Adsorption of Ni(II) by a thermo-sensitive colloid: methylcellulose/calcium alginate beads

Heavy metals are defined as metal elements with a relative atomic mass between 63.5 and 200.6 and a density greater than 5.0 kg·dm⁻³. Nickel is one of the most common heavy metals in industrial effluents and is associated with toxicity problems. Heavy metals are well known to be harmful to most organisms when they are present in excessive concentrations. Hence, effective removal of toxic heavy metal ions from aqueous systems is of great importance for the protection of natural ecosystems. Many technologies have been used for removing Ni(II) from water, including chemical precipitation, membrane filtration, ion exchange, activated carbon adsorption and electrolysis. However, these conventional technologies have disadvantages, such as secondary pollution, short service life, low treatment capacity and high cost, and they are highly energy consuming (Zhou et al. 2017). Therefore, techniques that are low-cost, economic, effective, pollution-free and with high adsorption capacity are being widely studied and applied in the removal of Ni(II) from wastewater. Especially, biosorption is a viable alternative, with low cost, high efficiency, and minimization of waste generated during wastewater treatment (Volesky 1987).

Adsorbent selection, which determines the adsorption effect on the adsorbate, is very important in the application process of adsorption technology. At present, various adsorbents have been reported to remove heavy metal ions from industrial or agricultural wastewater, for example, active carbon (Depci et al. 2012), resins (Zhao et al. 2018), clays (Kara et al. 2017) and algae (Moino et al. 2017). In particular, algae biosorbents have attracted increased attention because they involve lower investments in terms of both initial and running cost, simple design, easy operation and no effect of toxic substances on wastewater treatment.

Algae are divided into several absolutely independent evolutionary pathways: a ‘red pathway’ with red algae (Rhodophyta), a ‘brown pathway’ with brown algae (inter alia, Chromophyta) and a ‘green pathway’ that includes green algae (Chlorophyta) along with mosses, ferns and several plants (Romera et al. 2007). The brown variety has been studied most as a biosorbent because their cell walls contain biopolymer alginate, which is responsible for the mechanical strength and high adsorption capacity when compared with other algae. Sodium alginate (SA) is a renewable, anionic, linear natural polysaccharide and a major component of the cell walls of brown algae, and consists of the 1,4-linked C5 epimers, β-d-mannuronic acid (M or ManA) and α-l-guluronic acid (G or GulA) found in homo or mixed blocks (Stender et al. 2018). Furthermore, SA has high reactive activities and can be easily modified by ionotropic gelation. For example, a low-cost and convenient way of crosslinking is immersing SA membranes into CaCl₂ solution for several minutes, thus the coupled Ca²⁺ with oxygen molecule of the G sequences in the polymer chain build the ‘egg-box structure’, and the formation of insoluble calcium alginate follows (Xu et al. 2018). Insoluble alginate beads are widely used as biosorbents to adsorb various heavy metals through their carboxylic groups, and are convenient for separating the beads from wastewater after adsorption. Calcium alginate beads, one kind of alginate gels, have been reported in the literature to remove heavy metals. For instance, Hong et al. (2016) reported that calcium alginate beads showed an excellent adsorption performance for Sr⁺; Ren et al. (2016) reported that the calcium alginate beads modified by adding carboxymethyl cellulose could increase the adsorption capacity for Pb²⁺.

In addition, calcium alginate beads can be better removed from the treatment system after adsorption than SA when used as an adsorbent to adsorb dyes or heavy metal ions. However, there is not much research on the regeneration of calcium alginate except a few reports of using chemical reagents such as acids and salts (Hong et al. 2016). Therefore, SA is modified by polymer blending in this study; that is to say, the functional material methylcellulose (MC) is added to the sodium alginate aqueous solution, and the obtained mixture is added to CaCl₂ solution for calcification. MC is a natural carbohydrate polymer and freely soluble in water. It forms aqueous solutions and demonstrates a unique property to form reversible physical gels due to hydrophobic interactions when heated above a particular temperature (Ren et al. 2016). In theory, adding MC to SA can increase the adsorption capacity of colloid for Ni(II) from water and boost the difference of adsorption capacity between low and high temperatures. In other words, MC may be used in adsorbing, desorbing, and recovering Ni(II) by controlling the temperature. A thermo-sensitive colloid composed of calcium alginate and methylcellulose has been prepared in our lab, and it has been found that the methylcellulose/calcium alginate beads (CAMCBs) had a high capacity for adsorption of cationic dyes from wastewater (Li et al. 2016). Based on the results of adsorption of cationic dyes by CAMCBs and the characteristics of SA and MC, the prepared CAMCBs might also have a good adsorption ability for Ni(II) in water.

