Removal of Ammonium Ions from Aqueous Solutions Using Alkali-Activated Analcime as Sorbent

: Five alkali-activated analcime (ANA) sorbents (ANA-MK 1, ANA 2, ANA 3, ANA-MK 4, and ANA-MK 5) were developed for ammonium (NH 4+ ) ion removal. Acid treatment and calcination were used as pre-treatments for analcime, and metakaolin (MK) was used as a blending agent in three sorbents. Sorption experiments were performed to evaluate the effects of sorbent dosage (1–20 g L − 1 ), initial NH 4+ ion concentration (5–1000 g L − 1 ), and contact time (1 min–24 h). ANA-MK 1, ANA 2, and ANA-MK 4 were the most efﬁcient sorbents for NH 4+ ion removal, with a maximum experimental sorption uptake of 29.79, 26.00, and 22.24 mg g − 1 , respectively. ANA 3 and ANA-MK 5 demonstrated lower sorption capacities at 7.18 and 12.65 mg g − 1 , respectively. The results for the sorption of NH 4+ ions onto the alkali-activated analcime surfaces were modeled using several isotherms. The Langmuir, Freundlich, Sips, and Bi-Langmuir isotherms were the best isotherm models to represent the studied systems. The results of the kinetic studies showed the maximum NH 4+ ion removal percentage of the sorbents was ~80%, except for ANA-MK 5, which had a ~70% removal. Moreover, the pseudo-ﬁrst-order, pseudo-second-order, and Elovich models were applied to the experimental data. The results showed that the sorption process for ANA-MK 1, ANA 2, ANA 3, and ANA-MK 4 followed the Elovich model, whereas the pseudo-second-order model provided the best correlation for ANA-MK 5.


Introduction
Ammonium (NH 4 + ) is an essential nutrient for plants, but its excessive amount in water streams may result in eutrophication in water reservoirs, such as lakes, rivers, and seas, causing undesirable effects on the environment, aquatic fauna, and humans. In protecting water resources, the effective removal of NH 4 + from industrial wastewater prior to its discharge into natural water bodies is critical [1][2][3]. NH 4 + pollution comes from agricultural, industrial, or municipal sources. Wastewater treatment plants also contribute to NH 4 + pollution [4]. In response to this problem, many countries have established strict regulations concerning acceptable ammonia concentrations in surface waters [5]. For example, the EU urban wastewater directive has set the 15 mg/L (10,000-100,000 population equivalent) total-N limit in urban wastewater [6].
Various methods have been employed to remove nitrogen compounds from aqueous solutions. These methods include adsorption [7][8][9], air stripping [10,11], biological treatment (nitrification/denitrification process) [12,13], chemical co-precipitation [14,15], electrolysis [16,17], catalytic wet-air oxidation [18,19], ion-exchange [9,20,21] and membrane processes [22,23]. Many of these methods have several limitations, including high cost, low removal rate, high sensitivity to pH and temperature, and introduction of new pollutants. However, compared with other methods, ion exchange and adsorption techniques offer several advantages: high removal efficiency, low energy consumption, simplicity ANA a more efficient sorbent by subjecting it to different modification treatments. At first, the preliminary experiments were performed by modification analcime with different treatment methods. Based on these experiments, five alkali-activation methods were selected as modification methods for ANA. Namely, ANA-MK 1, ANA 2, ANA 3, ANA-MK 4, and ANA-MK 5 were prepared and compared in terms of their sorption efficiencies for NH 4 + ion removal. Sorption kinetics, sorption isotherms, effects of pH, sorbent dosage, and NH 4 + ion concentration were investigated. The prepared sorbents were characterized by X-ray diffraction (XRD), X-ray fluorescence (XRF), Fourier transform infrared (FTIR) spectroscopy, and field emission scanning electron microscope (FESEM) to determine their mineralogical and physicochemical characteristics. The specific surface areas (SSAs), pore sizes, and pore volumes of the samples were also determined.

Materials
Analcime was obtained from a Finnish mining company. Metakaolin (MK) was obtained from a Finnish supplier and was used as a blending agent in some of the prepared sorbents. Technical grade sodium hydroxide (NaOH; VWR International, Radnor, PA, USA), sodium silicate (7.5-8.5% Na 2 O; 25.5-28.5% SiO 2 ; VWR International), potassium silicate (K 2 SiO 3 ; VWR International), and phosphoric acid (H 3 PO 4 , 85%, VWR International)) were used to synthesized alkali-or acid-activated analcime sorbents. A synthetic NH 4 + ion stock solution was prepared from analytical grade ammonium chloride salt (NH 4 Cl, Merck KGaA, Darmstadt, Germany) and distilled water. The pH of the solution was adjusted by adding HCl and/or NaOH (FF-Chemicals, Haukipudas, Finland). All chemicals used in the experiments were of analytical grade.

