1. Introduction
During wastewater treatment, contaminants are removed from water using biological, physical and mechanical processes [
1,
2]. Even though ammonium is an essential nitrogen-containing nutrient, a high amount of ammonium affects water quality and ecosystems [
3]. Agricultural, industrial or municipal wastewater sources are major sources of ammonium-caused pollution problems. Industrial processes need to be improved by treating wastewater to prevent pollution [
1,
4]. High levels of ammonium in wastewater affect natural nitrification activity. There are also European Council directives regarding urban wastewater treatment, which state that 70–80% of nitrogen from wastewater should be removed [
5].
Currently, various methods are used for the removal of ammonium from wastewater, such as ion exchange [
1], biological nitrification–denitrification [
6], chemical precipitation [
7] and air strip** [
8]. They all have their advantages and disadvantages. Air strip** is widely used under alkaline conditions in wastewater treatment but it is not an energy-efficient method, and it is also not sensitive to substances that are toxic [
2]. The chemical precipitation method is a simple and cost-effective process for wastewater treatment but uses extra reagents for the treatment process and generates new pollutants in water due to the usage of salt for precipitation [
2,
9,
10,
11]. Widely used biological methods are effective but they are not suitable for high ammonium load because removal of ac-cumulated nitrate cause additional costs [
2,
12]. Biochar-based materials should be integrated with biological treatments to improve the effective removal of various contaminants (such as nitrate, phosphate, ammonia nitrogen and chloride) [
13]. Chemical precipitation and microbiological methods are characterized by processing complexities and high operational costs, and it has also been demonstrated that in practical applications microorganisms have a lower rate of survival [
2,
14,
15]. Air strip** and similar denitrification processes are not simple to set up and are affected by low temperatures [
7]. Due to the simplicity of the processes and the fact that they are not affected by low temperatures, ion exchange and adsorption methods are effective in the removal of ammonium ions [
2,
7]. These are a few examples of the advantages and disadvantages of the currently available methods for the separation of ammonium from wastewater.
The adsorption method is a simple process for reducing and removing contaminants where the adsorbent is usually used to adsorb pollutants from wastewater [
16]. Adsorption could be based on ion exchange, physisorption or chemisorption [
17,
18]. Gao et al. (2020) have reported that ammonium removal was based on chemisorption [
19]. Compared to traditional ammonium-removal methods, the adsorption and ion exchange methods have more advantages because they are environmentally friendly processes, use simple adsorbent production technology and are also cost-effective when adsorbents can be prepared from industrial by-products [
2,
20]. Activated carbon has been commonly used as an effective adsorbent [
2]. In activated carbon, the adsorption capacity can be improved, e.g., by using different surface modification methods like surfactants [
21]. However, geopolymers are considered to be more cost-effective adsorbents and contribute to sustainable wastewater treatment processes [
17,
18].
Geopolymers are described as amorphous, non-crystalline structures produced through the reaction between aluminosilicate-containing precursors and alkaline activators at an ambient temperature. The chemical structure of geopolymers makes them suitable as efficient adsorbents in wastewater treatment since their structure consists of an alumina silicate framework which is negatively charged and is characteristic of interchangeable charge-balancing behaviour in the activator solution [
22,
23,
24,
25]. The process of charge balancing in the Na
+ cations of the geopolymer alumino-silicate results in the exchange of ammonium ions up to around 100%, which shows the high affinity nature of geopolymers for ammonium ion [
4,
18,
23,
26]. The geopolymer preparation process is simpler than that of traditional adsorbents (e.g., zeolite, activated carbon and resin) due to the low temperature used and, for example, the use of non-hazardous chemicals. In addition, the adsorption process is simple and effective [
17,
23,
27,
28,
29].
Zeolites can be used in wastewater treatment to remove heavy metals using the ion exchange method. These adsorbents are cost-effective and found naturally in salt lakes, volcanic environments and sediment layers. Na-zeolite is abbreviated to analcime in this paper. Analcime (ANA), clinoptilolite, phillipsite and dachiardite are some of the most common types of zeolites. Zeolites belong to the crystalline hydrated aluminosilicates family and have properties such as the exchangeability of anions and cations. Currently, clinoptilolite is widely used for wastewater treatment due to its good availability [
30,
31,
32]. Recent experiments show that analcime has adsorbent properties and can be used to remove ammonium ions from wastewater. ANA [Na
16(Al
16Si
32O
96)·16H
2O] is formed as a side stream during the production of lithium carbonate from spodumene (LiAlSi
2O
6) using a sodium pressure-leaching process [
33]. In recent years, metakaolin, a type of calcinated clay, has also been used as an adsorbent. Although it is not efficient as an adsorbent in the removal of ammonium ions without any treatment, geopolymerisation improves the ammonium removal capacity [
4]. In this work, analcime from industrial mining and metakaolin have been studied and used as the raw materials for geopolymerization.
