Use of Biowaste for Sodium Removal in Mediterranean Irrigation Water: A Sustainable Approach

Abstract

The Mediterranean region faces significant water scarcity, a challenge intensified by climate change, impacting both agricultural productivity and water quality. High sodium levels in irrigation water compromise soil structure, leading to reduced crop yields and economic strain. This study investigates the use of sustainable adsorbents derived from agricultural residues (almond shell, eggshell, and pumice) for the removal of sodium from irrigation water. These materials, widely available in the Mediterranean, support circular economy principles by repurposing biowaste to address agricultural challenges. Adsorption experiments were conducted using real irrigation water, capturing the complexity of its physicochemical properties to evaluate the effectiveness of these biosorbents under practical conditions. A Central Composite Rotational Design (CCRD) was applied to optimize adsorption parameters, focusing on adsorbent concentration, agitation, and contact time. Kinetic studies indicated that sodium adsorption adhered to a pseudo-second order model, suggesting a chemically controlled process. Isotherm analysis, with a strong fit to the Jovanovic model, confirmed a predominantly monomolecular adsorption mechanism across all adsorbents, while the Freundlich model highlighted site heterogeneity. Microscopy and energy-dispersive X-ray spectroscopy (EDX) revealed structural modifications in the adsorbents before and after treatment. The porous internal structure of the almond shell displayed significant sodium retention, while the calcified eggshell surface showed high initial adsorption efficiency but rapid site saturation. Pumice, noted for its extensive porosity, sustained adsorption capacity even with surface deposits formed during treatment. This research demonstrates the potential of biowaste-derived adsorbents for efficient sodium removal from complex aqueous systems, offering a viable solution for sustainable agriculture and improved soil and water management in Mediterranean regions.

1. Introduction

The Mediterranean region faces critical water challenges that are being exacerbated by climate change [1,2]. As temperatures rise and precipitation patterns become more irregular, the availability of both surface and groundwater is decreasing, directly affecting agriculture. It is estimated that the Mediterranean basin will experience an increase in the frequency and intensity of droughts, along with a reduction in annual average precipitation, threatening aquifer recharge and reducing river and reservoir flows [3,4]. This situation is especially concerning in a region where up to 80% of available water is allocated to agriculture, making irrigation essential to secure food production [5].

Aquifer overexploitation, driven by the growing demand for water in agriculture, has led to a significant decline in water quality [4,6]. Intensive groundwater extraction not only depletes resources but also contributes to saline intrusion in coastal aquifers, adversely impacting irrigation water quality and soil health. The use of low-quality water, characterized by high concentrations of salts such as sodium, is common in many agricultural areas of the Mediterranean, exacerbating soil salinization [1,5]. This phenomenon reduces soil permeability and hinders nutrient absorption by plants, limiting the growth of essential crops, such as cereals, vegetables, and fruit trees, and compromising the sustainability of agriculture in the region [7,8,9,10].

The degradation of irrigation water quality has a direct impact on agricultural production, resulting in significant economic consequences [11,12]. Agriculture, a key source of employment and income in rural Mediterranean areas, faces lower productivity, crop loss, and reduced product quality due to water scarcity [13]. This issue affects not only small farmers but also has repercussions on the agri-food industry, a vital sector in many Mediterranean economies. Furthermore, competition for water resources between agriculture, urban demands, and tourism has generated economic and social tensions, forcing farmers to adopt more efficient, despite being costly, irrigation technologies to remain competitive in an environment of increasing water scarcity.

In response to these challenges, the search for sustainable solutions has become a priority in the Mediterranean region. The circular economy offers a promising strategy, promoting the reuse of biowaste and the valorization of residual materials to improve irrigation water quality [14]. Using biowaste as adsorbent materials for sodium removal represents an innovative and economical way to address the problem of soil salinization [15,16]. In addition to being sustainable, these materials are widely available in the region, reducing dependence on external inputs and promoting more efficient local resource management.

Sodium removal from aqueous matrices employs a variety of mechanisms, including adsorption, precipitation, and ion exchange. For example, acid-activated local date pits demonstrated remarkable adsorption capacities for sodium, achieving a maximum specific surface area of 825.03 m² g⁻¹ and exhibiting efficiency under optimal conditions such as pH 9, an adsorption time of 90 min, and a sodium concentration of 600 mg L⁻¹. The process adhered to Freundlich isotherm and pseudo-second order kinetics, highlighting its high adsorption potential [17]. Similarly, sodium-modified vermiculite effectively adsorbed calcium ions through mechanisms validated by Langmuir and Freundlich isotherms, suggesting its applicability for water softening in sodium-dominant aqueous systems [18].

Another innovative approach uses the fixation properties of plant-inspired calcium phosphates to remove sodium salts such as NaCl, NaNO₃, and Na₂SO₄. This method achieves optimal efficiency at neutral pH and moderate temperatures, relying on ion interaction to form stable precipitates [19]. These studies demonstrate that the physicochemical properties of adsorbents, such as specific surface area and chemical modification, play crucial roles in enhancing sodium removal efficacy.

The diversity of techniques allows tailoring solutions to the specific composition and characteristics of wastewater. While activated carbons and zeolites dominate conventional approaches, these emerging alternatives underscore the innovation potential in environmental remediation technologies.

These conventional methods, although effective, are associated with high operational costs and significant energy consumption, making them less viable for small- and medium-sized farmers in the Mediterranean region [20,21]. Furthermore, the use of synthetic adsorbents such as activated carbons and zeolites, while efficient, requires complex and costly manufacturing processes that limit their sustainability and accessibility [18]. Another significant challenge is that most studies were conducted using synthetic or ultrapure solutions under controlled conditions that do not reflect the complexity of real waters, which contain mixtures of salts, organic matter, and various contaminants that affect the efficiency of the proposed methods [17]. Methods such as chemical precipitation or coagulation also generate secondary residues, such as sludge, which require additional treatment, increasing costs and environmental impact [19]. Finally, these approaches do not always consider integration with circular economy principles, missing the opportunity to use local materials or agricultural residues as sustainable resources [22]. These limitations highlight the need to explore more economical, sustainable alternatives that are better adapted to the real conditions of irrigation water in the Mediterranean region, as proposed in this study.

This study focuses on evaluating the sodium adsorption capacity of biowaste materials under real-world conditions, using untreated irrigation water with complex and variable physicochemical compositions. Unlike previous research conducted under controlled conditions with synthetic solutions, this study aims to assess the effectiveness of sorbents in a setting more representative of agricultural reality. The complexity of irrigation water, containing a varied matrix of salts, organic matter, and other contaminants, presents a significant challenge for evaluating sorbent efficiency, but it also facilitates better transferability of results to the agricultural sector.

The approach of this research contributes to promoting sustainable solutions in agricultural water resource management, providing a solid scientific basis for adopting low-cost and effective sodium removal techniques. Applying natural and sustainable materials, such as biowaste, not only addresses the urgent need to improve irrigation water quality but also aligns with circular economy principles, promoting more efficient resource use in the Mediterranean region. Validating these materials under real conditions ensures that the proposed solutions are viable for field implementation, thus supporting the resilience of the Mediterranean agricultural sector in the face of climate change and water scarcity challenges.

