1. Introduction
As an essential component of the nitrogen cycle, the nitrate ion is ubiquitous in the environment. Being highly water soluble, it is one of the most widespread contaminants of groundwater and surface water, posing a serious threat to global supplies of drinking water and aquatic ecosystems (as a promoter of eutrophication) [
1]. Although it occurs naturally, the increase in nitrate levels in water are especially promoted by the excessive use of nitrogenous fertilizers or manure in intensive agriculture [
2,
3,
4]. Increased concentrations of nitrate in potable water have been associated with many types of cancer, diabetes, infectious diseases, cyanosis in children, and the possible formation of nitrosamines, which are carcinogenic and can cause baby blue syndrome [
4,
5]. The WHO has limited the nitrate concentration in potable water to 10 mg L
−1 (NO
3−−N).
In addition to its high solubility in water, the nitrate ion is also very stable, so its removal could be challenging [
6,
7]. For this reason, various physical and chemical methods for nitrate removal have been studied and developed, such as adsorption [
8], ion exchange [
9,
10], reverse osmosis [
11], electrodialysis [
12], denitrification [
13], catalytic reduction [
14], and many others. Removal of nitrate from water is of utmost importance and therefore optimization of existing technologies is also crucial [
15].
The most popular, simple, and efficient methods for nitrate removal are adsorption and ion exchange [
3,
16]. From an economic perspective, adsorbents should be efficient, cheap, and highly selective for pollutants. With this in mind, various adsorbents have been tested for the removal of nitrate and other pollutants from water, such as clay, chitosan, zeolite, carbon-based adsorbents, and agro-industrial waste materials (many of which contain lignocellulose) [
1]. However, of all the adsorbents listed, activated carbon is still considered the most efficient, probably due to its large surface area and versatility. However, activated carbons available on the market are mainly based on coal, which is a non-renewable resource and often expensive. Therefore, the long-term sustainability of coal-based activated carbon may be in question. To address this issue, more sustainable options are currently being investigated, such as the use of lignocellulosic and other waste materials from the agro-food industry [
17]. Lignocellulosic materials are mainly composed of lignin, cellulose, and hemicellulose [
18] and can bind a variety of substances due to their structure and chemical composition. These materials have demonstrated great potential for water and wastewater treatment. However, to increase the adsorption capacity or to favorably influence the selectivity of materials to be used as adsorbents, various modifications of their surface are often required. Modification techniques are usually divided into chemical, physical, biological, and electrochemical [
4]. Chemical modifications are usually carried out with acids (inorganic and organic), salt and alkali solutions, oxidizing agents, and other chemicals [
19].
As a waste material/by-product of the food industry, hazelnut shells (HS) are available in significant quantities in some countries, often not only during the harvest season. The annual production of hazelnuts in 2019/2020 was about 528,070 tons, of which roughly 67% (i.e., 353,897 tons) are shells [
20,
21]. Hazelnut shells are mostly used as fuel (thermal utilization), since their calorific value is comparable to that of wood. They also have the similar chemical composition, such as wood, HS consists mainly of the lignocellulosic polymers, lignin, cellulose, and hemicellulose. Another possible use of untreated HS based on its chemical composition, is as an adsorbent. Native (unmodified) hazelnut shells have been used for adsorptive removal of copper ions from water [
22], dye removal [
23] and chlorophenols [
24], while modified hazelnut shells (in form of activated carbon or chemically modified) have been found to be effective for lead [
25], cadmium, zinc, copper [
26], uranium (VI) [
27], arsenic (III) [
28], chromium (VI) [
29], methylene blue [
30], crystal violet [
31], and taxol (anticancer drug) [
32] removal. After adsorptive removal of pollutants, the same material (now loaded with adsorbate) can be used directly as a fuel, eliminating the often difficult step of regeneration and disposal of the used adsorbent [
23]. Moreover, HS can also be used for the production of activated carbon after a suitable thermal treatment.
The objective of this research was to prepare a novel adsorbent by chemical modification of hazelnut shells and evaluate its potential for the removal of nitrate from wastewater. The effects of various process parameters on the adsorptive removal of nitrate using the modified HS were investigated in a batch process, namely, initial adsorbent concentration, contact time, initial nitrate concentration, and pH. The regeneration capacity of the novel adsorbent and the possibility of using it in real water treatment systems were tested in fixed bed column experiments
2. Materials and Methods
2.1. Materials
All chemicals used were of analytical grade. Epichlorohydrin (ECH) and ethylenediamine were purchased from Sigma Aldrich (Sigma Aldrich, St. Louis, MO, USA), triethylamine Fisher Scientific (Leicestershire, UK) and N,N-dimethylformamide (DMF) were from GramMol (GramMol, Zagreb, Croatia). The nitrate solutions were prepared using potassium nitrate (Merck, Darmstadt, Germany). To prepare a 1000 mg L−1 (as N-NO3−) nitrate stock solution, 7.218 g KNO3 was dissolved in 1 L demineralized water. The adsorbate solutions of the desired concentration (10–300 mg L−1) were prepared by appropriate dilution of the stock solution.