Screening of Different Alkali-and Acid Activation Methods for Analcime
Analcime was pre-treated with acid washing, drying, and/or calcination prior to alkali and acid activation. The pre-treated analcime was alkali-activated using 10 M NaOH or with an alkaline solution containing 10 M NaOH and sodium silicate (7.5-8.5% Na 2 O; 25.5-28.5% SiO 2 ) at a 1:1 weight ratio. Acid activation was performed using phosphoric acid (85%). Moreover, the effect of metakaolin as a blending agent was investigated. Three alkali-activation methods and one acid activation method were employed to prepare an effective analcime sorbent for NH 4 + ion removal. The objective was to prepare a thoroughly consolidated mixture that would sufficiently harden and could be crushed and sieved to a certain particle size. The preparation methods are presented in Table 1. In addition, a schematic diagram of the preparation methods is shown in Figure 1.

Pre-Treatment of Analcime
Analcime was subjected to different pre-treatments prior to alkali or acid activation. These pre-treatments included acid washing with 2 M HCl at an L/S (liquid/solid) ratio of 100 for 24 h, drying at 105 • C overnight, and/or calcinating for 2 h at 400 • C or 700 • C. These pre-treatments are presented in Table 1 and Figure 1.

Alkali-and Acid Activation
Method 1: The pre-treated analcime was alkali-activated using 10 M NaOH or alkaline solution containing 10 M NaOH and sodium silicate (7.5-8.5% Na 2 O; 25.5-28.5% SiO 2 ) at a 1:1 weight ratio. The alkaline solution was prepared a day before use. The analcime (dried at 105 • C) was mixed with the alkaline solution at a 1:1 weight ratio. In one sample, metakaolin was used as the blending agent. The mixture was mixed for 15 min, vibrated for 2 min, and allowed to consolidate for 3 days at room temperature. The resulting solid material was crushed to a particle size of 63-125 µm. The obtained materials were washed with deionized water, dried at 105 • C, and stored in a desiccator. The alkali-activation recipes are shown in Table 1 and Figure 1.
Method 2: The pre-treated analcime was mixed with phosphoric acid (85%) at a 1:1 weight ratio. A small amount of water was added as needed. The mixture was transferred into a mold, which was placed and sealed in a small plastic bag; the mixture was then allowed to consolidate at 60 • C for 24 h. The resulting solid material was crushed to a particle size of 63-125 µm. The obtained materials were washed with deionized water, dried at 105 • C, and stored in a desiccator. The acid activation recipes are shown in Table 1 and Figure 1.

Alkali-Activation
Alkali-activated analcime sorbents were prepared using five recipes (Table 2, Figure 2). The analcime was pre-treated as in the screening tests (Section 2.2.1). ANA 2 and ANA 3 were acid-washed, and ANA 3 was further calcinated at 400 • C prior to alkali activation. Metakaolin was used as a blending agent in the case of ANA-MK 1, ANA-MK 4, and ANA-MK 5. The alkali-activated analcime sorbents were prepared using the method described in Section 2.2.2. Sodium silicate was used as an alkaline chemical, but in the case of ANA-MK 4 and ANA-MK 5, potassium silicate was used. Alkali-activated analcime sorbents were prepared using five recipes (Table 2, Figure  2). The analcime was pre-treated as in the screening tests (Section 2.2.1). ANA 2 and ANA 3 were acid-washed, and ANA 3 was further calcinated at 400 °C prior to alkali activation. Metakaolin was used as a blending agent in the case of ANA-MK 1, ANA-MK 4, and ANA-MK 5. The alkali-activated analcime sorbents were prepared using the method described in Section 2.2.2. Sodium silicate was used as an alkaline chemical, but in the case of ANA-MK 4 and ANA-MK 5, potassium silicate was used.

Characterization Methods
The SSA and pore volumes of the samples were determined from nitrogen gas adsorption−desorption isotherms at the temperature of liquid nitrogen (−196 • C) using a Micromeritics ASAP 2020 instrument. The SSA was calculated based on the Brunauer-Emmett-Teller (BET) isotherm, and the pore size distributions were calculated from the desorption data using the Barrett-Joyner-Halenda (BJH) method. The XRD patterns were recorded by a PANalytical X'Pert Pro X-ray diffractometer (Malvern Panalytical, Almelo, The Netherlands) using monochromatic CuKα1 radiation (λ = 1.5406 Å) at 45 kV and 40 mA. Diffractograms were collected within the 2θ range of 10 • -90 • at 0.017 • intervals at a scan step time of 100 s. The crystalline phases and structures of the adsorbents were analyzed using the HighScore Plus software (Version 4.0, PANalytical B. V., Almelo, The Netherlands). The peaks were identified according to the International Centre for Diffraction Data (ICDD) (PDF-4+ 2022 RDB). The phases were quantified through Rietveld analysis using HighScore.
The XRF spectra were recorded by a PANalytical Axios mAX 4 kW wavelength dispersive X-ray fluorescence spectrometer (Malvern Panalytical, Almelo, The Netherlands), wherein samples were prepared as loose powders using a mylar film under a helium atmosphere at 4 kW. The FTIR spectra at the wavenumber range of 650-4000 cm −1 were collected using a PerkinElmer Spectrum One FT-IR spectrometer (PerkinElmer, Walthman, MA, USA) equipped with an attenuated total reflectance unit. The morphology and microstructure shown in the FESEM images were obtained using a Zeiss Sigma field emission scanning electron microscope (FESEM, Carl Zeiss Microscopy GmbH, Jena, Germany) at the Centre for Material Analysis in the University of Oulu operated at 5 kV.

