Adsorption/Desorption Capability of Potassium-Type Zeolite Prepared from Coal Fly Ash for Removing of Hg 2+

: The feasibility of using potassium-type zeolite (K-type zeolite) prepared from coal ﬂy ash (CFA) for the removal of Hg 2+ from aqueous media and the adsorption/desorption capabilities of various potassium-type zeolites were assessed in this study. Potassium-type zeolite samples were synthesized by hydrothermal treatment of CFA at different intervals (designated CFA, FA1, FA3, FA6, FA12, FA24, and FA48, based on the hours of treatment) using potassium hydroxide solution, and their physicochemical characteristics were evaluated. Additionally, the quantity of Hg 2+ adsorbed was in the order CFA, FA1 < FA3 < FA6 < FA12 < FA24 < FA48, in the current experimental design. Therefore, the hydrothermal treatment time is important to enhance the adsorption capability of K-type zeolite. Moreover, the effects of pH, temperature, contact time, and coexistence on the adsorption of Hg 2+ were elucidated. In addition, Hg 2+ adsorption mechanism using FA48 was demonstrated. Our results indicated that Hg 2+ was exchanged with K + in the interlayer of FA48 (correlation coefﬁcient = 0.946). Finally, adsorbed Hg 2+ onto FA48 could be desorbed using a sodium hydroxide solution (desorption percentage was approximately 70%). Our results revealed that FA48 could be a potential adsorbent for the removal of Hg 2+ from aqueous media.


Introduction
The 2030 agenda for sustainable development, such as clean water and sanitation (Goal 6) and life below water (Goal 14), were adopted by all member states of the United Nations in 2015 [1], to establish a sustainable society, which is a matter of global concern. In particular, heavy metal pollution has become a severe global environmental issue, including in the developing countries. Among them, mercury (Hg 2+ ), lead (Pb 2+ ), and cadmium (Cd 2+ ) are referred to as the "big three" heavy metals with the greatest potential risk to human health and water environment [2][3][4]. They are highly toxic to organisms [5]. Mercury (Hg) and its compounds can cause serious threats to organisms, including humans, because of their bioaccumulative properties, damaging the bones, liver, kidney, and nervous system [6][7][8]. The Minamata Convention on Mercury was adopted by the Intergovernmental Negotiating Committee in 2017. The International Agency for Research on Cancer categorizes methylmercury compounds as group 2B (possibly carcinogenic to humans), and metallic Hg and inorganic Hg compounds as group 3 (unclassifiable as to carcinogenicity in humans) [9]. In addition, the maximum permissible limit of Hg in drinking water as recommended by the U.S. Environmental Protection Agency and many
We had previously reported the physicochemical characteristics of K-type zeolites [28]. X-ray diffraction (XRD) and morphology analyses were performed using MiniFlex II (Rigaku, Osaka, Japan) and SU1510 (Hitachi High-Technologies Co., Tokyo, Japan), respectively. The cation exchange capacity (CEC) and pH pzc were measured using the Japanese Industrial Standard Method (JIS K 1478) and the method previously reported by Faria et al. [31]. Additionally, the specific surface area and pore volume were measured using NOVA4200e (Quantachrome Instruments Japan G.K., Tokyo, Japan). The binding energy was measured using a JXA-8530F (JEOL Ltd., Tokyo, Japan). Finally, the solution pH was measured using an F-73S digital pH meter (HORIBA, Ltd., Kyoto, Japan).

Amount of Hg 2+ Adsorbed Using FA Series
Approximately 0.01 g of each pretreated adsorbent, namely CFA, FA1, FA3, FA6, FA12, FA24, and FA48, was mixed with 50 mL of 50 mg/L Hg 2+ solution. Subsequently, the reaction mixture was shaken at 100 rpm and 25 • C for 24 h. The resulting sample was filtered through a 0.45 µm membrane filter. The concentration of Hg 2+ was measured using an inductively coupled plasma optical emission spectrometer (iCAP-7600 Duo, Thermo Fisher Scientific Inc., Osaka, Japan). The quantity of Hg 2+ adsorbed was calculated using the levels before and after adsorption in Equation (1).
where q is the quantity adsorbed (mg/g); C 0 is the initial concentration (mg/L); C e is the equilibrium concentration (mg/L); V is the solvent volume (L); and W is the weight of the adsorbent (g).

