption process was agitated at 45 °C for 400 min.

The effect of initial Fe³⁺ concentration on the adsorption of Fe³⁺ ions by adsorbents within the MAP solutions was evaluated by contacting FCS/AC and FCS (0.3 g) with Fe³⁺ solutions (200 mL) having initial concentration ranging from 15 to 75 mg/L, respectively. The mixture was agitated at 45 °C having a constant pH of 4.0 for 400 min.

The effect of temperature was performed by placing 0.3 g of FCS/AC in MAP solutions containing 200 mL (55 mg/L) of Fe³⁺ solutions at pH 4.0. The adsorption process was agitated at 15, 25, 35 and 45 °C for 400 min, respectively.

The adsorption kinetics studies were carried out by contacting FCS/AC and FCS (0.3 g) with Fe³⁺ solutions (200 mL) of the initial concentration 55 mg/L, respectively. The adsorption process was agitated at 45 °C and pH of 4.0. The solution samples of 2 mL were taken from the suspension at pre-determined time intervals (0–600 min) for the detection of residual Fe³⁺ concentration within MAP solutions.

In this work, the Fe³⁺ concentration was detected by phenanthroline spectrophotography (HJ/T 345-2007) equipped with a ultraviolet spectrophotometer (UV-1100). The method’s lowest detectable concentration for Fe³⁺ ions was 0.12 mg/L and the linear equation of standard curve (Y = 481.74X + 1.5507, the correlation coefficient of 0.9965) was obtained in the experiment. Iron amount q_e (mg/g) loaded on adsorbents after adsorption experiments and the percentage removal (Removal%) of Fe³⁺ ions from MAP solutions were calculated using the following equations:

2.5 Desorption experiments

In order to estimate the reversibility of Fe³⁺ adsorption, desorption studies were carried out by mixing FCS/AC (0.3 g) with Fe³⁺ solutions (55 mg/L, 200 mL). The mixture was stirred at 130 rpm, pH of 4.0 and 45 °C for 400 min. Solid residue was thoroughly washed 3 times with deionized water to remove any unabsorbed Fe³⁺ ions, and dried at 105 °C for 6 h. After that, the Fe³⁺ loaded adsorbents was regenerated by adding separately 50 mL of eluents (CH₃COOH, Ca(NO₃)₂, EDTA-2Na, and NaOH) into a beaker stirring at 130 rpm at 45 °C for 120 min. The Fe³⁺ concentration present in the eluents was directly measured. The percentage desorption (Desorption %) of Fe³⁺ ions from loaded adsorbents in MAP solutions was calculated from the following equation:

3.1 Preparation and characterization of FCS/AC

FCS held excellent adsorbability for phosphate, the process realized by converting FCS into apatite on the surface. Furthermore, the apatite adsorbed many metal ions very well due to the excellent ions adsorbability and exchangeability on its surface (Corami et al., 2008). The adsorbent AC held huge specific surface area, whereas the adsorption of AC for Fe³⁺ ions was poor due to the physisorption process. Therefore, composites of FCS blended AC were prepared to make full use of their advantages. Table 1 listed the properties of FCS/AC composites with different preparing conditions (weight ratios of diatomaceous earth/hydrated lime/activated carbon (DE/HL/AC), temperature and reaction time). The mean diameter values D50, D90 and surface mean diameter (SMD) were 23.63–37.30 μm, 43.30–72.36 μm and 11.96–17.61 μm, respectively, which showed a significant difference with changing conditions. According to Table 1, FCS/AC composites exhibited higher specific surface area than pure FCS and DE due to the introduction of AC, which provided large interface between adsorbents and metal ions in MAP solutions. Researchers (Zhu et al., 2016; Karchemsky et al., 2013) found that a suitable mean diameter promoted the dispersion of adsorbents in solution, and the high specific surface area not only provided more adsorption and ion exchange sites, but also enhanced the adsorbability of the material. Based on the single factor analytic approach (Li et al., 2012), a suitable mean diameter 11.96 μm and the large specific surface area 155.35 m²/g of FCS/AC could be obtained.