Batch Sorption Experiments
The sorption properties of the alkali-activated analcime sorbents were characterized by batch experiments. Experiments that aimed to determine the effect of sorbent dose and initial NH 4 + ion concentration were performed in 50 mL centrifuge tubes, which were placed in a shaker for 24 h. At the end of the sorption experiment, the sorbents were separated through centrifugation (3500 rpm, 2-5 min), and the supernatant was sampled with a pipette. NH 4 + -N concentration was analyzed using Hach HQ30d equipped with an NH 4 + ion-selective electrode. Therefore, the concentration values presented are those of NH 4 + -N, although the phenomenon being studied (including trends and conclusions) is the removal of NH 4 + ions. Given that NH 4 + ions exist as ammonia in alkaline conditions, the pH was maintained below 9 during the batch experiments by adjusting the initial pH (0.1/1M HCl) to 2.5 [50,51].
Kinetic studies were performed in a 1 L reactor vessel equipped with a magnetic stirrer with an agitation speed of 800 rpm. Then, 30 mL of the sample was taken out from the vessel at different time intervals from 0 h to 24 h of sorption and then centrifuged prior to analysis. The detailed sorption conditions are presented in Table 3. Table 3. Target parameters used in the batch sorption experiments. All experiments were performed at room temperature (21 • C-23 • C). The initial pH of the synthetic NH 4 + ion solution was adjusted to 2.5. The removal efficiency (%) and sorption capacity for NH 4 + ion were calculated using Equations (1) and (2), respectively:

Parameter
C 0 and C e (mg L −1 ) are the initial and equilibrium liquid phase NH 4 + ion concentrations, respectively, V (L) is the volume of the solution, and m (g) is the mass of the sorbent.

Sorption Isotherms
Several non-linear isotherm models (Langmuir, Freundlich, Sips, Bi-Langmuir, Toth, Temkin, and Dubin-Radushkevich) were applied to the experimental data. Four models provided accurate results. These models were the traditional Langmuir and Freundlich isotherms, the Sips model, which is suitable for heterogeneous sorption, and the Bi-Langmuir model, which is suitable when sorption occurs in two types of sorption sites.
The other models did not provide any additional information about the sorption of NH 4 + . The isotherm parameters were obtained through non-linear regression using OriginPro 2018, and the equations used are as follows: The non-linear form of Langmuir's [52] equation is where b L (L mg −1 ) represents the energy of sorption, q m (mg g −1 ) is the Langmuir constant, which is related to sorption capacity, and C e (mg L −1 ) is the NH 4 + ion concentration of the solution in equilibrium. The Freundlich model [53] can be written as where K f (L g −1 ) and n F are Freundlich constants, which are related to sorption capacity and intensity, respectively. The Sips isotherm [54] combines the properties of both the earlier mentioned models and is given as follows: Analogous to the Langmuir model, in this equation, b S (L mg −1 ) is a constant related to sorption energy. The equation for the Bi-Langmuir model [55] is similar to that for the Langmuir model, except that it includes one term for both sorption sites: To evaluate the fit of the isotherm equations to the experimental data, the residual root mean square error (RMSE) was determined. The smaller the error function value, the better the curve fitting. The calculated expression of the error function is as follows: where n is the number of experimental data points, p is the number of parameters in the isotherm model, and q e(exp) (mg g −1 ) and q e(calc) (mg g −1 ) are the experimental and calculated values of the sorption capacity in equilibrium, respectively.

Reaction Kinetics of the Sorption Process
The kinetic parameters of the sorption experiments were solved using the non-linear forms of three models, namely, the traditionally used pseudo-first-order and pseudosecond-order models and the Elovich model, which could describe well slow sorption processes and which has gained more popularity in recent years.
The non-linear form of the pseudo-first-order equation [56] is where q e (mg g −1 ) and q t (mg g −1 ) are the amounts of NH 4 + ion sorbed at equilibrium and at time t (min), respectively; in addition, k 1 (min −1 ) is the pseudo-first-order rate constant. The pseudo-second-order equation [57] can be written as ChemEngineering 2023, 7, 5 8 of 23 where k 2 is the pseudo-second-order rate equilibrium constant (g mg −1 min −1 ). The non-linear form of the Elovich equation [58] is where υ 0 (mg g −1 min −1 ) is the initial sorption rate and β (g mg −1 ) is the desorption constant.

Intra-Particle Diffusion Model
To identify the diffusion mechanism, the intra-particle diffusion model based on the theory proposed by Weber and Morris [59] was used. The Weber-Morris equation can be written as follows: where K id (mg g −1 h −1/2 ) is the intra-particle diffusion rate constant, and C is the intercept.