Effect of pH, Temperature, and Contact Time on the Adsorption of Hg 2+
First, in order to evaluate the effect of pH, FA48 (0.01 g) was added to 50 mL of the Hg 2+ solution at 10, 30, 50 mg/L. The pH of the solution was adjusted to 2, 5, 7, 9, 11 using either nitric acid or sodium hydroxide solutions. The suspension was shaken at 100 rpm at 25 • C for 24 h, and filtered using a 0.45 µm membrane filter. Second, in order to evaluate the temperature effect, FA48 (0.01 g) was added to a 50 mL Hg 2+ solution at 10, 20, 30, 40, 50 mg/L, and the suspension was shaken at 100 rpm at 7, 25, 45 • C for 24 h. The 7 • C of solution was prepared as follows. The sample solution was set at 5 • C in a water bath shaker personal-11 (TAITEC Co., Nagoya, Japan) in the low-temperature room at 6 • C. Finally, to evaluate the effect of contact time, FA48 (0.01 g) was added to 50 mg/L Hg 2+ solution (50 mL). The suspension was shaken at 100 rpm and 25 • C for 0.5, 1, 3, 6, 12, 18, 21, 24, 30, 42, 48 h. The amount of Hg 2+ adsorbed was calculated as described in Section 2.2. In addition, to evaluate the Hg 2+ adsorption mechanism, the concentration of potassium ions released from FA48 in the adsorption isotherm experiment was measured using an iCAP-7600 Duo (Thermo Fisher Scientific Inc., Osaka, Japan).

Effect of Coexistences on the Adsorption of Hg 2+
In order to evaluate the selectivity of Hg 2+ adsorption, FA48 (0.01 g) was added to the binary solution of 50 mL. The two components were Hg 2+ and Na + , Mg 2+ , K + , Ca 2+ , Ni 2+ , Cu 2+ , Zn 2+ , Sr 2+ , or Cd 2+ , and Hg 2+ or individual cation concentration was 10 mg/L in a binary solution. The sample solution was shaken at 100 rpm at 25 • C for 24 h and filtered through a 0.45 µm membrane filter. The concentration of each metal was measured using an iCAP-7600 Duo. The amount adsorbed was calculated by comparing the levels before and after adsorption.

Adsorption/Desorption of Hg 2+ Using Sodium Hydroxide Solution
To evaluate the recycling of FA48 in Hg 2+ adsorption/desorption, FA48 (0.15 g) was added to a 150 mL Hg 2+ solution at 250 mg/L. The suspension was shaken at 100 rpm, 25 • C for 24 h, and filtered through a 0.45 µm membrane filter. The concentration of Hg 2+ was measured using an iCAP-7600 Duo. The amount of Hg 2+ adsorbed was calculated as described in Section 2.2. After adsorption, FA48 was collected, dried, and used for the desorption experiment. The collected FA48 (0.05 g) was added to 50 mL sodium hydroxide solution at 10, 100, 1000 mmol/L. The suspension was shaken at 100 rpm, 25 • C for 24 h, and filtered through a 0.45 µm membrane filter. The concentration of Hg 2+ released from FA48 was also measured using an iCAP-7600 Duo. The amount of Hg 2+ desorbed was calculated using the levels before and after desorption. All results in this study are expressed as mean ± standard error (n = 2-3, Sections 2.2-2.5). In addition, each Figure was prepared using Microsoft Excel.