The effect of different preparing conditions on the adsorption of FCS/AC for Fe³⁺ ions was shown in Table 1. It was found that the Fe³⁺ adsorption of 10.12 mg/g for FCS/AC was more than 7.07 mg/g, 1.12 mg/g for FCS and AC, respectively. Which was attributed to a synergy between FCS and AC. For the hydrothermal process, the material ratio, temperature and reaction time had a significant effect on the properties of products. Analyzing the aforementioned factors including the mean diameter, specific surface area and Fe³⁺ adsorption amount, the optimum preparation conditions for adsorbent FCS/AC were determined as DE/HL/AC weight ratio of 10/8.5/2.5, temperature of 160 °C and reaction time of 10 h in this work and used for further adsorption experiments. Moreover, the pure FCS was studied experimentally to show the excellent adsorption properties of FCS/AC by contrasting both adsorbents.

From Fig. 1(a and b), there were similar patterns for FCS/AC and FCS, and the diffraction peaks of tobermorite (PDF card 29-0329, Ca₅Si₆O₁₆(OH)₂) appeared at around 29.6°, 28.3° and 50.0°. Moreover, the diffraction peaks at around 26.6° and 20.9° were characteristic peaks of the SiO₂ and gismondine (PDF card 20-0452, CaAl₂Si₂O₈·4H₂O). These results indicated that the principal phase compositions of both adsorbents were tobermorite, gismondine and some residual SiO₂. According to Fig. 1(c), the principal peaks of giniite (PDF card 45-1436, Fe₅(PO₄)₄(OH)₃·2H₂O) emerged at around 26.7° and 27.9° after Fe³⁺ adsorption, and the peaks of dufrenite (PDF card 22-1143, CaFe₁₂(PO₄)₈(OH)₁₂·4H₂O) appeared at around 28.3°, 21.6° and 46.8°. Analyzing the XRD patterns of adsorbents before and after Fe³⁺ adsorption, it was found that Fe³⁺ adsorption process was realized by converting flour(calcium silicate) into giniite and dufrenite on surface.

Close

Fig. 1 XRD patterns of adsorbents: (a) FCS and (b) FCS/AC adsorbents before adsorption; (c) FCS/AC adsorbent after adsorption.

Export to PPT

FTIR spectra of AC and FCS/AC were obtained for characterization of their chemical structures, and the results given in Fig. 2. As shown in Fig. 2, the FTIR spectrum of FCS/AC differed from that of AC, indicating that the surface functional groups of FCS/AC were changed partially with blending FCS. There was a broad characteristic peak at around 3402 cm⁻¹ consistent with the stretching vibration of O—H and N—H bonds. The peak at around 2922 cm⁻¹ was attributed to the C—H stretching vibration of methylene group. The peaks at around 1578 and 1400 cm⁻¹ were the N—H in-plane bending vibration of primary amino group and the C—O stretching vibration of carboxyl group, respectively. In contrast with AC, the peaks related to FCS/AC shifted to higher wave numbers, which can be representative of introducing the electron withdrawing fluorine anion, thus producing an inductive effect (Larkin, 2011). Furthermore, the peak of AC at around 1168 cm⁻¹ was the C—O stretching vibration of alcohol group, however, the peak of FCS/AC moved to 983 cm⁻¹ attributed to the C—O bond of alcohol group substituted to the C—F bond. This could be deduced that the surface functional groups of FCS/AC were shifted and changed improving the Fe³⁺ adsorption from MAP solutions. In that the affinity of FCS/AC for Fe³⁺ ions enhanced by introduction of electrondrawing group, and the clathrate [FeF₆]³⁻ developed by the chelating ability between F^- anion and Fe³⁺ cation.

Close

Fig. 2 FTIR spectra of AC and FCS/AC adsorbents.

Export to PPT

Fig. 3 showed the N₂ adsorption-desorption isotherms of the FCS/AC and AC. It was clearly seen that both adsorbents adsorbed high nitrogen in the low relative pressure domain (P/P₀ < 1), which suggested the presence of textural porosity (Singh et al., 2017). Both adsorbents indicated type I isotherm as defined by IUPAC classification with the H₃ hysteresis loop, indicated the presence of mesopores. Fig. 4 exhibited the porous morphologies of FCS/AC and FCS. It was found that both adsorbents exhibited well-developed space connected cellular structure with an average diameter around 10 μm corresponding to the mean diameter 11.96 μm (from Table 1), and showed a serious aggregate. The surface of FCS was smoother and the agglomerated particle diameter was larger than FCS/AC, because the AC was intercalated into the surface and pores of FCS in the hydrothermal process thus forming smaller diameter aggregate.

Close

Fig. 3 N₂ adsorption-desorption isotherm curves of adsorbents.

Export to PPT

Close

Fig. 4 The porous morphologies of FCS/AC (a) and FCS (b) by scanning electron microscope.