The model (synthetic) wastewater (MW) was prepared as in Kosjek et al. [
33] and the appropriate amount (10–30 mg L
−1) of KNO
3 solution was added. The local confectionery factory (CW) and dairy industry (DW) provided the 24 h composite samples of the real wastewater.
2.2. Adsorbent Preparation
PP Orahovica d.o.o., Croatia, kindly provided the hazelnut shells (HS). A laboratory knife mill (MF10 basic, IKA Labortechnik, Germany) equipped with a 1 mm sieve was used to grind the material. The procedure described by Keränen et al. [
34], i.e., the epichlorohydrin–triethylamine (ETM) method, was slightly adapted and used for the chemical modification of HS. Here, 2 g HS was mixed with 16 mL DMF and 13 mL ECH at 70 °C. After 45 min, 2.5 mL of ethylenediamine was added to the mixture and stirred for another 45 min at 80 °C. The introduction of amine groups was achieved by adding 13 mL of trimethylamine to the mixture and stirring at 80 °C for 120 min. The obtained modified material was washed with ultra-pure water and dried at 100 °C for 24 h.
2.3. Adsorbent Characterization
HS and MHS morphology and surface characteristics were studied using a field emission scanning electron microscope (FE SEM, JOEL, JSM-7000 F, Akishima, Tokyo, Japan). The Perkin Elmer CHNS/O analyzer (II series, Waltham, MA, USA) was used for elemental composition analysis (C, H, N), while the Fourier transform infrared spectrometer (FT-IR) (Cary 630, Agilent Technologies, Santa Clara, CA, USA) was used for the identification of functional groups on the MHS surface involved in nitrate adsorption.
Determination of pHpzc of MHS
The point of zero charge was determined according to Khan and Sarwar [
35]. Briefly, in a series of Erlenmeyer flasks (different for each pH of the solution), 0.5 g of MHS reacted with 20 mL of 0.01 M NaCl after adjusting the pH of the solutions (from 2 to 10) with NaOH or HCl. The flasks were shaken for 24 h at 25 °C and 130 rpm. Then, the solutions were filtered and the final pH was measured in each flask, and the difference between the initial and final pH (ΔpH) was calculated. The pH
pzc value was determined from the ΔpH versus pH
initial plot.
2.4. Batch Adsorption Experiments
The adsorption experiments were carried out in a batch-type procedure using a shaker water bath (Bioblock Scientific, Poly-test 20). The adsorption experiments were performed in different aqueous media, namely, a model nitrate solution (MS), a model (synthetic) wastewater (MW), and two real wastewater samples—one from the confectionery industry (CW) and one from the dairy industry (DW).
A description of the batch adsorption experiments is given in
Table 1, where
γnitrate is the initial nitrate concentration,
γadsorbent is the adsorbent concentration,
t is the contact time, and
V is the volume of the aqueous phase.
Samples (in Erlenmeyer flasks) were removed from the shaking water bath at predetermined time intervals. After filtration, the residual nitrate concentration was determined spectrophotometrically at 324 nm using a UV/Vis spectrophotometer (Specord 200, Analytic Jena, Germany).
The amount of nitrate removed was expressed as a percentage removal R (%) and calculated as follows:
where
γ0 is the initial nitrate concentration (mg L
−1) and
γ is the nitrate concentration after a predetermined contact time.
The amount of nitrate that was adsorbed onto MHS at equilibrium was calculated as follows:
where
qe is the amount of adsorbed nitrate (mg g
−1),
γ0 is the initial concentration of nitrate,
γe is the concentration of nitrate at equilibrium,
V is the volume of solution (L), and
m (g) is the mass of MHS. The effect of temperature on the adsorption process was studied at 25, 35, and 45 °C. Adsorption data were analyzed using the nonlinear form of Langmuir and Freundlich adsorption models, while kinetic data were analyzed using the pseudo-first order, pseudo-second order, and intraparticle diffusion models.
All experiments described above were performed in duplicates and proved to be reproducible.
2.5. Column Experiments
The column adsorption experiments were performed using fixed-bed column. The adsorbent MHS (1 g) was packed in a glass column (13 mm × 220 mm). Then, 2 L of a nitrate solution (30 mg L
−1) was continuously fed to the top of the column using a peristaltic pump (Masterflex L/S, Cole-Palmer Instrum Company, Vernon Hills, IL, USA) at a constant flow rate (10 mL min
−1) and at 25 °C and natural pH (i.e., the pH was not adjusted and was 6.3 (MS), 7.5 (MW), 5.7 (CW), and 9.4 (DW)). Samples (100 mL) were taken from the bottom of the column to determine nitrate concentration; 200 mL of 0.1 M NaCl and 500 mL of demineralized water were used to regenerate nitrate-loaded MHS (at flow rate of 10 mL min
−1) in situ. The Equation (3) was used to calculate the saturation capacity (
qs, mg g
−1):
where
γ0 is the initial concentration of nitrate (mg L
−1),
V0 is the initial volume of nitrate solution (L),
γ0 is the concentration of nitrate in fraction n (mg L
−1),
Vn is the volume of fraction n (L), and
m is the mass of MHS (g).