Properties of Potassium-Type Zeolite
Zeolites are characterized by physicochemical properties, such as specific surface area, pore volume, and CEC (Table 1). These characteristics are related to the parameters of hydrothermal treatment, such as heat temperature, pressure, solution alkalinity, activation solution to CFA ratio, and formation process [18]. In this study, six types of potassium-type zeolites were prepared using the above-mentioned method [28]. In addition, our previous study reported the physicochemical properties of potassium-type zeolites in detail [28]. The XRD patterns indicate that CFA was mainly composed of mullite and quartz. The XRD patterns of FA1, FA3, FA6 and FA12 were similar to those of CFA under our experimental conditions. Zeolite F appeared in FA24 and FA48 structures. We observed changes in the surface of FA series with the treatment time. Aluminosilicate gels were clearly produced on FA24 and FA48 surfaces. These processes were in the following order: Al and Si dissolution, geopolymer formation, crystalline structure nucleation, finally zeolite crystal growth [18]. The CEC of FA48 (8.98-11.77 mmol/g) was the highest compared to other FA series. This value of FA48 was 26-69 times higher than that of CFA. The pH pzc of FA was 9.8, and that of FA24 and FA48 were 10.4. Finally, specific surface area and pore volume (d 20 Å) of FA48 (47.3 m 2 /g and 10 Å) was 34 and 100 times higher than CFA. Additionally, the value of FA48 (potassium-type zeolite) was greater than that of sodium-type zeolite [32]. Thus, these results indicate that potassium-type zeolite (FA48) can be prepared from coal fly ash by conventional hydrothermal treatment using potassium hydroxide.  Figure 1 shows the quantity of Hg 2+ adsorbed by the FA series. The adsorbed Hg 2+ was in the order CFA, FA1 (0-0.48 mg/g) < FA3 (2.2 mg/g) < FA6 (3.5 mg/g) < FA12 (4.0 mg/g) < FA24 (7.5 mg/g) < FA48 (11.6 mg/g) in the current experiment conditions. The adsorption capability of Hg 2+ using the FA series depended on the duration of the hydrothermal activation treatment using potassium hydroxide solution. Next, we evaluated the relationship between the adsorption capacity of Hg 2+ and the physicochemical properties of the FA series. As a result, the positive correlation coefficient between the quantity of Hg 2+ adsorbed and CEC, specific surface area, and pore volume (d 20 Å) were 0.928, 0.659, and 0.882, respectively. These results indicate that CEC and pore volume strongly affect the adsorption of Hg 2+ from aqueous solutions. Additionally, in this study, FA48 was selected to evaluate the adsorption capability for Hg 2+ removal from aqueous solutions. A comparison of the Hg 2+ adsorption capability of FA48 with that of the other adsorbents is listed in Table 2 [24,26,[32][33][34][35]. FA48 exhibited potential in Hg 2+ adsorption from aqueous solutions compared to other reported adsorbents (except for coal gangue and multifunctional mesoporous material).  Figure 2 shows the adsorption isotherms of Hg 2+ using FA48 at different temperatures. The quantity of Hg 2+ adsorbed using FA48 did not significantly vary with different temperatures. Therefore, in this study, the adsorption temperature did not strongly affect the adsorption capability of FA48. Additionally, to investigate the adsorption properties and interactions, the adsorption isotherm data were evaluated using the Freundlich and Langmuir isotherm models. The Freundlich isotherm model was applied to multilayer adsorption, while the Langmuir isotherm model showed monolayer adsorption at specific homogenous sites [24]. The Freundlich isotherm model can be represented as follows [36]:

Adsorption of Hg 2+
where q is the quantity of Hg 2+ adsorbed (mg/g), K F and 1/n are the Freundlich isotherm constants, C is the equilibrium concentration (mg/L). In general, the adsorption reaction in the aqueous phase fits this model. In the Freundlich isotherm model, the isotherm curve depends on the value of n. In particular, when the value of 1/n is 0.1-0.5, adsorption occurs easily, when 1/n is over 2, it is difficult to adsorb [37]. The Langmuir isotherm model can be represented as follows [38]: where K L is the Langmuir isotherm constant (L/mg) and q max is the maximum quantity adsorbed (mg/g). The Langmuir isotherm model is a theoretical model that can explain monolayer adsorption onto homogenous surfaces. In addition, this model considers adsorption sites. Table 3 shows the Freundlich and Langmuir model constants for the adsorption of Hg 2+ using FA48. The obtained data fitted both models (correlation coefficient of the Freundlich and Langmuir equations were ≥ 0.960 and ≥ 0.904, respectively). The maximum quantity adsorbed at 7 to 45 • C was not significantly different in this study, which is supported by the adsorption isotherm data in Figure 2. In addition, the value of 1/n was from 0.27 to 0.33 in this study. Therefore, the adsorption of Hg 2+ using FA48 from aqueous solutions is more favorable. Finally, adsorption properties were evaluated using Sips equation (Equation (4)). The Sips model was derived from the Langmuir and Freundlich equations. This model predicts the heterogeneous adsorption system and overcoming the drawback associated with Freundlich model [39]. The Sips equation was expressed as follows: where K S is the Sips equilibrium constant (L/mg), Q max is the maximum quantity adsorbed (mg/g). n is the Sips model exponent, which can be employed to describe the system's heterogeneity. If the value of n is equal to 1, this equation will become a Langmuir equation. It means a homogeneous adsorption process [40,41]. Table 4 shows the Sips model constants for the adsorption of Hg 2+ . The value of correlation coefficient of Sips equation was from 0.841 to 0.959 under our experimental conditions. The values of Q max at 7-45 • C was not significantly changed, which is similar trends to the adsorption isotherm data (Figure 2). In addition, the heterogeneous factor values (n = 0.4-1.1) indicate that heterogeneous adsorption process is related to the adsorption mechanism of Hg 2+ using FA48. Moreover, to evaluate the adsorption mechanism of Hg 2+ using FA48, more detailed investigations were conducted in this study ( Figure 3). First, the relationship between the quantity of Hg 2+ adsorbed and the quantity of K + released from FA48 was evaluated in this study. As a result, the correlation coefficient value (r) was positive at 0.946, indicating that ion exchange with K + in the interlayer of FA48 was one of the mechanisms of Hg 2+ adsorption from aqueous media. As mentioned in Section 3.2, the positive correlation coefficient between the quantity of Hg 2+ adsorbed and the value of CEC was 0.928. These trends were similar to those reported in previous studies [28,32]. Additionally, the X-ray photoelectron spectroscopy analysis was conducted in this study. The peak intensity of Hg(5p) at 67 eV was newly detected after the adsorption of Hg 2+ , indicating that Hg 2+ was present on the FA48 surface after adsorption, and was not detected before adsorption. Generally, Hg(4f) peaks at 101 and 105 eV were detected after adsorption. However, Si(2p) and Hg(4f) peaks overlapped in this study. Therefore, it was difficult to elucidate and/or detect these peaks in our experiments.