The aim of this paper was to provide the fundamental knowledge for adsorption of Ni(II) by the thermo-sensitive CAMCBs. The effects of various operating parameters have been investigated in detail. The process of Ni(II) adsorption by CAMCBs was investigated by kinetics, isotherms and thermodynamic methods. Furthermore, depending on the nature of the adsorbent and adsorbate, and also in order to maintain the structure of the adsorbent, CaCl₂ was chosen as the desorption agent to regenerate the used CAMCBs (Li et al. 2019), and the effect of CaCl₂ concentration on the desorption of Ni(II) was studied. In addition, the effects of regeneration times on the adsorption capacity and on the mechanical strength of CAMCBs were discussed as well.

Sodium alginate and KI were purchased from Xilong Chemical Co. Ltd (Guangdong, China). Methylcellulose and anhydrous CaCl₂ were purchased from Sinopharm Chemical Reagent Co. Ltd (Shanghai, China). HCl, NaOH, H₁₂N₂NiO₁₂ and ammonium hydroxide were purchased from Guanghua Science and Technology Co. Ltd (Guangdong, China). C₄H₇N₃O₇ and I₂ were purchased from Bodi Chemical Co. Ltd (Tianjin, China). C₄H₈N₂O₂ was purchased from East China normal university Chemical Plant (Shanghai, China). EDTA-Na₂ was purchased from Coolaber Science and Technology Co. Ltd (Beijing, China). All chemicals used in the study were of analytical reagent grade.

First, 0.1 g of MC and 0.1 g of SA were added to 10 mL of deionized water at 30 °C, and then the mixed solution was placed in the ultrasound wave to remove the bubbles. After that, the formed MC/SA solution was added dropwise into 40 mL of CaCl₂ solution (5 wt%) under the stirring speed of 200 rpm; it took about 10 minutes for all MC/SA mixture to be dropped into CaCl₂ solution. Then the drops of MC/SA solution were calcified in the CaCl₂ solution. The calcified reaction was allowed to proceed for 4 h to form CAMCBs. Finally, the CAMCBs were separated from the solution and washed several times with deionized water, and then the CAMCBs were kept in deionized water to maintain their structure. The diameter of the prepared CAMCBs was about 2.0 mm.

The adsorbents before and after adsorption were characterized by scanning electron microscopy (SEM) and Fourier transform infrared spectroscopy (FT-IR). SEM micrographs of the adsorbents were recorded with a Hitachi S-3400N scanning electron microscope. FT-IR analysis was collected on a 1752X FT-IR spectroscopy using a KBr pellet in the range of 4,000–500 cm⁻¹ with 4 cm⁻¹ resolution.

Moreover, the mechanical strength of CAMCBs (F, N) was evaluated by the value of bearing force of CAMCBs which was recorded when the bead was pressed to a thickness of 0.8 mm in a hardness-testing device (Li et al. 2019).

All experiments were conducted in duplicates, and the mean results were obtained from these data. Linear regression was used to evaluate the fitness of the prediction models to the experimental data in this study using Origin 9.1. The suitability of the model fitting was assessed using R²  values.

The adsorbent dosage is an important parameter in adsorption studies because it determines the capacity of adsorbent for a given initial concentration of Ni(II) solution. The effect of CAMCBs dosage on the adsorption of Ni(II) is shown in Figure 1(a). As seen from Figure 1(a), the adsorption capacity of Ni(II) decreased with the increase of CAMCBs dosage. A lower adsorbent dosage meant that a smaller overall total surface area of CAMCBs was exposed, and hence, more Ni(II) was adsorbed on the surface per gram unit of CAMCBs, which led to the higher adsorption capacity (Jalil et al. 2010). However, it was observed that the adsorption ratio initially increased rapidly and then almost reached a constant value after the CAMCBs dosage reached 2 g·L⁻¹. The adsorption ratio of Ni(II) increased from 55.1% to 69.6% with the dosage increase from 1 to 5 g·L⁻¹. The increase in adsorption ratio of Ni(II) was attributed to more available adsorption surface and more adsorption sites provided by added CAMCBs (Chen et al. 2010). Compared with that at the dosage of 2 g·L⁻¹, the adsorption ratio of Ni(II) showed no significant change when the dosage was over 2 g·L⁻¹. Therefore, following comprehensive consideration of changes in adsorption capacity and adsorption ratio with increasing dosage, 2 g·L⁻¹ of CAMCBs was chosen for later studies on the adsorption of Ni(II).