Screening of the Preparation Methods
In this study, several sorption materials were prepared under different preparation conditions. The preparation methods for the materials used in the batch sorption experiments were developed based on the information obtained during the screening test. The results showed ( Table 1) that either NaOH or phosphoric acid did not successfully consolidate the materials regardless of the pre-treatment employed. This can be because alkali and acid concentrations were too high, damaging the zeolite surface and internal pore structure [44]. By contrast, the treatment with an alkaline solution (10 M NaOH + SiO 2 :Na 2 O) resulted in consolidation when metakaolin was used as a blending agent, when analcime was acid washed, or when analcime was acid washed and calcinated (400 • C or 700 • C) before activation. Calcination without acid washing did not lead to consolidation. In other recipes (Section 2.3), analcime was pre-treated with acid washing and calcination. Calcination was performed only at 400 • C, as the objective was to prepare sorbents using the least amount of energy and chemicals. In addition, the preliminary sorption experiments (data not shown) indicated that sorption capacity was slightly lower during calcination at 700 • C than at 400 • C. Alkali activation using an alkaline solution (sodium silicate and NaOH) was performed to improve the sorption properties of analcime. Apart from sodium silicate, potassium silicate was also tested as an alkaline chemical.

Characterization of the Sorbents
The SSA, average pore size and volumes of the alkali-activated analcime sorbents are shown in Table 4. The results showed that ANA-MK 1 had the lowest SSA (4.182 m 2 /g) and the highest average pore diameter (31.281 nm). By contrast, ANA-MK 5 showed the highest SSA (46.712 m 2 /g) and the lowest average pore diameter (5.384 nm). The results showed that ANA-MK 5 has considerably higher SSA (46.712 m 2 /g) compared with ANA-MK 1 (4.182 m 2 /g) and ANA-MK 4 (7.784 m 2 /g). In all these sorbents, metakaolin was used as a blending agent. The ratio of analcime and metakaolin was 3:1 in the case of ANA-MK 1 and ANA-MK 4. For ANA-MK 5, the ratio was 5.4:1. Alkaline solution (10 M NaOH + Na-silicate) was used in the case of ANA-MK 1, and alkaline solution (10 M NaOH + K-silicate) in the case of ANA-MK 4 and ANA-MK 5. Based on the results, sorbent prepared using an alkaline solution consisting of potassium silicate in the alkaline solution and using a ratio of analcime and metakaolin 5.4:1 leads to the highest SSA value (ANA-MK 5: 46.712 m 2 /g) in this study. A mesopore structure was clearly dominant in each sorbent. Moreover, the surface areas of the alkali-activated analcime sorbents are higher than those of the non-modified sorbents (SSA: 3.01 m 2 /g) [60]. The chemical and mineral compositions of the alkali-activated analcime sorbents are presented in Table 5. The alkali-activated sorbents mainly consisted of silica and alumina with Si/Al mole ratios of 4.05, 2.31, 1.30, 1.36, and 1.44 for ANA-MK 1, ANA 2, ANA 3, ANA-MK 4, and ANA-MK 5, respectively. The mole ratio of raw analcime has previously been found to be 1.57 [49]. Raw analcime [47] and the alkali-activated analcime sorbents (Table 5) also contained exchangeable cations, such as Fe 3+ , Na + , Ca 2+ , Mg 2+ , and Mn 2+ . These cations play an important role in ion exchange with NH 4 + ions [3]. In addition, the sorbents contained some impurities and volatile compounds (e.g., water).  8 ). In addition, analcime may contain silicon oxide (quartz, SiO 2 ) and lithium aluminum silicate (spodumene, LiAlSi 2 O 6 ) [49,60]. The obtained XRD patterns in the present study are shown in Figure 3. According to the software there are many candidates, and many peaks are overlapping. All of the alkali-activated sorbents contained silicon oxide (SiO 2 , ICDD 04-014-7568), iron oxide ( The results showed that ANA-MK 1 considerably differed from the other alkali-activated analcime sorbents. Peak intensities were lower in ANA-MK 1; moreover, a high-intensity peak at position 27, which is associated with sodium aluminum silicate hydrate (NaAl(Si 2 O 6 )(H 2 O), ICDD 01-074-2219), was not observed in ANA-MK 1 compared with the other preparations. In the case of ANA-MK 1, SiO 2 was crystallized very well, and the intensity peak at 2θ of 26.6 • was very high.
ICDD 00-056-0629), and potassium sodium calcium aluminum silicate (K0.05Na0.94Ca0.01Al1.01Si2.99O8, ICDD 04-023-4722). The results showed that ANA-MK 1 considerably differed from the other alkali-activated analcime sorbents. Peak intensities were lower in ANA-MK 1; moreover, a high-intensity peak at position 27, which is associated with sodium aluminum silicate hydrate (NaAl(Si2O6)(H2O), ICDD 01-074-2219), was not observed in ANA-MK 1 compared with the other preparations. In the case of ANA-MK 1, SiO2 was crystallized very well, and the intensity peak at 2θ of 26.6° was very high. The FTIR spectra of the alkali-activated analcime sorbents are shown in Figure 4. The peaks between ~1621 and 1628 cm −1 can be attributed to the hydroxyl group with the stretching and bending vibrations of the adsorbed water [46,61]. The band between ~979 and 1031 cm −1 can be assigned to the internal stretching vibrations of Si-O(Si) and Si-O (Al) [46,62]. The bands between 697 and 880 cm −1 belong to the T-O (T=Al, Si) symmetric stretching vibrations [60,[63][64][65][66]. These vibrations are typical for zeolite materials. The FTIR spectra of alkali-activated analcime sorbents exhibit similar vibrations to raw analcime and Ba-modified analcime [60]. The FTIR spectra of the alkali-activated analcime sorbents are shown in Figure 4. The peaks between~1621 and 1628 cm −1 can be attributed to the hydroxyl group with the stretching and bending vibrations of the adsorbed water [46,61]. The band betweeñ 979 and 1031 cm −1 can be assigned to the internal stretching vibrations of Si-O(Si) and Si-O (Al) [46,62]. The bands between 697 and 880 cm −1 belong to the T-O (T=Al, Si) symmetric stretching vibrations [60,[63][64][65][66]. These vibrations are typical for zeolite materials. The FTIR spectra of alkali-activated analcime sorbents exhibit similar vibrations to raw analcime and Ba-modified analcime [60]. Figure 5 shows the FESEM images of the raw materials analcime and metakaolin, as well as prepared sorbents ANA-MK 1, ANA 2, ANA 3, ANA-MK 4, and ANA-MK 5. It can be seen from Figure 5a-c that analcime consists of crystalline phases, which is typical for zeolite materials [67,68]. Figure 5d-f presents a FESEM image of metakaolin showing mainly flakes structure; on the flake surface, there are heterogeneous materials consisting of irregularly shaped particles [69,70]. After alkali activation, samples seem to have different morphological structures compared with raw materials analcime and metakaolin. This can be affected by different pre-treatments for analcime and alkali-activation processes. As can be seen from Figure 5, alkali-activated samples contain voids, unreacted particles, and cracks. In addition, it is possible to see a gel network characteristic of alkali-activated materials, which has also been reported in other studies [70][71][72][73][74][75][76][77].  Figure 5 shows the FESEM images of the raw materials analcime and metakaolin, as well as prepared sorbents ANA-MK 1, ANA 2, ANA 3, ANA-MK 4, and ANA-MK 5. It can be seen from Figure 5a-c that analcime consists of crystalline phases, which is typical for zeolite materials [67,68]. Figure 5d-f presents a FESEM image of metakaolin showing mainly flakes structure; on the flake surface, there are heterogeneous materials consisting of irregularly shaped particles [69,70]. After alkali activation, samples seem to have different morphological structures compared with raw materials analcime and metakaolin. This can be affected by different pre-treatments for analcime and alkali-activation processes. As can be seen from Figure 5, alkali-activated samples contain voids, unreacted particles, and cracks. In addition, it is possible to see a gel network characteristic of alkaliactivated materials, which has also been reported in other studies [70][71][72][73][74][75][76][77].

