Evaluation of phosphate ion adsorption from aqueous solution by nickel-aluminum complex hydroxides

Phosphates, as essential and often limiting nutrients in most aquatic environments, can result in the acceleration of eutrophication, which leads to an increase in costs associated with water treatment and poses a potential threat to human health and the ecological balance (Dodds et al. 2008). At the same time, phosphorus resources in the world may only last 50 to 100 years. Therefore, for both these reasons, it is necessary to effectively remove (or adsorb) phosphate ions from aqueous solutions (Laroussi et al. 2003; Choe et al. 2004; Bennett & Elser 2011).

Currently, numerous methods such as chemical precipitation, electrocoagulation, crystallization, adsorption, reverse osmosis, and biological treatments are available for the removal of phosphate ions from wastewater (Den & Shi 2015). Among these methods, the adsorption process shows advantages of being operationally straightforward, highly efficient, and relatively inexpensive. Various types of adsorbents have been tested for phosphate ion adsorption (Yang et al. 2014). In recent years, the development of composite adsorbents containing two or more different metal oxides for phosphate ion adsorption has gained increasing attention (Li et al. 2014). Previous studies have reported on the use of Fe-Co binary oxide, Zn-Al LDH, Zr-Fe oxide nanoparticles, Fe-Mn binary oxide, and Fe-Mn-Al trimetal oxide for the adsorption of phosphate ions from aqueous solution. Composite adsorbents exhibit superior adsorption performance, since they possess the inherent advantages of their parent materials (Zhang et al. 2009; Lü et al. 2013; Li et al. 2014; Yu et al. 2015; Zhang et al. 2017).

In our previous studies, which focused on enhancing phosphate ion adsorption, we found that nickel-cobalt binary hydroxide showed excellent adsorption of phosphate ions from aqueous solution (Ogata et al. 2017). However, cobalt is expensive and has the potential to adversely affect health and the aquatic environment (including the biota). Consequently, we investigated materials to replace cobalt in the adsorbent and reported that aluminum oxides have excellent potential to adsorb phosphate ions (Kawasaki et al. 2008). In the aforementioned study, we anticipated that nickel-aluminum complex hydroxides would be promising as adsorbents for phosphate ions. However, until now, no systematic investigation of phosphate ion adsorption by nickel-aluminum complex hydroxides has been reported.

Therefore, the object of this study was to explore the feasibility of using composite metal oxides derived from nickel and aluminum for the adsorption of phosphate ions. An investigation of the physicochemical characteristics of nickel-aluminum complex hydroxides was conducted. Their kinetic and isothermal behavior was investigated, and the parameters influencing phosphate ion adsorption, including solution pH and adsorption temperatures, were also studied. Additionally, desorption of phosphate ions from nickel-aluminum hydroxides was performed to assess their regeneration capacity.

Nickel-aluminum complex hydroxides with molar ratios of Ni²⁺ to Al³⁺ of 4.0, 3.0, 2.0, 1.0, and 0.5, referred to as NA41, NA31, NA21, NA11, and NA12, respectively, were prepared by the co-precipitation method. The characteristics of the adsorbents were investigated: the specific surface area, morphologies, and crystallinities were measured using a NOVA 4200e specific surface analyzer (Yuasa Ionic, Japan), scanning electron microscopy (SEM, SU1510, Hitachi High-Technologies Co., Japan), and X-ray diffractometry (XRD, MiniFlex II, Rigaku, Japan), respectively. The number of hydroxyl groups was measured using the fluoride ion adsorption method reported by Yamashita et al. (1978). The surface pH of the adsorbent was measured using a digital pH meter F-73 (HORIBA, Ltd, Japan). Briefly, the adsorbent (0.1 g) was added to distilled water of pH 7.0 (50 mL), the suspensions were shaken for 2 min at 25 °C, and filtered through a 0.45 μm membrane filter. Elemental analysis and electron spectroscopy were carried out using an electron microanalyzer (EPMA, JXA-8530F, JEOL, Japan) and an X-ray photoelectron spectroscopy system (AXIS-NOVA, Shimadzu Co., Ltd, Japan), respectively.

The sample solution of phosphate ions was prepared by potassium dihydrogen phosphate. The solution pH was adjusted by hydrochloric acid and sodium hydroxide solution. All reagents (special reagent grade) were purchased from Wako Pure Chemical Industries, Ltd, Japan.