Temperature and time are both important parameters for the adsorption process, and the effects of temperature and time on the adsorption of Ni(II) on CAMCBs were investigated. As shown in Figure 1(b), the adsorption capacity of CAMCBs decreased with the increase of temperature in the range of 30 °C to 60 °C, which manifested that higher temperature was not conducive to the adsorption of Ni(II) on CAMCBs. The higher temperature significantly activated Ni(II) in the solution, which made it more difficult for Ni(II) to be adsorbed on the CAMCBs surface. Moreover, the higher temperature increased macromolecule cross-linking in the fiber, which narrowed the amorphous region and reduced the amount of Ni(II) that entered the fiber (Ren et al. 2016).

Furthermore, as shown in Figure 1(b) the adsorption capacity of Ni(II) on CAMCBs increased with the elapsed adsorption time and reached equilibrium as the adsorption time exceeded 60 min. This outcome suggested that the adsorption equilibrium time is approximately 60 min. It should be emphasized that the equilibrium time of this type of adsorbent was shorter than those reported in the literature (Xu et al. 2016; Liao et al. 2016).

The effects of initial concentration of Ni(II) and adsorption time on the adsorption of Ni(II) were investigated. As shown in Figure 1(c), the maximum adsorption capacity of CAMCBs increased with the increase of initial Ni(II) concentration in the range of 20 to 140 mg·L⁻¹. When the initial Ni(II) concentration was higher, the adsorption rate increased quickly at first, and then went down slowly until the adsorption equilibrium. This was due to the fact that higher initial concentration of adsorbate provided more driving power of adsorption of Ni(II) on CAMCBs, which made the adsorbate easier to capture (Liao et al. 2016).

where q_e (mg·g⁻¹) is the adsorption amount of Ni(II) on CAMCBs at equilibrium (values of calculated q_e were obtained from Equation (7)), and q_t (mg·g⁻¹) is the adsorption amount of Ni(II) on CAMCBs at time t. k₁ (min⁻¹) is the ratio constant of pseudo-first-order mode, and k₂ (g·mg⁻¹·min⁻¹) is the constant of pseudo-second-order mode. C_S (g·L⁻¹) is the concentration of adsorbent. α and k₀ are the constant.

The pseudo-first-order equation, pseudo-second-order equation and Bangham equation were applied to fit the experimental data. Plots of pseudo-first-order, pseudo-second-order, and Bangham kinetic models under different temperatures and different Ni(II) initial concentrations are presented in Figures 2 and 3, respectively. Meanwhile, on the basis of the intercept and slope of the linear fitting, these kinetic model parameters could be calculated. These data are summarized in Tables 1 and 2, respectively. It can be found that the linear regression correlation coefficient values (R²) with the pseudo-second-order model were higher (R² > 0.995 for all temperatures and initial concentrations) than the pseudo-first-order model, suggesting that the experimental data fitted better with the pseudo-second-order kinetic model than the pseudo-first-order model. A similar phenomenon has been observed in the adsorption of Ni(II) on chitosan (Liao et al. 2016) and novel silica-based hybrid adsorbents (Xu et al. 2016). In addition, comparison of q_e values obtained via adsorption experiments with that obtained by linear fitting of the pseudo-second-order model showed that these values were similar. This result confirmed that the pseudo-second-order kinetic model can be used to describe the adsorption behavior of Ni(II) on CAMCBs. Moreover, plots of Figures 2(c) and 3(c) were found to be linear both for different temperatures and different Ni(II) initial concentrations, which confirmed that pore-diffusion was involved in Ni(II) adsorption onto CAMCBs (Namasivayam & Sangeetha 2006).

Plots and calculated parameters of the intra-particle diffusion model under different temperatures are shown in Figure 4 and Table 3, respectively. According to Figure 4, the plots were straight lines that didn't intersect the origin. This finding demonstrated that the adsorption of Ni(II) onto CAMCBs was not solely controlled by the intra-particle diffusion and that membrane diffusion processes might be involved (Vimonses et al. 2009), which was consistent with the conclusion obtained by the Bangham model. The entire adsorption process was divided into three phases. The first phase was the surface diffusion process (membrane diffusion), that is, the adsorbate diffused to the surface of the adsorbent. The second phase was the intragranular diffusion process, namely the diffusion of the adsorbate in the pores of the adsorbent. The last phase was the adsorption equilibrium stage. In addition, the diffusion rate constants k₂ and k₃ are much smaller than k₁, demonstrating that the surface diffusion proceeded much faster. When the surface of the CAMCBs reached adsorption saturation, Ni(II) entered the internal pores of the adsorbent particles. And then the diffusion resistance gradually increased, resulting in a decrease in diffusion rate and finally reaching adsorption equilibrium (Zhang et al. 2019).