Therefore, this study aims not only to contribute to scientific knowledge in the field of contaminant adsorption in irrigation water but also to offer concrete, applicable solutions that enhance the sustainability and productivity of the Mediterranean agricultural sector.

2. Materials and Methods

2.1. Biowaste

In this study, three biowaste materials were selected as adsorbents for sodium removal in irrigation water: almond shells, eggshells, and pumice. The selection of these materials is based on several key criteria aligned with the principles of circular economy and sustainability in agricultural resource management [15,16].

The use of biowaste as adsorbent materials offers multiple advantages in terms of sustainability, economy, and availability [15]. In the Mediterranean region, agricultural residues such as almond shells and eggshells are abundant by-products generated on a large scale in the agri-food industry [23,24,25,26,27]. Valorizing these residues represents an efficient strategy to minimize environmental impact, reducing waste accumulation and transforming them into valuable resources. Pumice, a volcanic-origin material, is also abundant in many Mediterranean areas, facilitating its low-cost acquisition [28]. Its porous structure, which provides a high specific surface area, makes it a promising adsorbent for capturing ionic contaminants [29,30]. Additionally, the high local availability of these materials reduces dependency on external resources and minimizes acquisition and transportation costs. This strategy aligns with circular economy principles, promoting a zero-waste approach and material reuse in agriculture.

The biowaste materials were appropriately processed before use as adsorbents. The almond shells were washed with distilled water to remove surface impurities and dried at 60 °C for 24 h in a forced-air oven. They were then ground to achieve a uniform particle size [23,24]. For the eggshells, a similar cleaning process was applied, followed by drying at 105 °C for 4 h to ensure the removal of residual moisture. After drying, the eggshells were ground to a uniform particle size [26,27]. The pumice was washed to remove fine particles and adhered contaminants. Particles sized 2–5 mm were selected to maximize the specific surface area available for adsorption [29,31]. All samples were stored in sealed polyethylene bags until being used in the experiments.

2.2. Irrigation Water Samples

The irrigation water samples were collected from an irrigation pond located at the Escuela Politécnica Superior de Orihuela, part of the Miguel Hernández University (Orihuela, Spain). The choice to use these “real” waters for the experiments was driven by the need to assess the adsorption capacity of the materials under conditions representative of an agricultural environment. Unlike synthetic solutions, which allow precise control over contaminant concentrations, natural waters have a complex and variable composition that introduces additional challenges in adsorption processes. This approach ensures that the results obtained are directly applicable in the field, facilitating the transfer of conclusions to the actual agricultural sector.

Water samples were collected at accessible points around the irrigation pond in non-sterile polypropylene containers with a maximum capacity of 15 L. These samples were transported and maintained at a constant temperature of 4 °C until they were characterized and used in adsorption experiments. All irrigation water samples were characterized both before and after the experiments using Inductively Coupled Plasma Mass Spectrometry (ICP-MS, Model 2030, Shimadzu, Kyoto, Japan). This technique enabled the precise identification and quantification of elements and compounds present, providing a detailed assessment of the initial water composition and the changes induced by the adsorption process.

Throughout the experimental period, the composition of the irrigation water remained relatively constant, with some minor variations within an expected range due to natural environmental conditions. Calcium (Ca) levels averaged 81 ± 3 mg L⁻¹, while chloride (Cl) concentrations were at 188 ± 5 mg L⁻¹. Potassium (K) remained around 21 ± 2 mg L⁻¹, and sodium (Na) showed an average concentration of 42 ± 3 mg L⁻¹. Magnesium (Mg) was maintained at 27 ± 2 mg L⁻¹, and the pH fluctuated slightly within a range of 7.5 ± 0.1. This stability was attributed to the irrigation pond being filled to its maximum level at the start of the experiments, with minimal usage during the evaluation period, reducing variations due to recharge. Additionally, no significant precipitation events were recorded that could have affected the irrigation water composition during the study period.

The use of real irrigation water samples allowed for evaluating the adsorbent materials’ efficiency under conditions reflecting the natural complexity and variability of irrigation water, including the presence of salts, organic matter, and other contaminants typical of the Mediterranean agricultural environment. This approach enhances the relevance of the results obtained, ensuring that the study’s conclusions are directly applicable to real-world water management situations in agriculture, contributing to sustainable water resource management in the region.

2.3. Factor Design: Central Composite Rotatable Design (CCRD)

In this study, a Central Composite Rotatable Design (CCRD) was applied to optimize the sodium removal process using three sorbents: almond shell, eggshell, and pumice. This factorial design allows for the evaluation of interactions between multiple variables and precise modeling of adsorption system behaviors, reducing the number of necessary experiments and enhancing experimental efficiency [32,33,34].

The independent variables considered in this study were sorbent concentration (g L⁻¹), agitation speed (rpm), and contact time (min), while the dependent variable was sodium removal. These factors were chosen based on their fundamental roles in influencing adsorption processes, as supported by the existing literature and preliminary experimental trials [32,35,36,37]. Adsorbent concentration (g L⁻¹) was selected because the amount of adsorbent directly affects the availability of active sites for sodium adsorption. A higher concentration typically increases the probability of ion–sorbent interactions, enhancing adsorption efficiency, although excessive amounts may lead to aggregation and reduce the effective surface area. Stirring speed (rpm) was included as it influences the mixing efficiency and mass transfer rate of sodium ions to the adsorbent surface. Proper agitation minimizes boundary layer resistance, ensuring effective ion transport, while excessive stirring may disrupt adsorbent structures or cause desorption of adsorbed ions. Contact time (min) was considered crucial as it determines whether the adsorption process reaches equilibrium, allowing sufficient time for sodium ions to interact with the active sites. Identifying the optimal contact time is essential to avoid inefficiencies in practical applications. These factors were also selected because they consistently emerged as critical in prior studies on adsorption processes, and their interplay, analyzed through the CCRD, enables a comprehensive understanding of the adsorption system [35,36,38]. The CCRD matrix consisted of 17 experiments for each sorbent, with levels that included central, axial, and factorial points to cover the full range of variables (

Table 1. Factorial design matrix (2³) with levels for sorbent concentration (g L⁻¹), agitation speed (rpm), and contact time (min) for three sorbents.

Run	X1	X2	X3
1	1	1	−1
2	1	−1	1
3	−1	1	1
4	−1	−1	−1
5	0	0	0
6	1	−1	−1
7	0	0	0
8	1	1	1
9	−1	1	−1
10	−1	−1	1
11	−1.682	0	0
12	1.682	0	0
13	0	1.682	0
14	0	0	−1.682
15	0	0	0
16	0	−1.682	0
17	0	0	1.682

) [39].