In addition, the concentration of sulfate ions in the filtrate was measured by ion chromatography (DIONEX ICS-900, Thermo Fisher Scientific Inc., Japan). The measurements were performed using the IonPac AS12A system (4 × 200 mm, Thermo Fisher Scientific Inc., Japan). The mobile phase and the regenerant comprised 2.7 mmol/L Na₂CO₃ + 0.3 mmol/L NaHCO₃ and 12.5 mmol/L H₂SO₄, respectively. The flow rate was 1.0 mL/min at ambient temperature. The micro membrane suppressor was an AMMS 300 system (4 mm, Thermo Fisher Scientific Inc., Japan) and the sample volume was 10 μL.

First, a number of samples were prepared by adding 0.05 g of adsorbent (NA11) to 50 mL of a 50 mg/L phosphate ion solution. The suspension pH was adjusted by hydrochloric acid or sodium hydroxide solution to obtain a range of 2.0–12.0, and then the sample solutions were shaken at 100 rpm for 24 h at 25 °C. Secondly, the adsorbent (NA11, 0.05 g) was added to a 50 mL of a 50 mg/L phosphate ion solution. The suspensions were shaken for 1, 3, 5, 8, 12, 16, 20, and 24 h at 100 rpm at 25 °C. Finally, the adsorbent (NA11, 0.05 g) was added to 50 mL of phosphate ion solution that had concentrations ranging from 10 to 100 mg/L. The suspensions were shaken for 24 h at 100 rpm at 5, 25, and 45 °C. The amount of phosphate ion adsorbed in each case was calculated using Equation (1).

Figure 1 shows SEM images of the adsorbents. A hexagonal crystal system was observed in NA41 and NA31, but an irregular crystal system was noted with increasing amounts of aluminum. These phenomena indicated that the constituent elements (aluminum and nickel) affect the crystal system. Additionally, the changes of crystal system were slightly between NA12 and NA41. The XRD patterns of the adsorbents (shown in Figure 2) were consistent with powder diffraction file 2010 obtained from the International Center for Diffraction Data, indicating that all the adsorbents consisted mainly of nickel hydroxide. In addition, NA41 had a highly crystalline nature, which decreased with increasing amounts of aluminum, suggesting that NA12 and NA11 (amorphous nature) are highly reactive. Table 1 shows the characteristics of the adsorbents. Specific surface areas and the concentration of hydroxyl groups in NA were greater with increasing amounts of aluminum. NA12 (specific surface area: 26.4 m²/g, concentration of hydroxyl groups: 1.62 mmol/g) and NA11 (specific surface area: 22.8 m²/g, concentration of hydroxyl groups: 1.92 mmol/g) have high specific surface areas and hydroxyl group concentrations. However, the constituent element ratios did not appear to significantly affect the surface pH in this study (pH 6.50−7.98).

The amount of saturated phosphate ions adsorbed is shown in Figure 3; NA41 adsorbed the least (70.2 mg/g), and adsorption increased in the order NA31 (82.1 mg/g) < NA21 (121.1 mg/g) < NA12 (134.9 mg/g) < NA11 (146.8 mg/g). The adsorption capacity of NA11 was also greater than those reported previously for other adsorbents (Kawasaki et al. 2008, 2010; Ogata et al. 2016, 2017). These results indicate that the nickel-aluminum complex hydroxide is an efficient adsorbent of phosphate ions from aqueous solutions.