Equilibrium data, commonly known as adsorption isotherms, describe the relationship between the amount adsorbed by a unit weight of adsorbent and the amount of adsorbate remaining in a test medium at equilibrium. Two kinds of adsorption isotherms, namely Langmuir and Freundlich models, were used to analyze the adsorption data of Ni(II) adsorbed by CAMCBs. The parameters obtained from the different models provide important information on the surface properties of the adsorbent and the affinity of adsorbent to the adsorbate.

The experiment data of Ni(II) adsorbed by CAMCBs under a temperature of 30 °C are presented in Figure 5(a). The experiment data were fitted by linear plots of Langmuir and Freundlich adsorption isotherm models as C_e/q_e versus C_e and log q_e versus log C_e, respectively (Figure 5(b) and 5(c)). Table 3 reports the various parameters obtained from the four isotherm plots mentioned above. As shown in Table 3, the R² values of Langmuir and Freundlich isotherm models are greater than 0.9, which manifested that both isotherm models fitted very well (Dada et al. 2012). The Langmuir isotherm model gave a higher R² value (0.9991) than the Freundlich isotherm model, which was evidence that the adsorption of Ni(II) on CAMCBs was better fitted by the isotherm model of Langmuir (see Table 4). The applicability of the Langmuir isotherm model demonstrated that the adsorption of Ni(II) on CAMCBs was monolayer adsorption and adsorption occurred at specific homogeneous sites on the adsorbent (Chen et al. 2010). In addition, the R_L value (0.3782) of the Langmuir isotherm model was between 0 and 1, and the n value (2.2634) of the Freundlich isotherm model was greater than 1, both indicating that the adsorption process of Ni(II) by CAMCBs was favorable (Dada et al. 2012; Toor & Jin 2012).

Furthermore, comparisons of Ni(II) adsorption by different absorbents are given in Table 5. It can be seen that the optimum adsorption temperatures of these adsorbents for adsorbing Ni(II) are between 20 and 30 °C, which are relatively close to room temperature. All the adsorption behaviors of these adsorbents are consistent with Langmuir. Therefore, the q_m values obtained by linear fitting of Langmuir are compared, and the result reveals that the adsorption capacity of CAMCBs has the best performance.

As shown in Figure 6, the plot of lnK as a function of 1/T yields a straight line. ΔH and ΔS were calculated from the slope and the intercept of the plot, respectively. The thermodynamic parameters at various temperatures are presented in Table 6. The ΔG values were negative over the temperature range 303.15–323.15 K, confirming the feasibility of the process and the spontaneous nature of Ni(II) adsorption on CAMCBs. When the temperature decreased from 323.15 to 303.15 K, the magnitude of ΔG shifted to a high negative value (from −0.154 to −1.663 kJ·mol⁻¹), suggesting that the adsorption of Ni(II) on CAMCBs was rapid and more spontaneous at low temperature (Kara et al. 2003). However, at 333.15 K, the value of ΔG was positive, which showed that the adsorption was no longer a spontaneous one and that the adsorption system gained energy from an external source (Jalil et al. 2010). The negative value of ΔH manifested that the adsorption of Ni(II) on CAMCBs was exothermic (Safa Özcan et al. 2005). This was in agreement with the conclusion that the adsorption capacity increased with the decrease of temperature, which was observed in the section above, titled ‘Effects of adsorption temperature and adsorption time on adsorption of Ni(II)’. Furthermore, the heat of adsorption was lower than 40 kJ·mol⁻¹, indicating that the adsorption of Ni(II) on CAMCBs was physical in origin (Kara et al. 2003). In addition, the ΔS value was negative, which demonstrated the decreased randomness at the solid–solution interface during the adsorption of Ni(II) (Kong et al. 2017), and the adsorption was a process of entropy reduction.