Export to PPT

3.2 Effect of initial solution pH

The pH values in the range of 2.0–5.0 were chosen based on the reason that Fe³⁺ ions would precipitate abundantly when the pH was higher than 5.0 (Khan et al., 2013). The effect of solution pH on the adsorption of FCS/AC for Fe³⁺ ions was carried out and the results were presented in Fig. 5. It was found that the Fe³⁺ adsorption was highly dependent on solution pH, because solution pH affected the solubility of metal ions and ionization state of the functional groups (Li et al., 2012). The Fe³⁺ adsorption was 10.12 mg/g with the solution pH up to 4.0, and further increased to 10.12–17.34 mg/g within the range of pH 4.0–5.0. At low pH, metal ions and protons competed for binding sites and adsorbents would dissolve partially, leading to low adsorption amount for Fe³⁺ ions. However, there was still a certain amount of adsorption of Fe³⁺ ions at low pH based on the surface adsorption process provided by the high specific surface area and the surface functional groups of adsorbents. Moreover, the adsorption amount reached the maximum within the range of pH 4.5–5.0, because the number of negatively charged active sites (P—O⁻, —OH, —NH₂ and —CH₂F) and the electrostatic attraction increased resulting in an increase in adsorption amount. The pH range of 3.5–4.5 was feasible in MAP industry, and the pH of 4.0 was chosen for adsorption process in this study and used for further adsorption experiments.

Close

Fig. 5 The effect of solution pH on the adsorption of Fe³⁺ onto FCS/AC. u r ( q e ) ⩽ 13.43 % .

Export to PPT

3.3 Effect of initial Fe³⁺ concentration

Fig. 6 showed the effect of initial Fe³⁺ concentration on the adsorption of Fe³⁺ onto adsorbents. It indicated that the amount of Fe³⁺ ions absorbed onto FCS and FCS/AC both increased with the increasing initial Fe³⁺ concentration. Furthermore, the Fe³⁺ adsorption increased steadily in initial concentration up to 55 mg/L and then increased slowly. The reason was that the adsorbents provided enough adsorption sites with low initial Fe³⁺ concentration resulting in the Fe³⁺ adsorption increasing steadily, and excessive Fe³⁺ ions competitively adsorbed on the same adsorption sites leading to the reduction of adsorption rate per unit adsorbent. This phenomenon was similar to the results of other researchers (Zhu et al., 2017; Soliman et al., 2011). The maximum Fe³⁺ adsorption of 11.07 mg/g for FCS/AC was more than that of pure FCS, which further confirmed that the adsorbents intercalated AC exhibited excellent Fe³⁺ adsorption in MAP solutions. The suitable concentration below 30 mg/L had a neglectable effect on MAP crystal morphology, whereas the Fe³⁺ concentration in MAP solutions commonly located in the range of 45–75 mg/L. Therefore, the cycling adsorption process for the high concentration of Fe³⁺ ions in MAP solutions was carried out. As the initial Fe³⁺ concentration up to 55 mg/L, the final concentration was below 30 mg/L after twice adsorption and even as low as 18.34 mg/L after thrice. The removal efficiency of Fe³⁺ ions was more than 66% after cycling adsorption.

Close

Fig. 6 Effect of initial Fe³⁺ concentration on the adsorption of Fe³⁺ onto FCS/AC. u r ( q e ) ⩽ 6.27 % .

Export to PPT

Table 2 listed a comparison of the maximum Fe³⁺ adsorption for FCS/AC with some commonly used adsorbents reported in the literatures. Although direct comparison of FCS/AC with other adsorbents was difficult because of the differences in experimental conditions, it showed that the maximum Fe³⁺ adsorption for FCS/AC was higher than most of the other adsorbents listed in Table 2. The Fe³⁺ adsorption of 19.8 mg/g and 69.8 mg/g was obtained in other researchers. The reason was that the adsorbents were different and the adsorption process in aqueous solutions differed from the MAP solutions of this work, resulting in a great difference in adsorption amount. Moreover, the Fe³⁺ adsorption for DE and AC was only 0.32 mg/g and 1.12 mg/g from Table 1, respectively. Therefore, FCS/AC showed an excellent adsorption for Fe³⁺ ions and provided a potential application in the adsorption of Fe³⁺ ions from MAP solutions.

3.4 Adsorption isotherms

The equilibrium isotherm is fundamental in describing the interactive behaviors between adsorbents and solutes, together with illustrating adsorbent properties and affinities (Langmuir, 1918). In this work, the experimental data were fitted using adsorption models including Langmuir, Freundlich and Dubinin-Kaganer-Radushkevich (DKR) isotherms.