Effect of Sorbent Dosage
The sorbents used in this study are alkaline in nature, so they increase the pH during sorption experiments. The pH was maintained below 9 during the experiment by adjusting the initial pH to 2.5 after the addition of the sorbent to the solution. This was done because, as has been reported, NH 4 + ions evaporate as ammonia at high pH values [2,50,51,78]. pH 2.5 was selected as the initial pH based on preliminary experiments (data not shown). However, a pH that is too low is not good because, under this condition, H + ions start to compete with NH 4 + ions [78]. The results ( Figure 6) showed that the removal efficiency of NH 4 + ions increased with increasing sorbent dosage. The same trend has been reported in several sorption studies. This observation can be explained by the fact that when sorbent content increases, the surface area and the number of available sorption sites increase [2,50,78]. On the contrary, the sorption capacities (q e ) decreased when the sorbent dosage increased. This can be explained by the increase in the sorbent-to-adsorbate ratio. Moreover, these results showed that the removal percentage values were much higher when a lower NH 4 + ion concentration was used. This same phenomenon has been observed in concentration optimization studies (Section 3.3).
ANA-MK 1 was the most efficient among the prepared sorbents based on the removal percentage values. When a low NH 4 + ion concentration was used, the removal efficiency was approximately 30% at sorbent dosages 1 and 2 g L −1 . The removal percentage increased rapidly when the sorbent dosage was increased. The removal percentage was approximately 75-85% within the sorbent dosage range of 5-20 g L −1 . The decreasing order of the removal percentages of the sorbents at a sorbent dosage of 5 g L −1 and at a low initial NH 4 + ion concentration was as follows: ANA-MK 1 (75%) > ANA 3 (40%) > ANA 2 (35%) > ANA-MK 5 (20%) > ANA-MK 4 (15%). Although slightly higher removal percentages could be reached by using a higher sorbent dosage, 5 g L −1 was selected for the subsequent experiments, given the risk that pH will increase during the sorption experiment, driving the transformation of NH 4 + ions into ammonia. ChemEngineering 2023, 7, x FOR PEER REVIEW 12 o  approximately 75%−85% within the sorbent dosage range of 5−20 g L −1 . The decreasing order of the removal percentages of the sorbents at a sorbent dosage of 5 g L −1 and at a low initial NH4 + ion concentration was as follows: ANA-MK 1 (75%) > ANA 3 (40%) > ANA 2 (35%) > ANA-MK 5 (20%) > ANA-MK 4 (15%). Although slightly higher removal percentages could be reached by using a higher sorbent dosage, 5 g L −1 was selected for the subsequent experiments, given the risk that pH will increase during the sorption experiment, driving the transformation of NH4 + ions into ammonia.

