Selective removal of silicic acid by a gallic-acid modi ﬁ ed resin

To remove silicic acid from aqueous solutions, a novel gallic acid-type resin (GA-type resin) was prepared by a grafting method. The effects of the adsorption capacity, pH and presence of NaCl, NaNO 3 , Na 2 SO 4 , and NaCO 3 salts on the silicic acid removal were studied. The GA-type resin adsorbs monosilicic acid, silicate ions, and polymeric silicic acid. The adsorption capacity of 4.64 – 4.94 mg/g was achieved in a short adsorption time ( Q m of 8.99 mg/g) and is 30 – 40 times larger than that of the OH-type resin. The silicic acid removal ef ﬁ ciency was almost unaffected by the pH and common anions when the common anion and silicic acid contents were similar, proving the GA-type resin exhibits an excellent performance for selective adsorption of silicic acid. The Temkin isotherm model can well describe the adsorption process, which is chemical adsorption, and indicates that the adsorption heat decreases with the increasing adsorption amount. The adsorption mechanism of silicic acid on the GA-type resin involves dehydration condensation reactions of the hydroxyl groups in silicic acid and gallic acid. The GA-type resin can be ef ﬁ ciently regenerated and reused after treatment with an HCl solution.


INTRODUCTION
Mineral scale formation in pipelines, wells, heat exchangers, and membranes is one of the main problems that hinders the wide utilization of water resource regeneration and seawater and brackish water desalination (Semiat et al. ; Zarrouk et al. ; Menzri et al. ). In particular, silica scale formation in many industrial processes can significantly shorten the lifetime of the water treatment equipment, increase the maintenance costs and energy consumption of an operation due to water flow or membrane blockages, and cause equipment corrosion, unscheduled shutdown time and system efficiency loss (Nishida et al. ; Milne et al. ). Once formed, the scales are very difficult to remove and can damage equipment over prolonged periods, resulting in a reduced water production capacity (Durham ). Therefore, silica scale mitigation is a challenge for all water utilization systems because of the complicated deposition mechanism, insolubility in common acids, and hard, dense structures of silica scales. The purpose of this study was to find a new adsorbent that selectively removes silicic acid from water to achieve high water resource regeneration without silica scaling.
Our past study has found that the gallic acid (which has three adjacent hydroxyl functional groups and one carboxyl group) interacts with monosilicic acid to form a complex in the water solution. Based on this result, we modified a strong-base anion-exchange resin with gallic acid by grafting method, and the resin can be used under a broad range of pH conditions for silicic acid removal from aqueous systems. Additionally, the adsorbent can be regenerated.
The silicic acid uptake capacity, adsorption isotherms, adsorption kinetics, adsorption mechanism, regeneration ability, and other characteristics of the resins were tested by corresponding experiments.

EXPERIMENTAL Materials and methods
All of the chemicals used were of analytical reagent grade.
The silicic acid stock solution was prepared by dissolving sodium metasilicate (Na 2 SiO 3 ·9H 2 O) in a sodium hydroxide solution (0.1 mol/L), and its concentration was determined by inductively coupled plasma atomic emission spectrometry. The gallic acid stock solution was prepared by dissolving solid gallic acid in solution, and the solution was standardized by a Shimadzu UV-3600 UV-Vis spectrophotometer under 260 nm. The strong-base anion-exchange resin (Amberlite IRA402, anion-exchange capacity: 2.0 mequiv/dry g) was purchased from Shanghai Macklin Biochemical Co., Ltd. All solutions were prepared with ultra-pure water (Milli-Q SP system, Millipore, resistance >18 MΩ). All experiments were performed at room temperature (25 ± 1 C). The total silicic acid concentration (Si(T)) in the filtrates was measured by inductively coupled plasma atomic emission spectrometry. In addition, the monosilicic acid concentration (Si(M)) in the filtrates was measured by a spectrophotometric method (Shimadzu UV-3600 UV-Vis) based on the formation of yellow molybdosilicic acid with wave length 400 nm. Here, Si(T) À Si(M) represents the concentration of polysilicic acid. In this paper, the concentration of the silicic acid solution is represented as Si mg/L.