Each sorbent was evaluated under combinations of these levels, including both central and axial points, enabling a comprehensive exploration of variable ranges to optimize the sodium removal process. To ensure reproducibility and minimize experimental error, four repetitions were conducted at the central point of the design. The axial points of the design (−1.682 and +1.682) were calculated using interpolation methods, ensuring that the experimental space encompassed the most effective ranges of the independent variables. This approach allowed for the full exploration of the experimental space, ensuring accurate predictions of sodium removal. The factor levels were established through an exhaustive literature review and preliminary trials to define the most appropriate ranges for the three sorbents (

Table 2. Levels of Central Composite Rotatable Design (CCRD) for almond shell, eggshell, and pumice in sodium removal.

Sorbent	Content (g L−1)	Agitation (rpm)	Contact Time (min)	References
Almond Shell	0.1–24	100–300	5–360	[40,41,42,43]
Eggshell	0.4–175.3	100–300	60–2880	[44,45,46,47,48]
Pumice	2–255.36	33–266	5–3446	[30,31,49]

).

During the experiments, the temperature was maintained constant at 25 ± 1 °C. A 100 mL volume of sodium solution was used in polypropylene flasks with a maximum capacity of 250 mL (Kartell Labware, Noviglio, Italy). After each trial, the solutions were filtered through cellulose acetate membranes (0.20 μm, 47 mm diameter, Teknokroma, Barcelona, Spain) to remove sorbents before measuring residual sodium. It is noteworthy that factors such as pH and temperature were not considered, as the experiments were conducted with real irrigation water. Any modification to these parameters would have altered the chemical composition of the samples, potentially compromising the applicability of the results to real-world conditions.

The data obtained from the experiments were analyzed using STATISTICA 18 software (StatSoft, TIBCO Software Inc., Palo Alto, CA, USA) through multiple linear regression to fit a second order polynomial model, including linear, quadratic, and interaction terms for each response variable. An analysis of variance (ANOVA) was used to identify statistically significant model terms, with a significance level of p < 0.05, ensuring a 95% confidence level [32]. Additionally, 3D response surface plots were generated to visualize interactions between the independent variables and their combined effects on sodium removal. These plots provided a clear identification of optimal conditions to maximize process efficiency [39].

Once the optimal experimental conditions were identified using the CCRD, the results were employed to conduct kinetic trials. These trials were performed under the optimal conditions for each sorbent to study the temporal dynamics of the sodium removal process. This allowed for the evaluation of kinetic behavior and a better understanding of the adsorption mechanisms involved, providing a comprehensive view of the factors influencing the sodium removal process in each system.

2.4. Sorption Kinetics

The kinetic sorption trials were conducted under the optimal conditions identified in the previous phase through the Central Composite Rotatable Design (CCRD). These conditions were selected to maximize sodium removal efficiency using three sorbents: almond shell, eggshell, and pumice. Natural irrigation water (50 mL) was used instead of synthetic solutions, due to the complex physicochemical properties of real water, providing greater applicability of the results in practical contexts.

Irrigation water samples were placed in non-sterile polypropylene beakers with a maximum capacity of 100 mL and maintained at a controlled temperature of 22 ± 1 °C throughout the experimental process. Microcosms were adequately covered to prevent dust or impurities from entering and to minimize sample evaporation. The sorbent amount, agitation speed, and total contact time were defined based on the CCRD results, as detailed in

Table 3. Experimental conditions for sodium sorption kinetic trials using almond shell, eggshell, and pumice.

Sorbent	Content (g L−1)	Agitation (rpm)	Contact Time (min)
Almond Shell	6.5	200	360
Eggshell	50.2	200	1500
Pumice	71	150	3264

. At predefined times, samples were collected for residual sodium concentration analysis, using the IMACIMUS multiparameter probe, which allowed for the precise and rapid measurement of sodium concentration in the filtered samples.

To describe and analyze the kinetic behavior of sodium removal, the experimental data were fitted to six widely used kinetic models: pseudo-first order, pseudo-second order, Elovich, intraparticle diffusion, Lagergren, and fractional order. These models provided a more detailed understanding of the predominant adsorption mechanism for each sorbent under the optimal conditions determined in previous experiments.

The pseudo-first order model [50,51] is based on the assumption that the adsorption rate is proportional to the difference between the amount of sodium adsorbed at equilibrium (q_e) and the amount adsorbed at time t (q_t). This model is suitable for describing processes where adsorption occurs rapidly at the start and then decreases as equilibrium is approached, which is common in physisorption. The equation for this model is as follows:

$$log\left(q_e-q_t\right)=\mathrm{l}\mathrm{o}\mathrm{g} q_e-{\displaystyle \frac{k_1}{2.303}}t$$

where q_e and q_t represent the amounts of sorbate retained (mg) per gram of sorbent at equilibrium, and time t, respectively, and k₁ (min) is the rate constant of the pseudo-first order. This model fits well when the adsorption process involves weak and rapid interactions, as is characteristic of physisorption.

The pseudo-second order model [52,53] assumes that the adsorption rate depends not only on the available free sites but also on those already occupied, suggesting a chemisorption-based control. In this case, the adsorption rate is related to the adsorption capacity at the active sites on the sorbent surface. The governing equation for this model is as follows:

$${\displaystyle \frac{t}{q_t}}={\displaystyle \frac{1}{k_2{q}_e^2}}+{\displaystyle \frac{t}{q_e}}$$

where q_e (mg g⁻¹) is the amount of sorbate adsorbed per gram of sorbent at equilibrium, q_t (mg g⁻¹) is the amount of sorbate adsorbed at time t, t (min) is the adsorption time, and k₂ (g mg⁻¹·min⁻¹) is the rate constant of the pseudo-second order model. This model generally provides a better fit for systems where chemisorption is the predominant mechanism, involving stronger chemical bonds.

The Elovich model [51] is primarily used to describe adsorption on heterogeneous surfaces, where adsorption energy varies across the sorbent surface. This model is appropriate for adsorption that occurs in multiple phases with varying energy levels. The equation for this model is as follows:

$$q_t={\displaystyle \frac{1}{\beta }}ln\left(\alpha \beta \right)+{\displaystyle \frac{1}{\beta }}ln\left(t\right)$$

where q_t (mg g⁻¹) is the amount of sorbate adsorbed at time t, t (min) is the adsorption time, α (mg g⁻¹·min⁻¹) is the initial adsorption rate, and β (g mg⁻¹) is the constant related to the extent of surface coverage and energy requirements of the process. This model is useful for processes where surface heterogeneity significantly influences adsorption kinetics.

The intraparticle diffusion model [35,54] considers that the adsorption process is limited by the diffusion rate of sodium ions within the sorbent particles. This model is helpful in evaluating whether diffusion is the rate-limiting step in the sorption process. The governing equation is as follows:

$$q_t=k_i{t}^{{\displaystyle \frac{1}{2}}}+C$$

where q_t (mg g⁻¹) is the amount of sorbate adsorbed at time t, t (min) is the adsorption time, k_i (mg g⁻¹·min^−(1/2)) is the intraparticle diffusion rate constant, and C (mg g⁻¹) is a constant reflecting mass transport resistance. This model is appropriate when adsorption is controlled by the rate at which ions diffuse to the active sites within the sorbent particles.