Apart from the widely used batch-adsorption method, the fixed-bed column method is also broadly used in the contaminant removal process. In this method, the adsorbate is left to flow continuously over plastic or glass columns. This method is preferred because of its efficiency in managing differences in large concentrations; it can also be scaled up to meet industrial standards [
16,
34,
35]. Several empirical methods, such as the Thomas, Yoon–Nelson and Bohart–Adams models, are generally applied to column adsorption data to determine breakthrough behaviour [
16,
36,
37].
In this study, three geopolymer materials were prepared using analcime and metakaolin as raw materials. To measure the capacity of the prepared materials for ammonium ion removal, experiments were performed for ammonium ion removal and the effect of flow rate was studied. After the adsorption experiments, the stability of the adsorbent materials was studied by performing regeneration experiments. The experimental data obtained from the fixed-bed columns were applied using mathematical models: Thomas, Yoon–Nelson and Bohart–Adams. This study utilises inexpensive industrial side streams and investigates the possibility of the creation of a low-cost and simple approach to the removal of ammonium from synthesised wastewater.
2. Materials and Methods
2.1. Chemicals
Analcime was obtained from lithium carbonate production through a Finnish mining company. Metakaolin samples were obtained from Aquaminerals Finland Ltd (Paltamo, Finland). The synthetic wastewater solution was prepared using ammonium hydroxide (obtained through VWR International, Radnor, PA, USA). Hydrochloric acid and sodium hydroxide were used for pH value moderation and hydrochloric acid for the acid washing of the analcime (these were obtained through FF Chemicals, Werkendam, The Netherlands). An alkaline solution of sodium silicate, sodium hydroxide and potassium silicate (obtained through VWR international, Radnor, PA, USA) was utilised in the synthesis of the geopolymers. In this study, three different kinds of alkaline solution combinations have been used for the preparation of geopolymers. Sodium chloride and sodium hydroxide were purchased from VWR International Chemicals and used as regeneration agents.
2.2. Characterisation Techniques and Analytical Methods
The pore size, volume and the specific surface area were calculated from nitrogen gas adsorption−desorption isotherms. The experiment was carried out using Micrometrics ASAP 2020 equipment with liquid nitrogen (−196 °C). The Brunauer–Emmett–Teller (BET) formula was used to calculate specific surface area and the Barrett–Joyner–Halenda (BJH) method was used to calculate the pore surface size distribution from the desorption data. The NH4-N concentration was analysed using the Hach HQ30d equipped with an ammonium ion-selective electrode. pH measurements were also conducted using the Hach HQ30d.
2.3. Preparation of Analcime-Based Geopolymers
Sodium hydroxide, sodium silicate and potassium silicate alkaline solutions were used for synthesising geopolymers in this study. Before alkaline activation, the raw material particle size was <1 mm. GEOP1 was synthesised by adding sodium hydroxide and sodium silicate (1:1) to the analcime and metakaolin (ratio 3:1). GEOP2 was prepared from analcime by adding sodium silicate and water for the mixing of the sample. Before alkaline activation, the analcime sample was first washed in 2 M of hydrochloric acid for 24 h. After washing, the samples were kept in the oven at 105 °C for 24 to 48 h for drying. In the formation process of GEOP3, In the formation process of GEOP3, a weight ratio of 3:1 analcime and metakaolin were mixed with sodium hydroxide and potassium hydroxide (1:1). After alkaline activation the sample was poured into the silica mould and kept at room temperature for three days. De-moulded samples were crushed using a jaw crusher and sieved into <150 mm particles. All the prepared geopolymers were washed with distilled water to eliminate unreacted alkaline solution and were dried overnight in an oven at 105 °C.
2.4. Column Experiments
The fixed-bed column system was constructed using a plastic column (diameter 3 cm, height 7 cm), where the ammonium solution continuously flowed from bottom to top with the aid of a roller pump (Watson–Marlow 120 Series). At the bottom of the column, the deposited plastic sieve was followed by glass wool. Three grams (0.5 cm) of geopolymer adsorbent was placed in the fixed-bed column. To prevent the loss of adsorbent during the experiment, adsorbent was surrounded with acid-washed fine and normal sand [
28]. Adsorption behaviour was studied with flow rates of 5, 10 and 20 mL/min at room temperature. The target concentration of the ammonium solution was 40 mg/L and the bed height was 0.5 cm in all experiments. The column set-up has been presented in more detail in our previous work [
28]. The kinetics of the adsorption process were followed by collecting samples from the effluent at specific time intervals. The initial pH was adjusted to 2.5 but it increased up to 7–8 during the experiment. Alkaline pH was avoided since the NH
4+ ions tend to evaporate as ammonia in high pH values (NH
4+ pk
a value is 9.24) [
18,
38]. The adsorption method depends on the pH value of the effluent, as the H
+ ions compete with NH
4+ ions when the pH value is low. Ammonium ion concentrations were examined using an ammonium-selective electrode. When the ammonium ion concentration was above 99% of the incoming ammonium concentration, the continuous flow of the synthetic model solution was stopped and distilled water was poured through the column.