Effect of Solution pH on the Adsorption of Hg 2+
In aqueous media, pH is one of the most important parameters for heavy metal removal. Thus, the solution pH strongly and directly affects the degree of metal ionization and/or metal binding on the adsorbent surface sites [42]. In this study, the quantity of Hg 2+ adsorbed onto FA24 increased with an increase in the solution pH from 2 to 5, and decreased with further increase in pH of 5 to 7 (Figure 4). First, when the solution pH is below 3, Hg 2+ is the dominant species; when the solution pH is over 5, Hg(OH) 2 is the dominant species [43,44]. In addition, Hg(OH) + exists (1-13% of the total mercury Hg 2+ ) when the solution pH is between 2 and 6. Previous studies have reported that there are several stable Hg 2+ products related to either the equilibrium hydrolysis (such as Hg(OH) 3 − , Hg(OH) 2 , Hg(OH) + ) or to the complexation equilibrium with chloride (such as HgCl 4 2− , HgCl 3 − , HgCl 2 , HgCl + ). Moreover, mixed species such as Hg(OH)Cl exist in aqueous media [44]. In an acidic solution (pH 2), the FA48 surface is protonated and the electrostatic repulsion between the FA48 surface (positive charge) and Hg 2+ species (positive charge) such as Hg(OH) + and HgCl + easily occurred, resulting in low Hg adsorption. Additionally, the pH pzc value of FA48 was 10.4 in this study, which supports the availability of positive charge on the FA48 surface and the low adsorption of Hg 2+ from aqueous media. Next, similar to pH 2, at pH 5, the quantities of Hg 2+ species such as Hg(OH) + and HgCl + decreased. Conversely, HgCl 3 species increased in aqueous media. Therefore, the quantity of Hg 2+ adsorbed increased because of the electrostatic interaction between the FA48 surface (positive charge) and Hg 2+ species such as HgCl 3 − (negative charge). Finally, when the solution pH was over 7, the quantity of Hg(OH) 2 increased, and the hydroxyl ion (OH − ) also increased in the sample solution media [6,43]. Therefore, FA48 showed a low adsorption capability for Hg 2+ under alkaline conditions.

Effect of Contact Time on the Adsorption of Hg 2+
To investigate the effect of contact time on the removal of Hg 2+ from aqueous media using FA48, the duration was varied from 0.5 to 48 h ( Figure 5). Rapid adsorption was observed within 0.5 h from the start of the adsorption process, following which the rate of adsorption of Hg 2+ fluctuated with increase in adsorption time. Finally, adsorption equilibrium was achieved at approximately 3 h under our experimental conditions. In this study, the adsorption might be mainly attributed to two factors: the interaction between Hg 2+ and active adsorption sites, such as specific surface area and pore volume (mentioned in Section 3.2), and ion exchange with K + in the interlayer of FA48 (mentioned in Section 3.3). In addition, to evaluate the kinetic adsorption mechanism of Hg 2+ using FA48, pseudofirst-order and pseudo-second-order models were selected to interpret the kinetics data using Equations (5) and (6) [36,38].
where q e,exp and q t are the quantities of Hg 2+ adsorbed at equilibrium and at time t (mg/g), respectively, q e,cal is the quantity of Hg 2+ adsorbed in the calculation (mg/g), k 1 (1/h) and k 2 (g/mg/h) are the rate constants of the pseudo-first-order and pseudo-second-order models, respectively. The calculated results are shown in Table 5. From Table 5, it is evident that the correlation coefficient (r) in the pseudo-second-order model (0.996) was significantly higher than the pseudo-first-order model (0.515), indicating that the pseudo-second-order model is more suitable for describing the adsorption kinetics of Hg 2+ in this study. Additionally, the value of q e,exp was closest to the value of q e,cal of the pseudo-second-order model than that of the pseudo-first-order model. In addition, it is strongly suggested that the adsorption of Hg 2+ onto FA48 is because of chemisorption, as assumed by this model [45,46].
In addition, the Elovich model (Equation (7)) was also used to describe adsorption kinetic in this study. This model describes activated adsorption, and predicts an energetically heterogeneous solid surface of adsorbent which means adsorption kinetics is not affected by interaction between the adsorbent particles [26]. q t =1/β ln (αβ) + 1/β ln t where q t is the quantity of Hg 2+ adsorbed at time t (mg/g), α is the initial adsorption rate (mg/g/h), β is the related to the extent of surface coverage and activation energy for chemisorption (g/mg). From the result, the value of α, β, and r (correlation coefficient) was 8.4 × 10 3 mg/g/h, 1.1 g/mg, and 0.888, respectively. The Elovich equation is suitable to describe adsorption behavior of Hg 2+ using FA48 that relates to the nature of chemical sorption under our experimental conditions [47].