The Lagergren model [35] is a simplified version of the pseudo-first order model and was the first model proposed to describe liquid phase adsorption processes. The equation is as follows:

$$log\left(q_e-q_t\right)=log q_e-{\displaystyle \frac{k_{1}t}{2.303}}$$

This model, though similar to the pseudo-first order, is mainly used to describe physisorption processes where equilibrium is reached quickly, with less dependence on concentration conditions.

Finally, the fractional order model is applied in cases where adsorption does not strictly follow an integer order in kinetic terms. This model is useful when the adsorption mechanism is more complex and cannot be accurately described by integer order models. The corresponding equation is as follows:

$$q_t=k{t}^n$$

where k (mg g⁻¹·min⁻ⁿ) is the kinetic constant, and n (dimensionless) is the fractional order of the reaction, which can take non-integer values to reflect the complex nature of the adsorption process.

The experimental data were fitted to each of these kinetic models and evaluated using the coefficient of determination (R²) to determine which model provided the best fit to the experimental data.

2.5. Sorption Isotherms

The adsorption isotherm study was conducted following the determination of optimal experimental conditions obtained in the previous kinetic trials. Thus, the trials were performed under the best conditions of sorbent content and agitation speed, as determined by the kinetic results, with a fixed contact time of 24 h. To evaluate the behavior of the sorbents (eggshell, almond shell, and pumice) in sodium removal, various dilutions of real irrigation water and ultrapure water were used.

The solutions were prepared by mixing real irrigation water with ultrapure water, achieving four dilution levels, expressed as percentages of irrigation water (% v/v): 10%, 25%, 50% and 100%. These dilutions represented different levels of physicochemical complexity in the solutions, with real irrigation water serving as a source containing multiple dissolved elements that influence the adsorption process. The irrigation water used was neither chemically treated nor adjusted in terms of pH or temperature to maintain the natural field conditions.

For the experiments, 50 mL of each dilution were placed in 100 mL polypropylene beakers, and the amount of sorbent previously determined as optimal from the kinetic results was added. The sorbent concentration varied depending on the adsorbent material used and was defined according to the experimental planning. Throughout the experiment, the temperature was maintained constant at 22 ± 1 °C.

Each of the isothermal experiments was conducted in triplicate to ensure reproducibility and reliability of the results. Controls corresponding to each dilution without the addition of sorbents were used to measure the initial sodium concentrations in the solutions. The residual sodium concentration in the solutions was determined using the IMACIMUS multiparameter probe, which enabled quick and accurate measurements of sodium concentrations.

The experimental data were fitted to six isothermal models: Langmuir, Freundlich, Temkin, Dubinin–Radushkevich, Sips, and Jovanovic. Fitting to these models allowed for describing the relationship between sodium concentration and the adsorption capacity of the evaluated sorbents. The Langmuir model [53,55] assumes that adsorption occurs on a homogeneous surface with equivalent and non-interacting adsorption sites, describing a monolayer process where all active sites are occupied by one adsorbed molecule. This model is suitable for systems where adsorption saturates at a maximum level. Conversely, the Freundlich model [56] applies to heterogeneous surfaces, where adsorption sites have different energies, describing a multi-surface adsorption that does not reach a defined maximum, and is useful for heterogeneous systems.

The Temkin model [57] considers that the adsorption energy of all molecules in the layer decreases linearly with coverage due to sorbate–sorbent interactions, making it suitable for systems where adsorption energy varies during the process. The Dubinin–Radushkevich model [58,59] is useful for describing adsorption processes on microporous surfaces and differentiates between physical and chemical adsorption by evaluating the mean free energy of adsorption. The Sips model [60] is a combination of the Langmuir and Freundlich models, applicable to both homogeneous and heterogeneous surfaces, and is suitable for systems exhibiting saturation behavior at high adsorbate concentrations. Finally, the Jovanovic model [61,62] is similar to Langmuir but considers desorption effects in the advanced stages of the process, making it useful for describing monolayer adsorption on a homogeneous surface with the possibility of desorption.

The fitting of these models allowed for identifying the one that best described the sodium adsorption process on each sorbent, using nonlinear regression analysis to calculate the characteristic parameters of each isotherm model. This facilitated a comprehensive evaluation of adsorption capacity and characteristics of the different sorbents based on the various irrigation water dilutions.

2.6. Scanning Electron Microscopy and EDX Analysis

The morphological and compositional characterization of the sorbents used in this study (almond shell, eggshell, and pumice) was conducted using scanning electron microscopy (SEM). SEM analysis was performed on all pre- and post-treatment sorbent samples using the Sigma 300 VP field emission microscope (ZEISS, Jena, Germany). The equipment is equipped with a Schottky hot cathode, allowing for operation at voltages between 0.02 and 30 kV and obtaining high-resolution images, reaching up to 1.0 nm at 15 kV and 1.6 nm at 1 kV. Secondary electron (SE) and InLens detectors were used to capture surface images of the samples, maximizing spatial resolution and enabling detailed observation of surface microstructure. Additionally, the backscattered electron (BSE) detector was employed to contrast compositional differences on the sorbent surfaces.

The samples were mounted on metallic slides using conductive carbon tape to ensure proper electrical conductivity and prevent surface charging during observation. Images were acquired under high vacuum (HV) conditions, using accelerating voltages between 5 and 20 kV, depending on the material’s characteristics. Preliminary observations at low voltages (5 kV) were conducted on post-treatment samples to prevent potential damage to more delicate structures, gradually increasing the voltage to optimize resolution.

Furthermore, the elemental composition of the pre- and post-treatment samples was analyzed by energy-dispersive X-ray spectroscopy (EDX), using the EDX detector coupled to the Sigma 300 VP microscope. This analysis allowed for the identification and quantification of sodium and other elements present on the sorbent surfaces after exposure to irrigation water. Multiple areas of each sample were scanned to ensure the representativeness of the results. Experimental conditions for EDX spectrum acquisition included using accelerating voltages of 15 to 20 kV to ensure adequate excitation of elements present on the surface.

EDX analysis was performed on both the internal and external parts of the almond shell and eggshells, as well as on the entire surface of pumice. In each case, at least three representative areas of each sample were selected to obtain qualitative and semi-quantitative spectra, allowing for a comparison of the elemental composition before and after the sodium removal treatment.

3. Results

3.1. Central Composite Rotatable Design (CCRD)

The Central Composite Rotatable Design (CCRD) was employed in this study as a key statistical tool to optimize the sodium removal process using three sorbents: almond shell, eggshell, and pumice. The use of CCRD allowed for the simultaneous evaluation of multiple experimental factors and the identification of significant interactions between them, considerably improving the precision and robustness of the results compared to traditional trial-and-error methods [34,39]. This statistical approach also reduced the total number of experiments required, resulting in a more efficient design.

It is important to note that, in this study, parameters such as temperature and pH were not altered, as real irrigation water was used. Modifying these factors could have significantly changed the chemical composition of the samples, producing results that would not be applicable to real field conditions. Maintaining the original conditions of the water ensures that the results obtained accurately reflect the sorbents’ behavior under realistic conditions.