Effect of Initial NH4 + Ion Concentration
The effect of initial NH4 + ion concentration was investigated using model solutions within the concentration range of 5-1000 mg L −1 . The results are presented in Figure 6. The concentrations used were selected for practical reasons, as sorbents should be suitable for use in a wide range of ion concentrations. The NH4 + ion concentration in wastewater from the mining industry is typically quite low (below 40 mg L −1 ), whereas that from agricultural sources may be as high as 900 mg L −1 . It can be seen in Figure 7 that the removal percentage decreased with increasing initial NH4 + ion concentration. In the case of ANA-MK 1 and ANA 2, the NH4 + ion removal efficiencies increased rapidly at first. This can be explained by the fact that the driving force of the concentration gradient increases with increasing initial NH4 + ion concentration. When the initial NH4 + ion concentration was higher than 50 mg L −1 (NH4 + -N: ~40 mg L −1 ), the removal efficiencies increased more slowly, reaching a plateau; subsequently, the uptake started to decrease. This phenomenon can be explained by the fact that at low initial concentrations, a sufficient number of available pores on the sorbent surface may adsorb most of the NH4 + ions, which in turn would occupy the available sites. As a result, the sorption efficiency increases to a certain level. Several investigations have shown that the removal efficiency of NH4 + ions is concentration dependent and that removal efficiency decreases with increasing initial ion concentration [2]. Based on these results, the optimum initial NH4 + ion concentration for the studied sorbents was 50 mg L −1 . The decreasing order of removal percentages of the sorbents was as follows: ANA-MK 4 (~75%) > ANA-MK 1 (~55%) > ANA 2 (~50%) > ANA 3 (~40%) > ANA MK 5 (~40%).

Sorption Isotherms
The functionality of the Langmuir, Freundlich, Sips, Bi-Langmuir, Toth, Temkin, and Dubin-Radushkevich models were tested for the sorption of NH 4 + ions onto the alkaliactivated analcime surfaces. However, only the results obtained using the first four models are presented. The experimental results and the modeled isotherms are illustrated in Figure 8. The isotherm parameters, correlation coefficients, and calculated errors given by the models are shown in Table 6.

Sorption Isotherms
The functionality of the Langmuir, Freundlich, Sips, Bi-Langmuir, Toth, Temkin, and Dubin-Radushkevich models were tested for the sorption of NH4 + ions onto the alkaliactivated analcime surfaces. However, only the results obtained using the first four models are presented. The experimental results and the modeled isotherms are illustrated in Figure 8. The isotherm parameters, correlation coefficients, and calculated errors given by the models are shown in Table 6.   On the basis of correlation coefficients (R 2 ), all models represented quite well the studied systems. The Bi-Langmuir and Sips isotherm models showed the best fit for the sorption of NH 4 + ions onto the sorbent surfaces. Nevertheless, the Langmuir and Bi-Langmuir models gave more realistic q-values than the Sips model. The shape of the curves produced by the Sips and Bi-Langmuir models were quite similar for all the studied sorbents, as can also be seen in the R 2 values with the same order of magnitude. These findings support the conclusion that the surface of the adsorbent is heterogeneous; that is, it possesses two types of occupied sorption sites and exhibits different sorption energetics. Moreover, the RMSE values reinforced the conclusion that the RMSE values are smaller when the R 2 values are higher.
In the case of ANA-MK 1, the n S value approached unity, which means that the R 2 value of the Sips model approached that of the Langmuir model, leading to nearly identical R 2 values for these models. Still, the Langmuir and Bi-Langmuir models showed a slightly better correlation than the Sips model, indicating that the process apparently followed the single monolayer sorption. In the case of ANA-MK 4 and ANA-MK 5, the Freundlich model provided quite similar correlations compared with the Sips and Bi-Langmuir models, whereas the Langmuir provided a weak correlation. In terms of sorption capacities, the decreasing order of the effectiveness of the prepared sorbents was as follows: ANA-MK 1 (29.79 mg g −1 ) > ANA 2 (26.00 mg g −1 ) > ANA-MK 4 (22.24 mg g −1 ) > ANA-MK 5 (12.65 mg g −1 ) > ANA 3 (7.18 mg g −1 ). The only difference in the preparation of ANA-MK 1 and ANA-MK 4 was the activation with an alkali chemical ( Table 2), demonstrating that Na-silicate with NaOH produced a slightly better adsorbent than the NaOH + K-silicate combination. In the case of ANA-MK 1, the final pH was higher than other sorbents (Figure 4). NH 4 + starts to evaporate as ammonia at high pH values, and due to that, NH 4 + removal can be based partly on evaporation as ammonia in the case of ANA-MK 1. Considering that the adsorption capacity of ANA-MK 4 was nearly twice that of ANA-MK 5, we can conclude that a higher metakaolin content of the prepared adsorbent leads to a more effective removal. In fact, metakaolin is apparently a requirement for the consolidation of materials treated with Na-silicate (Table 1). However, an effective adsorbent may be produced using only analcime as raw material (ANA 2) only if analcime is acid-washed prior to alkali activation. In the preparation step, it seemed that calcination also had a positive effect on consolidation; however, the sorption experiment showed that the calcinated ANA 3 had the worst q-value. The difference between the q-values of ANA 2 and ANA 3 may be due to the different alkali-activation treatments employed. The use of Na-silicate without NaOH also seems to produce effective adsorbents. Meanwhile, the SSA and average pore diameter of ANA-MK 5 differed from those of the other sorbents. This led to a slower sorption capacity (12.65 mg g −1 ) compared with that of the most effective adsorbents, namely, ANA-MK 1 (29.79 mg g −1 ), ANA 2 (26.00 mg g −1 ), and ANA-MK 4 (22.24 mg g −1 ).
The NH 4 + removal mechanism of zeolites and alkali-activated is mostly based on ion exchange, which involves a reversible replacement of ions of the same charge [9,44,45,[79][80][81][82]. The aluminosilicate structure is negatively charged, and the exchangeable cations are located in the voids [83,84]. High meso-and macro-pore volumes in alkali-activated sorbent enhance active contact sites for NH 4 + sorption [84]. Several factors affect the ion-exchange behavior of zeolites and alkali-activated materials. These factors are the framework structure, ion size and shape, charge density of the anionic framework, ionic charge, and concentration of the external electrolyte solu-tion [45,83]. For example, based on FESEM images (Figure 5), the porosity of the prepared alkali-activated sorbents has been increased compared to analcime. This can be one explanation for why these alkali-activated sorbents have higher q-values compared to raw analcime (<10 mg/g) [49].