Selectivity for Hg 2+ Removal from Binary Solution System
Considering the field application of FA48, the selectivity for Hg 2+ adsorption is one of the critical parameters in this study. Therefore, the effect of coexisting ions on the adsorption capability of Hg 2+ is shown in Table 6. In our study, Na + , Mg 2+ , K + , Ca 2+ , Ni 2+ , Cu 2+ , Zn 2+ , Sr 2+ , and Cd 2+ were used as the components of the binary solution system, as these ions are ubiquitous in the water environment [48,49]. In this study, the removal percentage of Hg 2+ using FA48 in a single solution system was approximately 14.0% whereas, the removal of Hg 2+ in the binary solution system was over 11.4% (except for Na + and K + ), and the removal of other cations was significantly lower. A similar trend was reported in a previous study [6]. In addition, previous studies reported that the radius of the hydrated ion and/or the electronegativity of the adsorbate (Hg 2+ in this study) strongly and directly influenced the adsorption capability in aqueous media [48,49]. Therefore, similar phenomena were observed under our experimental conditions. Finally, our results show that FA48 is useful for the selective removal of Hg 2+ from aqueous media. Moreover, the Minamata Convention on Mercury was adopted by the Intergovernmental Negotiating Committee in 2017. Therefore, the development of removal techniques for Hg 2+ in wastewater from anthropogenic activities such as the steel industry is very important for establishing a sustainable society. Thus, FA48 could be applied for wastewater purification including Hg 2+ such as the steel industry.

Adsorption/Desorption Capability of Hg 2+ Using FA48
Finally, the adsorption/desorption capability of Hg 2+ using FA48 was demonstrated in this study ( Figure 6). The quantity of Hg 2+ desorbed increased with increasing concentration of sodium hydroxide solution from 10 to 1000 mmol/L (the quantity of Hg 2+ adsorbed was approximately 40 mg/g). The desorption percentages using 10, 100, and 1000 mmol/L sodium hydroxide solutions were 37.5%, 41.6%, and 68.3%, respectively. Therefore, adsorbed Hg 2+ onto FA48 could be easily desorbed using a sodium hydroxide solution under our experimental conditions. Further investigations are needed to elucidate the application of FA48 in these fields.

Conclusions
Six types of potassium zeolites (FA1, FA3, FA6, FA12, FA24, FA48) were synthesized by hydrothermal treatment using a potassium hydroxide solution. The values of CEC, specific surface area, and pore volume (d 20 Å) of FA48 were 26-29, 34, and 100 times higher than that of CFA, indicating that FA48 has a high potential for the removal of Hg 2+ from aqueous media. The quantity of Hg 2+ was in the order CFA and FA1 (0-0.48 mg/g) < FA3 (2.2 mg/g) < FA6 (3.5 mg/g) < FA12 (4.0 mg/g) < FA24 (7.5 mg/g) < FA48 (11.6 mg/g) under our experiment conditions. These adsorption behaviors were affected by the CEC and/or pore volume (d ≤ 20 Å). In addition, the effects of pH, temperature, contact time, and coexistences on the adsorption of Hg 2+ using FA48 were demonstrated. The optimal pH was approximately 5.0. The adsorption isotherm data or kinetics data were described by the Freundlich and Langmuir models or the pseudo-second-order model, respectively. Moreover, one of the adsorption mechanisms determined was the ion exchange with K + in the interlayer of FA48 (correlation coefficient = 0.946). FA48 showed selectivity for the adsorption of Hg 2+ from a binary solution system containing Na + , Mg 2+ , K + , Ca 2+ , Ni 2+ , Cu 2+ , Zn 2+ , Sr 2+ , or Cd 2+ . Finally, adsorbed Hg 2+ onto FA48 was easily desorbed using a sodium hydroxide solution. It is evident that FA48 is a useful adsorbent for Hg 2+ removal from aqueous media. These techniques may potentially aid in mitigating heavy metal pollution and thus contribute to the establishment of a sustainable society.