The three main factors analyzed in the study were sorbent concentration (mg L⁻¹), agitation speed (rpm), and contact time (min). These factors were selected for their influence on adsorption processes, and the levels of each variable were defined based on existing scientific literature. The CCRD statistical analysis was conducted independently for each sorbent, allowing clear conclusions to be drawn about the influence of each factor on sodium removal.

The statistical analysis for almond shell indicated that sorbent concentration was the most significant factor in sodium removal (F = 15.41, p < 0.0005), with a negative linear effect. This means that at higher almond shell concentrations, the sodium adsorption capacity increased significantly. However, agitation and contact time factors did not show a statistically relevant influence (p > 0.05). The interaction between factors was also marginal, suggesting that the removal process primarily depended on the number of active sites available on the sorbent. The maximum sodium removal point was reached at concentrations near the upper limit of the tested range (12 mg L⁻¹), indicating that once the active sites are occupied, increasing agitation or prolonging contact time does not significantly improve removal efficiency.

As with almond shell, concentration was also the most important factor for sodium removal with eggshell (F = 6.02, p < 0.004). Although the behavior of eggshell was similar to that of almond shell, its adsorption capacity was somewhat more moderate. This could be attributed to differences in the available active surface area or chemical composition between the two sorbents. Again, agitation and contact time factors were not significant (p > 0.05). The optimal concentration for eggshell was close to 100 mg L⁻¹, although the overall model showed a lower adjusted R², suggesting that other factors not considered in the analysis may influence the adsorption process.

In contrast to the other two sorbents, pumice showed a slight influence of both contact time and concentration (F = 4.09, p < 0.008). In this case, the interaction time between pumice and sodium appeared to slightly improve sodium removal, suggesting that adsorption equilibrium is not reached as quickly as with the other sorbents. However, as with almond shell and eggshell, agitation did not have a considerable impact. The optimal concentration of pumice for sodium removal was 140 mg L⁻¹, and although a slight influence of contact time was observed, the overall model yielded a low adjusted R², indicating that other important factors may be missing in fully describing this sorbent’s behavior.

3.1.1. Response Surface Analysis

The response surfaces generated provide a detailed view of how the experimental factors, specifically sorbent concentration, agitation, and contact time, influence sodium removal [63].

For almond shell, the surface plots indicate that sodium removal increased significantly with an increase in sorbent concentration. However, neither agitation nor contact time showed a notable effect on removal efficiency. The plots also suggest a slight interaction between concentration and contact time, where intermediate times seem to favor adsorption (Figure 1a).

For eggshell, the plots reflect a similar trend, where sorbent concentration was the predominant factor in sodium removal. Although agitation and contact time had a limited impact, the plots support the statistical analysis findings, confirming that concentration was the only statistically significant factor (Figure 1b).

In the case of pumice, the response surfaces indicated that both concentration and contact time contributed marginally to sodium removal. Unlike the other two sorbents, pumice showed a slight dependence on contact time, suggesting that longer times may improve adsorption in this case (Figure 1c).

The sorbent concentration was the most influential factor in all cases, as is consistent with the principles of adsorption by sorption. The availability of active sites on the sorbent surface is crucial for capturing sodium ions, and the surface plots show that removal efficiency consistently increased as concentration rose, reaching a peak at the highest levels of the range studied.

Although contact time was not significant in most cases, the surface plots indicate that in some instances, such as with pumice, a longer contact time may slightly improve adsorption. Conversely, agitation did not show a relevant effect for any of the sorbents, suggesting that diffusion in the liquid phase was not a limiting factor under the conditions studied.

3.1.2. Statistical Validation of the Models

To determine the statistical validity of the models, analysis of variance (ANOVA) was used at a 95% confidence level. A model is statistically significant when the p-value is less than 0.05, indicating that the factors evaluated have a real effect on sodium removal.

For almond shell, the model showed a satisfactory fit with an adjusted R² of 95%, indicating that the experimental data fit the model well. Concentration was the most significant factor (p < 0.0005), fully validating the model at the 95% confidence level.

However, for eggshell and pumice, although concentration was a significant factor, the overall models did not achieve a satisfactory fit at the 95% confidence level, suggesting that other factors or interactions might not be adequately captured by the model. Specifically, for eggshell, although the concentration was significant (p = 0.004), the overall model had a lower adjusted R² (21.8%), indicating that the model did not capture all the variability in the data and, therefore, does not fully validate the model at 95%.

Similarly, for pumice, the concentration was again significant (p = 0.008), but the overall model showed a low adjusted R² (26.5%), suggesting, as with eggshell, the possible absence of other relevant factors influencing the process.

The coefficient of determination (R²) and adjusted R² values were used to evaluate the fit of the CCRD model. For almond shell, the high R² (97.8%) and adjusted R² (95.0%) values indicate a strong correlation between the experimental and predicted values, validating the model and demonstrating its capacity to describe the data. However, for eggshell and pumice stone, the adjusted R² values were considerably lower (21.8% and 26.5%, respectively). These results suggest that while some factors were significant, the models do not fully capture the variability of the experimental data. This may indicate the influence of unconsidered factors or additional interactions affecting the adsorption process, highlighting the need for further refinement or exploration of other variables to improve the models’ predictive capacity.

3.2. Sorption Kinetic Study

To understand the mechanisms controlling the adsorption process, six kinetic models were evaluated: pseudo-first order, pseudo-second order, Elovich, intraparticle diffusion, Lagergren, and fractional order. The results obtained for each sorbent in relation to these models are discussed below. The sorption kinetic model graphs are presented in Figure S1, Supplementary Material.

The pseudo-first order model assumes that the adsorption rate is proportional to the amount of active sites still available on the sorbent [51]. This model showed a good initial fit for all three sorbents studied, with adjusted qe values of 0.027 mg g⁻¹ for eggshell, 0.277 mg g⁻¹ for almond shell, and 0.0255 mg g⁻¹ for pumice. The rate constants K₁ were 0.451 per min for eggshell, 0.907 per min for almond shell, and 46.25 per min for pumice, suggesting that initial adsorption was rapid. However, the model showed deviations at prolonged times, indicating that this mechanism cannot fully describe the adsorption process, especially during the equilibrium phase.

Conversely, the pseudo-second order model, which assumes that the adsorption rate is related to the square of the number of available active sites [52,53], provided the best fit for all three sorbents. This model was able to describe both the initial rapid adsorption phase and the system’s stabilization at equilibrium. For eggshell, the adjusted qe value was 0.026 mg g⁻¹, and the rate constant K₂ was 2.36 g mg⁻¹ min. In the case of almond shell, qe was 0.277 mg g⁻¹ and K₂ was 69.1 g mg⁻¹ min, while for pumice, the values were 0.0255 mg g⁻¹ and 8.4 g mg⁻¹ min, respectively. This consistent fit across all cases suggests that sodium adsorption in the three sorbents is controlled by chemical interactions rather than merely the availability of active sites, as proposed by the first order model.