In addition, we evaluated the relationship between the amount adsorbed and the characteristics of the adsorbent to elucidate the adsorption mechanism of phosphate ions. First, the relationship between the specific surface area or the number of hydroxyl groups and the amount of phosphate ions adsorbed was evaluated. The correlation coefficients were 0.818 and 0.918, respectively, which indicated that the amounts adsorbed are related to the specific surface area and the number of hydroxyl groups. Second, we observed the adsorbent surface before and after adsorption (Figure 4) and recorded an increase in phosphorus content after the adsorption treatment. We interpret this observation to mean that one of the adsorption mechanisms of phosphate ions using NA may involve a combination of electrostatic interaction, ligand exchange, and ion exchange (Yang et al. 2014). Third, we measured the binding energy of the adsorbent surface (NA11) before and after the adsorption treatment (Figure 5). After adsorption, P(2s) and P(2p) peaks were detected and the S(2p) peak intensity was significantly reduced, which suggested that phosphate ions combined with NA11 and, at the same time, sulfate ions were released (the NA was prepared using sulfates). Finally, we measured the concentration of phosphate and sulfate ions after adsorption (data not shown). We confirmed that the amounts of phosphate ions adsorbed by NA11 and the amounts of sulfate ions released were positively correlated in this experiment (correlation coefficient: 0.837), supporting the theory that sulfate ions interlayered in the NA are exchanged with phosphate ions in the sample solutions. The aforementioned mechanisms are similar to those reported by previous studies (Li et al. 2014; Lu et al. 2014; Yang et al. 2014; Yu et al. 2015; Wang et al. 2018). However, the release of sulfate ion is not good point in the wastewater treatment. Therefore, if NA11 is applied in the field, it must be considered the sulfate ion concentration after adsorption treatment. In addition, Table 2 shows the comparison of phosphate ion adsorption capacity of NA11 with other reported adsorbents. It is evident that the phosphate ion adsorption capacity of NA11 is higher than that of other adsorbents. Therefore, this result indicated that NA11 has the greatest potential for the adsorption of phosphate ion from the aqueous phase.

The effect of solution pH on the adsorption of phosphate ions is shown in Figure 6; the amount adsorbed at an initial pH of 4.0 to 10.0 (equilibrium pH is 6.2–7.7) was greater than that at an initial pH of 2.0 (equilibrium pH is 3.2). When the pH value is between 3.0 and 7.2, the main species in solution is monovalent H₂PO₄⁻, while at a pH between 7.2 and 11.0, the predominant species of phosphate is HPO₄²⁻ (Rodrigues & de Silva 2009). Previous studies indicated that the H₂PO₄⁻ species were more easily adsorbed on the metal (hydr)oxide surface than other species (Rodrigues & de Silva 2009; Lai et al. 2016). In addition, the results demonstrated that NA11 has a buffering capacity. This phenomenon could also explain why no obvious changes in the amount of adsorbed phosphate ions were observed in the pH range between 4.0 and 10.0. Consequently, it seems likely that NA11 would be suitable for the adsorption of phosphate ions from environmental water for a realistic range of operational conditions.

The kinetic data of the adsorption of phosphate ions onto NA11 were fitted to the pseudo-first-order kinetic model and pseudo-second-order kinetic model; the calculated parameters are listed in Table 3. From comparisons of the correlation coefficients (r), it is evident that the fit of the pseudo-second-order model (1.000) is better than pseudo-first-order model (0.969), although it can be seen from the correlation coefficients that both models fit the data well. In addition, the amount of adsorbent calculated, q_e,cal (50.3 mg/g), in the pseudo-second-order model was closer to the actual experimental amount of adsorbent used, q_e,exp (49.8 mg/g). These results indicate that chemisorption is possibly the dominant mechanism in this adsorption reaction (Sposite 1984; Lu et al. 2014).

The amount of phosphate ions adsorbed by NA11 increases with increasing temperature; the temperature dependence of the adsorption capacity of different aqueous equilibrium concentrations can be illustrated by adsorption isotherms (Figure 7). The Langmuir and Freundlich isotherm models were used to investigate the equilibrium adsorption data. The Langmuir isotherm model assumes that the adsorbent is structurally homogeneous; namely, the adsorption process results in a monolayer with uniform adsorption energies, and there is no interaction among the adsorbed molecules. On the other hand, the Freundlich isotherm model can be applied to multilayer adsorption on surfaces with heterogeneous energetic distribution of active sites; in contrast to the Langmuir model, there is interaction among the adsorbed molecules (Liu et al. 2016).

Table 4 shows the Freundlich and Langmuir constants for the adsorption of phosphate ions; the correlation coefficients of the respective isotherm models are 0.857–0.987 and 0.764–0.960. The Langmuir constant (W_s) increased with increasing adsorption temperatures (5–45 °C), which corresponded to the trend observed in the data in Figure 7. In addition, phosphate ions were easily adsorbed onto the surface of NA11 when 1/n was in the range 0.1–0.5, but not when 1/n> 2 (Abe et al. 1967). The data obtained in our research (1/n: 0.29–0.36) suggested that the adsorption of phosphate ions on the NA11 surface occurred readily, as a result of chemisorption. In the present study, NA11 was produced by the co-precipitation method, which might lead to a more heterogeneous surface of NA11; therefore, the Freundlich equation would be more suitable to describe the isotherm data (Lu et al. 2014).