Preparation of the GA-type resin
The strong-base anion-exchange resin (10 g) was added into the gallic acid solution (500 mL, 10 g/L, pH 7) and was stirred for 24 h. The resins that bound gallic acid were filtered with a qualitative paper filter, rinsed with water, and air dried at ambient temperature for 3 days and stored in a dryer. The concentration of gallic acid in the filtrate was determined by UV-Vis spectrophotometer under 260 nm. Then the amount of gallic acid on the resin was calculated by the difference in the gallic acid concentration between initial solution and the filtrate and was estimated to be 725.3 mg/g (dry).

Silicic acid uptake by the GA-type resin
The GA-type resin and OH-type anion-exchange resin (each after being sputtered with gold (Au). Images and data were then collected at accelerating voltage of 15 kV as presented.
Effect of coexisting anions on the uptake of silicic acid by the GA-type resin The GA-type resin (0.36 g) was added into silicic acid solutions (250 mL, 0.356 mmol/L, equal to 10 mg/L) with various concentration ratios of salts, including Na 2 CO 3 , Na 2 SO 4 , NaCl, and Na 2 NO 3 . The concentration ratios of silicic acid to the anions were 1:1, 1:10, and 1:100 (as Si: anion, mmol/L).

Adsorption isotherms and kinetics
Batch experiments were carried out by shaking 0.36 g of the GA-type resin in 250 mL solutions with different initial silicic acid concentrations (10-50 mg/L). The pH values of the solutions were adjusted and maintained at 8 ± 0.2. The silicic acid adsorption data were fitted to four two-parameter equations using nonlinear regression. Four isotherms, as described below in Equations (1)-(4), were used for fitting the results obtained from the adsorption experimental data (Langmuir, Freundlich, Temkin and Dubinin-Astakhov): where q is the amount adsorbed at equilibrium (mg/g), and C e is the equilibrium concentration of silicic acid in solution (mg/L). The other parameters are the different isotherm constants, which can be determined by a regression analysis of the experimental data. The curve fitting and statistical analyses were performed with Origin 8.5.
For the kinetic study, a nonlinear pseudo-second-order reaction model and an intra-particle diffusion model were used to fit the experimental data. The models are represented as follows in Equations (5) and (6), and the conformity between the experimental data and the model predicted values is expressed by the correlation coefficients (r 2 ).
where k 2 is the rate constants for the pseudo-second-order adsorption, k n is the intra-particle diffusion rate constant, and Q t and Q e are the amounts of silicic acid adsorbed at time t and at equilibrium, respectively.

Regeneration and reuse of the GA-type resin
To evaluate the possibility of regenerating and reusing the GA-type resin, sorption-regeneration cycles were performed with different eluents. Two GA-type resins, which had certain amounts of adsorbed silicic acid (silicic acid adsorption amount was 4.310 mg/g as Si), were eluted with HCl and NaOH solutions (0.005 mol/L) and stirred for 24 h.
The resins were filtered after washing with water and grafted again with a gallic acid solution (50 mL, 5 g /L) at pH 7 to prepare the regenerated resins. The regenerated resins were used as adsorbents to adsorb 10 mg/L silicic acid from solution.

RESULTS AND DISCUSSION
Silicic acid uptake by the GA-type resin In conclusion, the GA-type resin adsorbs monosilicic acid, silicate ions and the polymer form of silicic acid. The adsorption capacity was 30-40 times larger than that of the OH-type resin, which is commonly used as adsorbent to remove silicate ions from water. In addition, the adsorption capacity of the GA-type resin for silicic acid was independent of the pH and shows broad application prospects over a wide pH range for water treatment processes. In addition, we also investigated the effect of multiple anions on the adsorption amount using CO 3 2À , Cl À , SO 4 2À , and NO 3 À ions, and the anion to silicic acid (10 mg/L) molar ratio was 1:1. The adsorption amount decreased from 3.65 mg/g to 3.06 mg/g, which indicated the effect of multiple anions was larger than that of a single anion (not show here). Although the concentration ratio was the same as that used to examine the effect of a single anion

Effect of coexisting anions on silicic acid adsorption
(1:1), many kinds of ion species and different ion charges result in a different ionic strength and effect.