The Elovich model [51], commonly used to describe adsorption processes on heterogeneous surfaces, was not representative of the adsorption process for any of the three sorbents. The parameters adjusted for this model were extremely high and unstable, indicating that sodium adsorption on these sorbents was not controlled by heterogeneous chemisorption. This reinforces the idea that the predominant mechanism is chemical and that the interaction with active sites on the sorbent is homogeneous in behavior.

The intraparticle diffusion model [54] was evaluated to determine if sodium adsorption was controlled by ion diffusion into the sorbent pores. In all three cases, this model did not show an adequate fit, suggesting that adsorption was not limited by an intraparticle diffusion process. The intraparticle diffusion constants, kid, were low, and the model fit was weak, indicating that adsorption occurred rapidly and primarily on the sorbent surfaces, without significant limitations from internal diffusion.

The Lagergren model [50], a variant of the pseudo-first order model, exhibited similar behavior, fitting well in the initial stages of adsorption for all sorbents. However, this model also deviated at prolonged times, reinforcing the conclusion that sodium adsorption is not fully controlled by a first order mechanism, as it failed to accurately describe system stabilization at equilibrium.

Finally, the fractional order model [64], which allows for describing complex systems where more than one mechanism may be involved, provided a reasonable fit in the early stages of the adsorption process. However, this model was also unable to accurately describe the long-term behavior. For eggshell and almond shell, the model adequately predicted the initial rapid adsorption but failed to capture system equilibrium. In pumice, the fit was similar, with a good initial fit but an inability to describe the stabilization phase of the process. These results suggest that the fractional order model does not adequately represent the main adsorption mechanism for the evaluated sorbents.

Overall, of the six models evaluated, the pseudo-second order model best described the sodium adsorption kinetics in all three sorbents. This model not only fit the experimental data well but also provides a mathematically coherent explanation of the adsorption mechanism, suggesting that the process is controlled by chemical interactions. The adjusted kinetic constants K₂ indicate rapid adsorption in the initial minutes and stabilization as active sites become saturated, as is consistent with a chemisorption adsorption process.

In contrast, the pseudo-first order, Lagergren, and fractional order models only fit the data in the early stages of the process but were unable to capture equilibrium. The Elovich and intraparticle diffusion models were not representative of the adsorption mechanism for any of the sorbents, suggesting that these were not the dominant processes in sodium removal.

The results for the three sorbents indicate that sodium adsorption is dominated by chemical interactions between sodium ions and active sites present in the studied biowastes. The decision not to vary temperature or pH in this study was justified by the intent to replicate real conditions of natural irrigation waters, where these parameters are not easily modifiable. Since irrigation waters usually present a complex and fluctuating chemical composition, it was considered more relevant to analyze sorbent behavior under these conditions, so that the results obtained can be applied directly without requiring additional pre-treatments that would increase operational costs for farmers. In this context, the pseudo-second order model remained the best representation of the adsorption mechanism under these uncontrolled conditions.

3.3. Sorption Isotherm

The isothermal analysis was conducted based on the optimal conditions identified in the previous kinetic study for each sorbent. These experimental conditions, derived from the kinetic trials, included the most suitable contact time, sorbent concentration, and agitation speed, which were selected to study sodium adsorption behavior using various isothermal models. The experimental data were fitted to six isotherm models: Langmuir, Freundlich, Temkin, Dubinin–Radushkevich, Sips, and Jovanovic, allowing for a detailed characterization of the adsorption process in eggshell, pumice, and almond shell.

For eggshell, the Jovanovic model best described the adsorption process, with a coefficient of determination R² = 0.9999. This strong fit demonstrates that sodium adsorption follows a monomolecular behavior, with potential desorption in the advanced stages of the process [62]. The Jovanovic model is particularly relevant in this case, as it effectively describes adsorption in systems where a dynamic equilibrium between adsorption and desorption plays an important role, especially at higher sodium concentrations.

The Freundlich model [62], useful for describing adsorption on heterogeneous surfaces, showed a poor fit with R² = 0.1845, indicating that the eggshell surface does not exhibit significant heterogeneity in adsorption sites. This confirms that the adsorption process was controlled by a direct and homogeneous interaction with the sorbent’s active sites, as is consistent with the findings from the Jovanovic model. The Langmuir and Temkin models also showed poor fits, suggesting that sodium adsorption on eggshell did not follow a typical monomolecular saturation pattern nor a linear decrease in adsorption energy as sites became saturated, ruling out both mechanisms as predominant.

In the case of pumice, the best fit was again obtained with the Jovanovic model, with an R² = 1, indicating that the sodium adsorption process was monomolecular, with possible desorption in the later stages. This result highlights the efficiency and stability of the adsorption process on pumice, especially at low sodium concentrations. The high affinity between sodium and the sorbent’s active sites reflects pumice’s microporous structure and physical adsorption capacity.

The Freundlich model showed a reasonably good fit (R² = 0.8004), suggesting that pumice has some heterogeneity in active sites, though this is not the predominant mechanism. The presence of micropores could explain its adsorption capacity, but these do not appear to be completely heterogeneous, as the Freundlich model was not the best fit. Conversely, the Dubinin–Radushkevich model [58] showed a very low fit (R² = 0.2061), indicating that adsorption was not primarily controlled by micropores or chemical interactions, but by a physical adsorption process with more homogeneous characteristics.

For almond shell, the results were consistent with a monomolecular adsorption process, with the Jovanovic model showing the best fit (R² = 0.9947). This fit demonstrates that adsorption was controlled by direct interactions between sodium and the active sites on the almond shell, with the possibility of desorption at higher concentrations. The fibrous and porous structure of almond shell facilitated a stable and efficient adsorption process.

The Langmuir model also showed a strong fit (R² = 0.9367), suggesting that, in the initial stages, adsorption followed a saturation pattern on a homogeneous surface. However, the good fit of the Freundlich model (R² = 0.9103) indicates the presence of adsorption sites with varying energies, which aligns with the heterogeneity of almond shell, allowing for adsorption at different energy levels. The Temkin model (R² = 0.9058) indicated that adsorption energy slightly decreased as active sites were occupied, a typical behavior in systems where there is interaction between the sorbate and sorbent and competition for adsorption sites at higher concentrations [57]. This behavior is consistent with the physicochemical properties of almond shell, where a mixture of active sites with different adsorption energies is present.

The isotherm results provide a deeper understanding of the adsorption mechanisms for each sorbent in sodium removal. The predominance of the Jovanovic model across all three sorbents indicates that the adsorption process was largely monomolecular, with desorption in the advanced stages. This underscores the stable and predictable nature of the sorbents in sodium adsorption, even in complex water matrices like those used in this study.

However, almond shell and pumice showed significant correlation with the Freundlich model, indicating the presence of heterogeneous adsorption sites that can influence process efficiency at varying sodium concentrations. These heterogeneous sites are particularly advantageous for adsorption in systems with variable sodium concentrations, making these sorbents suitable for field conditions.

The Langmuir and Temkin models, although useful for describing certain aspects of the process, were not predominant for any of the sorbents, indicating that adsorption was not limited to homogeneous saturation or a simple linear decrease in adsorption energy. This reflects the complexity of the adsorption process, which depends on both the sorbent structure and sorbate characteristics.