The thermodynamic parameters for the adsorption of phosphate ions onto NA11 are shown in Table 5. The positive value of ΔH (2.23 kJ/mol) confirmed the endothermic nature of the adsorption and was in good agreement with the experimental results reported elsewhere (Yu et al. 2015). In addition, the ΔG value decreased from − 4.31 to −5.26 kJ/mol when the temperature was increased from 278 to 318 K, implying an increase in the spontaneity of the reactions (Liu et al. 2012). In addition, the ΔG values were greater than −20 kJ/mol, suggesting that electrostatic interaction played a significant role in the adsorption process (Weng et al. 2007). The positive ΔS value (23.4 J/mol K) implied that the randomness increased at the solution/solid interface during the adsorption process (Zhang et al. 2017).

To determine the reusability of the used NA11, its regeneration (using sodium hydroxide at different concentrations) and re-adsorption were investigated. It could be seen that the amounts of phosphate ions desorbed increased with increasing alkalinity. The phosphate ion desorption percentages were 0.5% (amount adsorbed and amount desorbed is 93 mg/g and 0.4 mg/g, respectively), 9.4% (amount adsorbed and amount desorbed is 93 mg/g and 8.7 mg/g, respectively), 25.3% (amount adsorbed and amount desorbed is 93 mg/g and 23.5 mg/g, respectively), and 74.2% (amount adsorbed and amount desorbed is 93 mg/g and 69.0 mg/g, respectively) for 1, 10, 100, and 1,000 mmol/L sodium hydroxide solutions, respectively. These results suggested that the bonds between the active sites and the adsorbed phosphate ions are breakable, and the phosphate ions adsorbed into NA11 can be exchanged by hydroxide under high-pH conditions; thus, NA11 has the potential to be used as a renewable adsorbent (Maor et al. 2011). These adsorption–desorption cycles using 1,000 mmol/L sodium hydroxide solution were carried out up to three times; the results are presented in Figure 8. The adsorption percentages of phosphate ions slowly decreased with an increase in the number of regeneration cycles (the total amount adsorbed, total amount desorbed, and total recovery percentage are 225 mg/g, 197 mg/g, and 88%, respectively). The adsorption loss might be due to the extraction of effective constitute (nickel and aluminum) in NA11 and the incomplete phosphate ion desorption using alkaline solution as an eluent, especially for inner phosphate ion complexes. Nevertheless, the regenerated NA11 was still cost-effective for phosphate ion adsorption and could be reused at least three times. Thus, the adsorption of phosphate ions onto NA11 is relatively reversible and it appears that NA11 can be regenerated via sodium hydroxide solution treatment (Maor et al. 2011; Liu et al. 2012; Li et al. 2014).

Adsorbents (nickel-aluminum complex hydroxides, NA12, 11, 21, 31, and 41) were prepared in order to investigate their phosphate ion adsorption capacity. NA11 had an amorphous structure with a large specific surface area (22.8 m²/g) and a high concentration of hydroxyl groups (1.92 mmol/g). We elucidated that the adsorption mechanism of phosphate ions was related to the above-mentioned properties. In addition, we measured the elemental composition and the binding energy before and after adsorption, which supported the hypothesis that adsorbent surfaces are an important factor for the adsorption of phosphate ions. Ion exchange with the sulfate ions in NA11 was also involved in the adsorption of phosphate ions (r = 0.837). The amount of phosphate ions adsorbed increased with increasing temperature, and these data were fitted to the Freundlich equation model (r = 0.857–0.987). In addition, thermodynamic and kinetic experiments demonstrated that the phosphate ion adsorption was a spontaneous, endothermic chemisorption. The phosphate-loaded NA11 could be effectively regenerated by sodium hydroxide solution and reused at least three times for phosphate ion adsorption. Thus, NA11 appears to be a promising candidate for the adsorption of phosphate ions from aqueous solution.

MEXT (Ministry of Education, Culture, Sports, Science and Technology)-supported Program for the Strategic Research Foundation at Private Universities, 2014–2018 (S1411037).