Adsorption isotherms and kinetics
The adsorption process can be understood through isotherms examining the adsorbate concentration in solution and the adsorbate amount adsorbed on a unit mass of the adsorbent at a constant temperature. Compared with the double-parameters of linear adsorption isotherm models, nonlinear models can more accurately describe adsorption processes. Therefore, to estimate the adsorption characteristics and mechanism of silicic acid on the GA-type resin, the experimental silicic acid adsorption data were fitted to Equations (1)-(4) using nonlinear regression.
The adsorption isotherms of silicic acid on the GA-type resin are shown in Figure 4. As seen in Figure 4, all four isotherm models well fit the experimental data. Among the models, the Temkin isotherm had the highest r 2 value 12 mg/g as SiO 2 ), which is higher than the silicic acid uptake by the TMAMo resin (13.82 mg/g as SiO 2 ), as shown in Table S1, Supplementary material (Alexandratos et al. ). The data in Table S1 (online   Table S1, which was consistent with the results in this study. Considering the complexity of the adsorption process, various simplified kinetic models have been reported in the literature, and each model has its own limitations due to simplified assumptions (Peydayesh & Rahbar-Kelishami ; Sadegh et al. ). In this study, to investigate the adsorption mechanism and potential rate controlling steps, a certain amount of the GA-type resin was contacted with three silicic acid solutions with different concentrations (10, 20, and 30 mg/L), and a pseudo-second-order reaction model and an intra-particle diffusion model (Equations (5) and (6)) were used to fit the experimental data.
The change in the adsorption capacity for different concentrations of silicic acid and the fitting of the kinetics models are shown in Figure 5. As seen in Figure 5(a), the pseudo-second-order model k values showed a decreasing trend as the initial concentration increased. The high r 2 values of the pseudo-second-order model indicate that the adsorption process is better fit by the pseudo-second-order rate mechanism. The suitability of the pseudo-second-order model suggests that chemical adsorption is the silicic acid adsorption mechanism on the GA-type resin.
The linearity of the curves in Figure 5

Adsorption mechanisms
According to the isotherm analysis, the adsorption of silicic acid on the GA-type resin occurs via chemisorption. As shown in Equation (7), the gallic acid molecule becomes negatively charged upon deprotonation of the carboxyl  (8)).
(8) Figure 6 shows (2.12 mg/g after three cycles) than those for regeneration with a NaOH solution (1.45 mg/g after three cycles), which suggested that an acidic eluent is more suitable for the GA-type resin regeneration. The silicic acid that is eluted from the resin can be in two forms, i.e., the molecular form of silicic acid (silicic acid only) and a gallatesilicate complex (C 7 O 5 H 3 -Si(OH) 2 ). In the case of HCl, silicic acid can be eluted as a complex, which frees the original ion-exchange site on the resin and is beneficial for further grafting. With NaOH, silicic acid can be eluted in its molecular form because silica is soluble in alkali. Therefore, some gallic acid remains on the surface of the resin, preventing repeated grafting. In addition, gallic acid can be easily oxidized to quinone under alkaline conditions (pKa 2 of gallic acid: 8.45), decreasing the probability of a gallate-silicate complex forming and subsequently decreasing the adsorption amount of silicic acid. However, the adsorption capacity after multiple regeneration cycles was still above 1.45 mg/g (with NaOH), which was much higher than that of the original OH-type resin (0.13-0.25 mg/g) and indicated the resin can be efficiently reused. Acidic eluent was more suitable for eluting the GA-type resin for regeneration. The regeneration cycle results were slightly lower than the ideal adsorption capacity and can be improved in future research.

CONCLUSIONS
A new, selective silicic acid removal resin was prepared by a grafting method, and the resin showed good adsorption properties for silicic acid from aqueous solutions with a wide pH range. The adsorption kinetics were faster than those of commonly available ion-exchange resins. The removal of silicic acid from solutions was studied over a concentration range of 10-400 mg/L (as Si), and the results showed that the GA-type resin adsorbs monosilicic acid, silicate ions, and polymeric silicic acid. The adsorption capacity was 4.64-4.94 mg/g with a short adsorption time, which is 30-40 times larger than that of the OH-type resin. The effects of coexisting anions on the silicic acid