In addition to the high correlation observed in the Jovanovic models for all three sorbents, specific parameters were analyzed to gain a better understanding of the adsorption process. For eggshell, the Jovanovic model showed an adsorption constant (K_j) of 1.435 L g⁻¹, indicating a strong affinity between sodium and the sorbent’s active sites. The maximum adsorption capacity (q_max) was 0.161 mg g⁻¹, reflecting the sorbent’s capacity to retain sodium in the initial phases. These parameters, together with the coefficient of determination R² = 0.9999, confirm that adsorption was efficient and primarily monomolecular, with a low likelihood of significant desorption.

For pumice, the Jovanovic model showed an adsorption constant (K_j) of 1.512 L g⁻¹ and a maximum adsorption capacity (q_max) of 0.173 mg g⁻¹, reinforcing the high affinity of sodium with pumice’s active sites, along with a highly efficient adsorption behavior. The perfect fit with R² = 1 in this case suggests that the model precisely describes both the adsorption process and potential desorption in the later stages.

In the case of almond shell, the Jovanovic model revealed a K_j of 1.320 L g⁻¹ and a q_max of 0.158 mg g⁻¹, indicating high sodium adsorption efficiency. The R² coefficient of 0.9947 also shows that the process follows a monomolecular behavior, though with the possibility of some desorption at higher sodium concentrations. The porous and fibrous nature of almond shell facilitates this process, providing multiple active sites for adsorption.

The Freundlich model, which showed a reasonable fit for pumice (R² = 0.8004) and a strong fit for almond shell (R² = 0.9103), provided additional information on the heterogeneity of adsorption sites. For pumice, the value of K_f, representing relative adsorption capacity, was 0.144 L g⁻¹, while the parameter n, describing adsorption intensity, was 1.32. These results suggest that although variations in adsorption energy exist, the process is efficient and stable. For almond shell, the K_f value was 0.131 L g⁻¹ and n was 1.40, indicating greater heterogeneity in adsorption sites and higher efficiency at varying sodium concentrations.

The Langmuir model, applied to almond shell, showed a theoretical maximum adsorption capacity (q_max) of 0.167 mg g⁻¹ and a Langmuir constant (K_L) of 0.425 L mg⁻¹. These values, together with the coefficient of determination R² = 0.9367, indicate that adsorption followed a saturation pattern on a homogeneous surface in the early stages. As the energetically favorable sites became saturated, a transition to lower-energy sites occurred, as is consistent with the fit observed in the Freundlich model.

Finally, the Temkin model, which showed a good fit for almond shell (R² = 0.9058), provided a B value (related to adsorption heat) of 5.83 J mol⁻¹. This value suggests a slight decrease in adsorption energy as active sites were occupied, which is typical in systems where there is interaction between the sorbate and sorbent. This behavior reinforces the idea that adsorption in almond shell is controlled both by the availability of active sites and by energetic interactions between sodium and the sorbent.

3.4. Scanning Electron Microscopy (SEM) and EDX Analysis

The analysis using scanning electron microscopy (SEM) and energy-dispersive x-ray spectroscopy (EDX) was essential to assess the sodium adsorption capacity of the sorbents studied: almond shell, eggshell, and pumice. These techniques allowed for the observation of structural and compositional changes in each material before and after sodium removal treatment in irrigation water. The observed structural differences provide detailed insights into the efficacy of each sorbent and its behavior in a complex environment. The specific results for each material are presented below.

3.4.1. Almond Shell

The analysis of almond shell focused on comparing the differences between the external and internal parts before and after treatment (Figure 2 and Figure 3, respectively).

The external part, before treatment, displayed a fibrous structure with heterogeneously distributed cavities and micropores. Although rough, this surface exhibited lower porosity compared to the internal part, suggesting a more limited availability of active sites. The initial EDX analysis confirmed the absence of sodium, demonstrating the material’s purity prior to exposure to irrigation water. After treatment, SEM images revealed increased surface roughness and the accumulation of deposits within cavities. However, the EDX analysis did not detect significant amounts of sodium, indicating that the external part played a limited role in the adsorption of this ion.

The internal part of the shell presented a highly porous surface before treatment, with fissures, microcracks, and irregularly distributed micropores. This structure maximizes accessibility to active sites, facilitating interaction with ions in irrigation water. After treatment, notable changes were observed: there was an increase in deposit density and partial obstruction of some micropores. Additionally, structures compatible with bacteria appeared, suggesting possible biofilm formation encouraged by the biological complexity of irrigation water. This phenomenon may have positively contributed to sorbent efficacy, enhancing adsorption.

The post-treatment EDX results confirmed the presence of sodium exclusively in the internal part of the shell, highlighting its superior adsorptive capacity compared to the external part. The correlation between these findings and the statistical analysis performed with the Central Composite Rotatable Design (CCRD) showed that sorbent concentration was a key factor in sodium removal, aligning with the high density of active sites in the internal layer of the dried pericarp. Observations also matched the pseudo-second order kinetics, indicating that adsorption was dominated by chemical interactions. The porous structure observed before treatment provided an optimal environment for the formation of stable chemical bonds with sodium, which was supported by the reduced porosity once active sites were occupied.

Isotherm models, especially the Jovanovic model, confirmed that adsorption in almond shell was primarily monomolecular, as is consistent with the fibrous and microporous arrangement of the internal structure. The fit with the Freundlich model also highlighted the heterogeneity of adsorption sites, reinforcing the importance of the accessibility of micropores and microcracks in the internal part for sodium retention in a complex aqueous matrix.

3.4.2. Eggshells

As with almond shell, the differences between the external and internal parts of eggshell were evaluated, considering the calciferous nature of the material (Figure 4 and Figure 5, respectively).

Before treatment, the external surface showed a rough structure with scattered micropores and micro fissures, revealing a calciferous texture with uniformly distributed microcrystals. This arrangement suggested a high specific surface area, favorable for capturing ions like sodium. After treatment with irrigation water, SEM images showed increased roughness and the appearance of deposits in the micropores, indicating early saturation of active sites. The EDX analysis detected sodium in limited quantities in the external part, reflecting that the accumulation of adsorbed material quickly reduced the surface’s efficacy in continued sodium adsorption. The presence of calcium, phosphorus, and magnesium suggested competition between these elements and sodium for active sites.

In contrast, the internal part presented a more heterogeneous structure with compact zones combined with filamentous structural zones before treatment, with scattered micropores providing accessibility to active sites. However, after treatment, the surface became rougher, and some micropores appeared partially obstructed. This indicated that despite initial accessibility, the internal part may contain highly effective active sites, as confirmed by EDX, which detected higher sodium concentrations (up to 6.6% by weight). These results emphasized the sustained capacity of the internal part to adsorb sodium, even in complex irrigation water conditions.