The Langmuir isotherm assumes a monolayer adsorption onto a homogenous surface where only a finite number of adsorption sites of uniform strategies for adsorption with no transmigration of adsorbate in the plane of surface (Allen et al., 2003). The linear form of the Langmuir isotherm is given by Eq. (4):

The Freundlich isotherm is based on the assumption of an exponentially decaying adsorption site energy distribution (Freundlich, 1906), which applied to describe heterogeneous system and the linearized Freundlich isotherm model is expressed as:

The DKR isotherm has been employed for description the adsorption of metal ions on adsorbent (Khan et al., 2012; Erdem et al., 2004; Mobasherpour et al., 2011). And the DKR equation is given by Eq. (7):

Fig. 7 showed the equilibrium isotherms for the adsorption of Fe³⁺ on FCS/AC and FCS, and the isotherm parameters obtained from Fig. 7 were presented in Table 3. The correlation coefficients R² (0.994, 0.973 and 0.987 for Langmuir, Freundlich and DKR models, respectively) provided good agreement between theoretical models and experimental data in this study, and the Langmuir isotherm model gave the highest R² value indicating the equilibrium data of Fe³⁺ on FCS/AC were best represented by this model. Similarly, the adsorption of Fe³⁺ onto FCS was well described by the Langmuir isotherm. The reason was attributed to the homogeneous nature of active sites on the surface of both adsorbents, and a monolayer adsorption occurred between solute and adsorbent. The greater affinity between adsorbent and adsorbate inferred when R_L was smaller giving an indication on the tendency of the adsorption process. The R_L value of FCS was smaller than that of FCS/AC inferred that the affinity of the adsorbent for Fe³⁺ ions reduced after blending with AC. The maximum adsorption q_m calculated from Langmuir model was 15.41 and 9.64 mg/g for FCS/AC and FCS, respectively. It could be found that the adsorption amount enhanced after blending with AC although the affinity reduced. The Frendlich model was applied to describe heterogeneous system characterized by a heterogeneity factor of n, and the value of n⁻¹ (0.49, 0.43) demonstrated that the adsorption amount was only slightly decreased at lower equilibrium concentrations indicating multilayer adsorption on the surface. It was known that magnitude of apparent adsorption energy E was used to estimate the adsorption type and if the value was between 8 and 16 kJ/mol the type could be explained by ion-exchange (Lin and Juang, 2002; Wang et al., 2004). Therefore, the values (10.00–11.18 kJ/mol) calculated for the DKR model showed an ion-exchange mechanism for both adsorbents.

Close

Fig. 7 (a) Equilibrium adsorption and (b, c, d) adsorption isotherms for Fe³⁺ ions adsorbed on adsorbents.

Export to PPT

3.5 Effect of temperature and adsorption thermodynamics

The effect of temperature on the adsorption of Fe³⁺ onto FCS/AC was researched at 288, 298, 308 and 318 K, and the results indicating that the adsorption amount of Fe³⁺ ions increased with increase in temperature.

Thermodynamic parameters are calculated by Eqs. (9) and (10):

The linear fit of thermodynamics for the adsorption process was presented in Fig. 8. The values of △G⁰ were calculated to be −5.92, −9.87, −12.27 and −14.64 kJ/mol for Fe³⁺ adsorption process at 288, 298, 308 and 318 K, respectively. Negative values of △G⁰ indicated the feasibility of the process and spontaneous nature of the adsorption, furthermore, the △G⁰ values decreased with an increase in temperature suggesting an increased trend in the degree of spontaneity and feasibility of Fe³⁺ adsorption. The △H⁰ value was obtained to be 76.63 kJ/mol for Fe³⁺ adsorption onto FCS/AC. This result confirms the endothermic nature of the adsorption process due to the positive values of △H⁰, and demonstrated the chemisorption process of Fe³⁺ adsorption because of the adsorption heat between 20 and 400 kJ/mol (Ünlü and Mustafa, 2007). The △S⁰ value of 288.02 J/(mol·K) suggested that the degree of freedom increased at the solid-liquid interface during the adsorption. The reason was that the activation energy of Fe³⁺ adsorption decreased with the increase in temperature. The adsorbent surface sites were activated to complex Fe³⁺ ions in MAP solutions, resulting in the increase affinity of Fe³⁺ onto FCS/AC and the randomness in the adsorbent-adsorbate system (Baraka et al., 2007).