Breakthrough curves were determined as an action of different ammonium concentration and the flow rate depended on the ratio of outlet and inlet concentration (
C/C0) at a function of outflow time (
t). The maximal adsorption capacity
qtotal (mg) in the column was estimated by integrating the plot of metal adsorbed concentration versus the outflow time (
t). The adsorbed concentration for metal can be determined from the breakthrough curve (
Cad =
C0 −
C) using the equation:
where
C0 and
C denote the initial inlet and outlet concentration (mg/L), respectively.
A is the area above the breakthrough curve,
Q refers to the flow rate (mL/min) and
ttotal represents the total time for the fixed-bed column to reach saturation of adsorbent with ammonium ions (min).
The equilibrium capacity (
qe) was determined using the fixed-bed column data (mg/g), through the ratio of the total amount of ammonium ions adsorbed into the adsorbent
qtotal (mg) and the dry weight of analcime geopolymer used in the adsorbent bed (g):
2.5. Regeneration
Analcime geopolymers were regenerated after each adsorption. In total, three adsorption–regeneration cycles were implemented during this study. After the adsorption experiment, the adsorbate solution was completely removed from the column by washing it with 1 L of distilled water for 1 h using a 20 mL/min flow rate. After washing, the mixture of 0.1 M of sodium hydroxide and 0.2 M of sodium chloride with a weight ratio of 1:3 was used to regenerate the adsorption material. The regeneration experiment was performed at room temperature. The experiment was stopped when the effluent reached inlet concentration.
2.6. Breakthrough Curve Modelling
Generally used models–the Bohart–Adams, Thomas and Yoon–Nelson models–were employed to describe the adsorption process. Details concerning these models are presented in the following sections. Models have been applied to an increasing part of the breakthrough curves and parameters have been implemented to draw conclusions from the experimental data and define the behaviour of the fixed-bed column. All the models used are based on the same general equation [
37] and therefore they all have equal
R2 value.
2.6.1. Thomas Model
The Thomas model is widely used for column experimental data to investigate the prospect of breakthrough curves. This model evolved assuming the Langmuir kinetics of adsorption–desorption and second-order reversible reaction kinetics [
39]. Any axial dispersion does not derive from adsorption. The Thomas model is more applicable for a sorption process when internal and external diffusion limitations are absent. The adsorption capacity and rate constant were appraised by the Thomas model in fixed-bed column methods [
36,
37,
40].
The linearised form of the Thomas model can be given as:
where
kT indicates the Thomas rate constant (mL/min
.mg),
q0 is the equilibrium uptake capacity of the adsorbent (mg/g), m is mass of adsorbent (g),
Q (mL/min) is the flow rate of the bed,
C is the outlet ammonium concentration (mg/g),
C0 is the inlet ammonium concentration (mg/g) and term
t is the sampling time (min). The values of
kT and
q0 are calculated from the plot of ln[(
C0/
Ct)
−1] against
t.
2.6.2. Bohart–Adams Model
The Bohart–Adams model [
41] is used to describe the ratio between concentration
(C0/C) as mg/g and time t (min) in the fixed-bed column system. The model has been used to describe the initial part of the breakthrough curve. It describes the adsorption rate dependent on the metal concentration of the adsorbing species and the residual capacity of the adsorbent in the column. The Bohart–Adams model was applied to determine the adsorption capacity and the service time based on the column bed height at different flow rates [
36,
37,
40]. The Bohart–Adams model can be expressed as follows:
where
BA is the rate constant (L/mg min),
N0 denotes the saturation concentration (mg/L),
h represents the bed height (cm),
vF refers to linear flow velocity (cm/min) and
t denotes time (min).
2.6.3. Yoon–Nelson Model
The Yoon–Nelson model considers that the decrease rate of each molecule’s adsorption probability will be directly proportional to the adsorbate’s adsorption probability and the breakthrough probability of the adsorbate on the adsorbent [
42]. Adsorption probability will be proportional directly to the adsorbate adsorption probability and the breakthrough probability of adsorbate on the adsorbent [
42]. This model does not emphasise the type of the adsorbent or the adsorbate characteristics or the fixed-bed column’s parameters. [
36,
37,
40]. The equation of this model is expressed as:
where
kYN represents the Yoon–Nelson rate constant (L/min),
τ refers to the time required for 50% of ammonium ion breakthrough (min) and
t shows the breakthrough time (min).