From the results of Ni(II) adsorption on CAMCBs, it could be concluded that high temperature was favorable for the desorption of Ni(II) in the experimental temperature range from 30 to 60 °C. Besides, CAMCBs was thermo-sensitive, which means it had better desorption at high temperatures than at low temperatures. Therefore, the regeneration of CAMCBs, via the desorption of Ni(II) by using CaCl₂ as desorption agent, was investigated at the temperature of 60 °C. In order to explore the effect of CaCl₂ concentration on the desorption of Ni(II) on CAMCBs, the CAMCBs after reaching adsorption equilibrium were placed into 20 ml of CaCl₂ solution. As shown in Figure 7, the desorption ratio of Ni(II) increased by 21.2% when the CaCl₂ concentration increased from 1 to 3 g·L⁻¹; however, it only increased by 3.8% when the concentration of CaCl₂ increased from 3 to 6 g·L⁻¹. Hence, 3 g·L⁻¹ was chosen as the concentration of CaCl₂ to be used in the subsequent experiments of regeneration. Besides, as shown in Figure 8(a), compared with the adsorption of initial CAMCBs, the adsorption capacity of regenerated CAMCBs in the subsequent four phases showed no obvious change, demonstrating that the CAMCBs had a good regeneration performance. In addition, as shown in Figure 8(b), the mechanical strength of CAMCBs decreased slightly after experiencing five regeneration cycles, indicating that CaCl₂ as a desorption agent showed no significant effect on the strength of the CAMCBs structure.

SEM images of CAMCBs before adsorption and after adsorption of Ni(II) are shown in Figure 9. As seen from Figure 9(a), CAMCBs before adsorption had a smooth surface, with obvious fissures and pores. However, after adsorption, the SEM image of CAMCBs exhibited a rough surface, with almost no fissures and pores. The differences of surface morphology between CAMCBs before and after adsorption were mainly due to the covering and filling by the adsorbed Ni(II) on the surface of CAMCBs. In addition, as shown in Figure 9(b), the SEM of CAMCBs after adsorption showed some white particles on the surface, which were likely to be nickel (Jiang et al. 2010).

The FT-IR spectra of CAMCBs before adsorption and after adsorption of Ni(II) are shown in Figure 10. The FT-IR spectra of CAMCBs before and after adsorption showed absorption bands at around 3,448 cm⁻¹ (-OH stretching vibration), around 1,636 cm⁻¹ (asymmetric -COO⁻ stretching vibration), and around 1,096 cm⁻¹ (C-OH stretching vibration). However, comparing the FT-IR spectra of CAMCBs before and after adsorption, one interesting thing observed was that the peaks of CAMCBs after adsorption shifted slightly to the right, and the intensity of the peaks also increased significantly. The differences of FT-IR spectra between CAMCBs before and after adsorption indicated that -OH, -COO⁻ and C-OH acted as adsorption sites for Ni(II) on the adsorbent, and Ni(II) was successfully adsorbed on the CAMCBs (Vijaya et al. 2008).

CAMCBs was found to be an effective adsorbent in the adsorption of Ni(II) from aqueous solution, and the maximum adsorption capacity occurred at the following operation conditions: 2 g·L⁻¹ adsorbent, 140 mg·L⁻¹ initial Ni(II) solution, temperature 30 °C, and adsorption time 2 h. It was observed that the pseudo-second-order kinetic model was a better fit for the experiment data than the pseudo-first-order kinetic model. The intra-particle diffusion equation was analyzed, indicating that the adsorption process might be controlled by both membrane diffusion and intra-particle diffusion. The best-fit adsorption isotherm was achieved with the Langmuir model, indicating that the behavior of the Ni(II) adsorption process on CAMCBs was monolayer adsorption and that adsorption occurred at specific homogeneous sites on the adsorbent. Furthermore, according to the Langmuir model, the adsorbent's greatest adsorption capacity was predicted to be 28.17 mg·g⁻¹ for Ni(II). The negative ΔH for the adsorption process confirmed the exothermic nature of the adsorption, and the negative value of ΔG showed the feasibility and spontaneity of Ni(II) adsorption on CAMCBs at 30 to 50 °C. The regeneration results demonstrated the excellent regeneration performance of CAMCBs. Hence, from the foregoing results, it can be safely concluded that CAMCBs is an effective potential adsorbent for removing Ni(II) from wastewater.

Financial support from National Natural Science Foundation of China (21667004) and Guangxi Science Foundation Funded Project (2018GXNSFAA281343, 2018GXNSFAA138039), and Ministry-province Jointly-constructed Cultivation Base for State Key Laboratory of Processing for Non-ferrous Metal and Featured Materials in Guangxi Zhuang Autonomous Region (GXKFJ16-06, GXKFJ16-04) is gratefully acknowledged.