Effect of Contact Time
Kinetic studies were performed for all the prepared sorbents at room temperature using the model solution with an NH 4 + ion concentration of 50 mg L −1 (c(NH 4 + -N), 40 mg L −1 ). As can be seen in Figure 9, the removal percentage increased quite rapidly, reaching the maximum removal percentage and equilibrium. The availability of a larger number of sorption sites can explain this observation. The maximum NH 4 + ion removal percentage of the sorbents was~80%, except for ANA-MK 5, which had a~70% removal.

Effect of Contact Time
Kinetic studies were performed for all the prepared sorbents at room temperature using the model solution with an NH4 + ion concentration of 50 mg L −1 (c(NH4 + -N), 40 mg L −1 ). As can be seen in Figure 9, the removal percentage increased quite rapidly, reaching the maximum removal percentage and equilibrium. The availability of a larger number of sorption sites can explain this observation. The maximum NH4 + ion removal percentage of the sorbents was ~80%, except for ANA-MK 5, which had a ~70% removal.

Kinetic Modeling
The pseudo-first-order, pseudo-second-order, and Elovich models were applied to the experimental data. The kinetic parameters and correlation coefficients are presented in Table 7, and the graphs are shown in Figure 10.

Kinetic Modeling
The pseudo-first-order, pseudo-second-order, and Elovich models were applied to the experimental data. The kinetic parameters and correlation coefficients are presented in Table 7, and the graphs are shown in Figure 10.
Comparison of the correlation coefficients and RMSE values revealed that sorption of NH 4 + ions onto the ANA-MK 1, ANA 2, ANA 3, and ANA-MK 4 surfaces were best represented by the Elovich model, indicating that sorption was relatively slow. However, sorption onto ANA-MK 5 differed from that observed in the other sorbents, as it showed a more angular curve. This pattern showed the best correlation obtained using the pseudofirst-order model. Although the pseudo-first-order and pseudo-second-order models rarely showed a poor correlation, the correlation between the experimental and calculated qvalues was reasonable. The differences in surface structures might be the reason behind having a different optimum kinetic model for ANA-MK 5 (pseudo-first-order) and for the other adsorbents (Elovich), an indication of differences in sorption behaviors.  Comparison of the correlation coefficients and RMSE values revealed that sorption of NH4 + ions onto the ANA-MK 1, ANA 2, ANA 3, and ANA-MK 4 surfaces were best represented by the Elovich model, indicating that sorption was relatively slow. However, sorption onto ANA-MK 5 differed from that observed in the other sorbents, as it showed a more angular curve. This pattern showed the best correlation obtained using the pseudo-first-order model. Although the pseudo-first-order and pseudo-second-order models rarely showed a poor correlation, the correlation between the experimental and calculated q-values was reasonable. The differences in surface structures might be the reason behind having a different optimum kinetic model for ANA-MK 5 (pseudo-first-order) and for the other adsorbents (Elovich), an indication of differences in sorption behaviors.

Weber and Morris Intraparticle Diffusion Model
The plot qt versus t0.5 for the sorption of NH4 + ions onto the prepared sorbents is shown in Figure 11. The results showed that sorption proceeds in three or four stages. The first stage proceeds quite fast, wherein NH4 + ions diffuse through the solution to reach the

Weber and Morris Intraparticle Diffusion Model
The plot q t versus t 0.5 for the sorption of NH 4 + ions onto the prepared sorbents is shown in Figure 11. The results showed that sorption proceeds in three or four stages. The first stage proceeds quite fast, wherein NH 4 + ions diffuse through the solution to reach the external surface of the sorbent. This stage is called the external surface sorption or the macro-pore diffusion. Internal surface sorption or micro-pore diffusion occurs in the second and third steps. In the final step, small slopes indicate slow intraparticle diffusion onto the internal surface due to the extremely low solute concentrations in the bulk phase and boundary layer [3].