Close

Fig. 8 Linear fit of experimental data obtained by plotting of lnK_d versus 1/T for Fe³⁺ adsorption on FCS/AC.

Export to PPT

3.6 Effect of contact time and adsorption kinetics

The effect of contact time on the adsorption of Fe³⁺ onto FCS/AC and FCS was shown in Fig. 9. From Fig. 9, the adsorption of Fe³⁺ onto both adsorbents increased with an increase in contact time, and the adsorption equilibrium was achieved in 400–600 min. Moreover, the Fe³⁺ adsorption rate increased rapidly in contact time up to 400 min and then increased slowly. Therefore, a time of 400 min for the Fe³⁺ adsorption process was selected for consideration pertaining to the feasibility and economics of the process.

Close

Fig. 9 The effect of contact time on the adsorption of Fe³⁺ onto adsorbents. The inset showed the fitted results by the pseudo-second order model. u r ( q e ) ⩽ 19.28 % .

Export to PPT

The different kinetic models including pseudo-first order, pseudo-second order and intraparticle diffusion are employed to investigate the mechanism of adsorption and potential rate controlling steps such as chemical reaction, mass transport and diffusion control processes (Li et al., 2012). The pseudo-first order and pseudo-second order are generally expressed as Eqs. (11) and (12), respectively:

Intraparticle diffusion equation suggests that intraparticle diffusion is the rate-limiting step in adsorption. The diffusion process may affect the adsorption of Fe³⁺ on FCS/AC and FCS due to the porous structure of both adsorbents. Therefore, the intraparticle diffusion equation is used to explore the behavior of intraparticle diffusion, and the equation is given by Eq. (13):

The experimental data have been fitted by the aforementioned kinetic models, and the kinetic parameters were shown in Table 4. Based on the R² values of the kinetic models, the pseudo-second order kinetic model provided preferable match between theoretical and experimental value. Moreover, the fitting results of the pseudo-second order model were presented in the inset of Fig. 9, which showed the theoretical simulated curve fitting the experimental data quite well. Thus, the pseudo-second order model was more appropriate to describe the kinetic behaviors for Fe³⁺ adsorption onto FCS/AC and FCS adsorbents, indicating that the chemisorption was the rate-limiting step in adsorption process.

The intraparticle diffusion model was widely used to predict the rate-limiting step (Weber and Morris, 1963), and the results were shown in Fig. 10 and Table 4. It could be seen that the plots were mult-linearity over the whole time range, suggesting that three steps were operational in the adsorption of Fe³⁺ onto both adsorbents. The first linear plot was the instantaneous adsorption or external surface adsorption attributing to the rapid consumption of the available adsorption sites on the adsorbent surface, the second stage was the gradual adsorption stage where the intraparticle was the rate-limiting step, and the third portion was attributed to the final equilibrium for which the intraparticle diffusion starts to slow down due to the extremely few adsorption sites left on adsorbent. This kind of multi-linearity in the shape of intraparticle diffusion plot was also observed in the adsorption of Fe³⁺ ions from aqueous solutions by silica gel sorbents (Sui et al., 2016) and in the removal of dye and heavy metal from aqueous solutions by poly(vinyl alcohol)/chitosan coposites (Li et al., 2012). Based on the results of the aforementioned kinetics, which indicated that both chemisorption and intraparticle diffusion affected the Fe³⁺ adsorption onto FCS/AC and FCS.

Close

Fig. 10 Intraparticle diffusion model for the adsorption of Fe³⁺ onto adsorbents.