Comparison with the Other Adsorbents
The sorption capacities of the aluminosilicate-based materials during NH4 + ion removal are presented in Table 8. The materials exhibiting a high sorption capacity are obtained from virgin sources. The material developed in this study demonstrated a higher sorption capacity compared with other side stream-based materials, such as a zeolite synthesized from fly ash. Therefore, the sorption capacity obtained in this study indicated a very good functionality towards NH4 + ion of the sorbent produced using industrial side stream as raw material. However, comparing q values, it is good to consider operational conditions that are not exactly the same in all studies. For example, the initial pH is quite high (7)(8) in some studies so ammonium can be partly evaporated during the adsorption experiment. In addition, the used adsorption temperature varies between 20-35 °C. Worth noting that 15 degrees may have a major impact on adsorption capacity. Initial concentration values and sorbent dosages vary between different studies even though values are mostly in the same scale compared to the values used in this study.
The adsorbents studied in this study have an advantage from the local circular economy point of view. Analcime is formed as a side stream in lithium production, and the needed amount of lithium will increase in the future due to the green transition through electric vehicles. Therefore, the solutions to utilize local side streams need to be developed, and the adsorbents developed in this study are one answer.

Comparison with the Other Adsorbents
The sorption capacities of the aluminosilicate-based materials during NH 4 + ion removal are presented in Table 8. The materials exhibiting a high sorption capacity are obtained from virgin sources. The material developed in this study demonstrated a higher sorption capacity compared with other side stream-based materials, such as a zeolite synthesized from fly ash. Therefore, the sorption capacity obtained in this study indicated a very good functionality towards NH 4 + ion of the sorbent produced using industrial side stream as raw material. However, comparing q values, it is good to consider operational conditions that are not exactly the same in all studies. For example, the initial pH is quite high (7)(8) in some studies so ammonium can be partly evaporated during the adsorption experiment. In addition, the used adsorption temperature varies between 20-35 • C. Worth noting that 15 degrees may have a major impact on adsorption capacity. Initial concentration values and sorbent dosages vary between different studies even though values are mostly in the same scale compared to the values used in this study.
The adsorbents studied in this study have an advantage from the local circular economy point of view. Analcime is formed as a side stream in lithium production, and the needed amount of lithium will increase in the future due to the green transition through electric vehicles. Therefore, the solutions to utilize local side streams need to be developed, and the adsorbents developed in this study are one answer.

Conclusions
A total of five alkali-activated analcime sorbents, namely, ANA-MK 1, ANA 2, ANA 3, ANA-MK 4, and ANA-MK 5, were prepared and studied for NH 4 + ion removal. Analcime was pre-treated with acid washing and/or calcination at a high temperature. ANA 2 and ANA 3 were acid-washed, and ANA 3 was further calcinated at 400 • C prior to alkali activation. In the case of ANA-MK 1, ANA-MK 4, and ANA-MK 5, metakaolin was used as a blending agent. ANA-MK 1, ANA 2, and ANA 3 were alkali-activated using sodium silicate. In the case of ANA-MK 4 and ANA-MK 5, potassium silicate was used as the alkaline chemical for activation. The effects of sorbent dosage (1-20 g L −1 ), initial NH 4 + ion concentration (5-1000 g L −1 ), and contact time (1 min-24 h) were studied. The results showed that ANA-MK 1, ANA 2, and ANA-MK 4 were the most efficient sorbents for NH 4 + ion removal, with a maximum experimental sorption uptake of 29.79, 26.00, and 22.24 mg g −1 , respectively. ANA 3 and ANA-MK 5 demonstrated lower sorption capacities at 7.18 and 12.65 mg g −1 , respectively. The results of the kinetic studies showed the maximum NH 4 + ion removal percentage of the sorbents was~80%. The exception was ANA-MK 5, which had a~70% removal. Different isotherm and kinetic models were applied to generate the experimental results for the prepared alkali-activated analcime sorbents. The equilibrium data were best represented by the Langmuir, Freundlich, Sips, and Bi-Langmuir isotherms. In the case of ANA-MK 1, ANA 2, ANA 3, and ANA-MK 4, the Elovich model provided the best correlation. ANA-MK 5 followed the pseudo-second-order model.
The results suggested that the alkali-activated analcime sorbents are feasible for NH 4 + ion removal for wastewater treatment (e.g., municipal wastewater treatment). Pre-treatment involving acid washing prior to alkali activation or pre-treatment involving calcination enhanced the NH 4 + ion removal efficiency. Moreover, the addition of metakaolin as a blending agent increased the NH 4 + ion removal efficiency. The sorption capacities of the sorbents previously developed for NH 4 + ion removal are similar to the results obtained in this study. Funding: This study was conducted as part of WaterPro (ERDF Project No. A74635, funded by the Central Ostrobothnia Regional Council, the European Union, the European Regional Development Fund, and the Leverage from the EU) and was supported by Maaja vesitekniikan tukiry and K.H. Renlund Foundation.
Institutional Review Board Statement: Not applicable.

Informed Consent Statement: Not applicable.
Data Availability Statement: All processed data supporting the findings of this study are included in this manuscript.