Export to PPT

3.7 Mechanism of Fe³⁺ adsorption

As seen in Fig. 11, the peaks of iron and phosphorus appeared in the EDS spectra after Fe³⁺ adsorption, indicating that the FCS/AC loaded with Fe³⁺ ions and phosphate. The calcium atomic ratio decreased sharply from 6.85% to 1.08% after Fe³⁺ adsorption according to Table 5. This phenomenon demonstrated that calcium cations were exchanged by iron cations and produced new substance on surface of adsorbent. Furthermore, the calcium silicate was apt to adsorb phosphate formation apatite in phosphate solutions, and calcium cations in apatite were easily ion-exchanged with many other metal ions (Smičiklas et al., 2006). Based on Fig. 1, the new substance of Fe₅(PO₄)₄(OH,F)₃·2H₂O and CaFe₁₂(PO₄)₈(OH,F)₁₂·4H₂O was observed after adsorption. In consideration of the new product CaFe₁₂(PO₄)₈(OH,F)₁₂·4H₂O, we believed that the intermediate apatite was existence in adsorption process. CaFe₁₂(PO₄)₈(OH,F)₁₂·4H₂O was obtained when iron cations substituted calcium cations partially, and Fe₅(PO₄)₄(OH,F)₃·2H₂O produced as substituted thoroughly. Based on the effect of pH on adsorbents (From Fig. 5) and EDS analysis, it was concluded that the adsorption process was realized by ion-exchange and surface adsorption mechanism. In addition, the number moles of Fe³⁺ absorbed onto FCS/AC were calculated to be 0.1812 mmol/g. The Fe³⁺ adsorption about 0.1614 mmol/g was absorbed onto FCS/AC by ion-exchange mechanism (Calculated from Table 5), and the proportion contributed by ion-exchange was 89.07%. Which was consistent with the experiment date of 88.93%, indicated that the adsorption mechanism of Fe³⁺ ions was mainly ion-exchange.

Close

Fig. 11 EDS spectra of FCS/AC (a) before and (b) after Fe³⁺ adsorption.

Export to PPT

In order to further investigate this adsorption process, the mechanism of Fe³⁺ adsorbed onto FCS/AC adsorbents was described in Scheme 1, which could be explained as following processes: firstly, FCS/AC adsorbed phosphate forming apatite preferentially in MAP solutions; secondly, calcium cations of apatite were exchanged by iron cations rapidly. In addition to the ion-exchange of Fe³⁺ ions with Ca²⁺ ions and chelation of Fe³⁺ ions with fluorinion, the adsorption process of Fe³⁺ onto FCS/AC was also taken place between Fe³⁺ ions and amino, hydroxyl and halogenated groups (From Fig. 2). So a stronger adsorbability was observed for FCS/AC than that of pure FCS.

Close

Scheme 1 The process of Fe³⁺ adsorbed onto FCS/AC.

Export to PPT

3.8 Desorption and reusability

The desorption-resorption experiments were employed to investigate the potential of FCS/AC for industrial application by the batch method using different eluents including CH₃COOH, Ca(NO₃)₂, EDTA-2Na and NaOH. Fig. 12 showed the efficiencies of different eluents releasing Fe³⁺ ions from loaded absorbent. It was observed that the highest desorption value for Fe³⁺ ions (99.86%) was obtained using EDTA-2Na (0.05 M) solutions. However, the resorption value of Fe³⁺ onto FCS/AC was below 20% because that Fe³⁺ ions and Ca²⁺ ions in the adsorbent skeleton released simultaneously resulting in the adsorbent skeleton breakdown, and the active sites occupied by excess EDTA-2Na reducing adsorbability due to the strong chelation of EDTA-2Na with cations on adsorbents. The effect of different concentration of EDTA-2Na on the adsorption of Fe³⁺ onto FCS/AC was discussed, and the results were shown in Fig. 12. It was found that the desorbed Fe³⁺ ions from loaded FCS/AC decreased with the decrease of EDTA-2Na concentration, while the Fe³⁺ ions resorbed on adsorbents was increased heavily with the concentration reducing to 0.0025 M and even down to 0.001 M. The resorption value of Fe³⁺ onto adsorbents was up to 75%. Table 6 listed the valuation of pH before and after desorption, it showed that a difference between initial pH and final pH. The solution pH increased after desorption with low EDTA-2Na concentration but opposite with high concentration due to the effect of excess EDTA-2Na, indicating that EDTA-2Na was an excellent eluent for Fe³⁺ adsorption within the range of 0.001–0.0025 M.

Close

Fig. 12 Desorption-resorption efficiencies of Fe³⁺ onto FCS/AC from MAP solutions.

Export to PPT

It was reported that CH₃COOH (HAc) and Ca(NO₃)₂ solutions were effective eluents for metal ions desorption, because that the adsorbent was unstable in acidic solution and the process of cationic exchange was reversible between the adsorbent surface and the solution (Zhu et al., 2017; Simon et al., 2008). From Fig. 12, the desorbed value of both eluents was below 10% with the resorption value up to 48%, this phenomenon could be explained that forces of Fe³⁺ onto FCS/AC were strong and the surface ions were desorbed only. However, enough active sites exposed after desorption due to the high specific surface area of adsorbents (Table 1 and Fig. 4) resulting in the increase of resorption value, which was similar to the desorption-resorption process of low EDTA-2Na concentration (0.001 M).