The CCRD and kinetic analysis matched the morphological observations. Sorbent concentration was confirmed as the most relevant factor, especially for the internal part, whose micropore density allowed for efficient adsorption at high concentrations. The pseudo-second order kinetic model indicated that chemical interactions dominated adsorption, reflected in the modification of the calciferous structure in both the external and internal parts. As with almond shell, the fit of the Jovanovic isotherm model, with an R² coefficient close to 0.9999, supported the hypothesis of monomolecular adsorption, which is in line with the observed micropore saturation. The Freundlich model also showed that the heterogeneity in the distribution of active sites in the internal part was significant for the process, corresponding with the observed morphological variability after treatment.

3.4.3. Pumice

Finally, the morphological analysis of pumice highlighted its porous characteristics and the influence of treatment on its adsorptive capacity (Figure 6). Before treatment, pumice exhibited a highly porous surface with a dense network of micropores and mesopores. This structure offered a high specific surface area, ideal for ion adsorption. The rough texture, along with the presence of cavities and fissures, indicated great accessibility to active sites, suitable for efficient sodium adsorption. After treatment, SEM images showed that part of the porous structure was covered by deposits, suggesting the accumulation of salts and compounds from irrigation water. Despite this reduction in surface roughness in certain areas, the overall porous integrity was maintained.

EDX analysis confirmed pumice’s effectiveness in sodium adsorption, detecting up to 11.25% by weight in specific areas. The presence of oxygen, silicon, and calcium reflected the typical composition of pumice and the material’s ability to retain other ions present in irrigation water. The results were consistent with previous kinetic and isothermal studies. The Jovanovic model, with a perfect fit (R² = 1), corroborated the monomolecular nature of adsorption, while the Freundlich model indicated heterogeneity in active sites. The pseudo-second order kinetics supported the idea that adsorption was dominated by chemical interactions, as is consistent with the high sodium retention observed. The initial porous structure facilitated rapid adsorption, while deposit formation during the process reflected the saturation of more accessible sites and system stabilization at equilibrium.

4. Discussion

The results of this study highlight the effectiveness of biowastes—almond shell, eggshell, and pumice—in sodium adsorption in real irrigation waters, comparing them with natural adsorbent materials documented in the literature. The selection of these biowastes addresses the need for sustainable, low-cost solutions in agricultural water quality management, aligning with the global trend of minimizing the environmental impact of using commercial adsorbents [11,12]. For example, natural zeolites, which are commonly used for cation adsorption, exhibit sodium adsorption capacities ranging from 0.15 to 0.30 mg/g under similar conditions [65]. In comparison, almond shell achieved a capacity of 0.28 mg g⁻¹, pumice had a capacity of 0.17 mg g⁻¹, and eggshell had a capacity of 0.16 mg g⁻¹ in this study. These values indicate that biowaste-derived materials, despite being less processed, can offer competitive performance in sodium removal while promoting sustainability and reducing costs.

The removal efficiencies observed in this study were 29% for almond shell, 19% for eggshell, and 12% for pumice, calculated based on the initial and final sodium concentrations in the irrigation water. While these percentages may appear lower compared to capacities reported for synthetic solutions or highly processed adsorbents, they are consistent with the challenges posed by real water matrices. For instance, natural waters contain a complex mix of salts, organic matter, and competing ions, which can hinder the adsorption process. Despite these challenges, almond shell demonstrated a competitive performance, aligning with its higher adsorption capacity (0.28 mg g⁻¹), comparable to lignocellulosic materials such as rice husk and sugarcane bagasse. Eggshell and pumice, while showing lower removal percentages, offer advantages in terms of material availability and stability in real conditions. These findings highlight the importance of considering real water complexities in adsorption studies and suggest that biowaste materials remain viable options for sustainable water treatment in agricultural settings.

The use of biowaste for sodium removal was widely studied in synthetic solutions due to the experimental control simplicity they offer [57,58]. However, results obtained under real conditions are crucial to validating practical applicability in the Mediterranean agricultural context, where water quality is a constant concern. This study focused on using real irrigation waters, allowing the evaluation of adsorbent efficiency under conditions that reflect the complexity and variability of water in agricultural settings. Unlike synthetic solutions, real waters present variable chemical compositions that introduce ionic competition and other uncertainties that may influence the availability of active sites on the adsorbents [36,66,67,68,69,70]. This approach adds relevance and direct applicability to the results obtained.

The adsorption mechanism of almond shell involves interactions between sodium ions and the functional groups present in the lignocellulosic matrix, such as hydroxyl and carboxyl groups. These interactions include ion exchange and hydrogen bonding, which align with the predominance of chemical interactions suggested by the pseudo-second order kinetic model. Additionally, the porous structure enhances sodium ion accessibility to active sites, further improving adsorption efficiency in real water matrices. Almond shell showed high efficiency in sodium removal, similar to that reported for other lignocellulosic residues such as coconut shell and sugarcane bagasse, which have proven effective in cation adsorption due to their high density of functional groups and porous structure [70,71,72,73,74]. This pattern was also observed in rice husk, which, due to the presence of amorphous silica and hydroxyl groups, has shown comparable results in sodium adsorption [75]. However, in real waters, the presence of organic and inorganic compounds affects adsorption efficiency, though almond shell remains competitive against other materials studied, such as natural zeolite [76].

For eggshell, the initial sodium adsorption efficiency can be attributed to its calciferous structure, similar to that documented for carbonate-rich residues such as crustacean shells and animal bones [77,78]. Sodium adsorption occurs primarily through ion exchange with carbonate groups, a mechanism that is particularly effective in the initial stages of adsorption. However, in real irrigation waters, the high ionic competition with other cations, such as calcium and magnesium, limits the availability of active sites, leading to rapid saturation [79]. Despite these limitations, eggshell remains comparable with natural materials in small- to medium-scale studies, particularly in agricultural environments where cost-effective solutions are essential.

Pumice, on the other hand, stood out for its stability in adsorption capacity, even in the presence of complex chemical matrices in the water. This behavior aligns with that observed for other volcanic materials, such as natural zeolite, which has shown high selectivity for cations in previous studies [49,80]. The microporous and siliceous nature of pumice facilitates a dual adsorption mechanism that combines physical adsorption through its pore structure with chemical interactions at the surface. Electrostatic attractions between sodium ions and silanol groups are particularly significant, allowing pumice to maintain its performance even under challenging conditions, such as high ionic competition in real agricultural waters. This stability makes pumice an attractive option for long-term applications in agricultural water treatment, where maintaining adsorption efficiency under variable conditions is essential.

Comparison with other biowastes, such as banana peel and sugarcane bagasse, underscores the robustness of pumice in real irrigation waters, where ionic competition and organic matter interference can reduce adsorption capacity in less stable materials [31,40,81]. Pumice’s efficiency in this study suggests that it could be suitable for long-term applications in agriculture, where sorbent stability is essential.

The kinetic models used in this study, particularly the pseudo-second order model, fit the experimental data well, suggesting that chemical interactions predominate in adsorption, even in real irrigation waters. This result is consistent with previous studies reporting a higher adsorption rate in synthetic solutions due to the lack of ionic interferences [35,67,82]. In terms of isotherms, the fit to the Jovanovic model in all three biowastes indicates that adsorption follows a monomolecular behavior, a finding consistent with previous research on lignocellulosic and calciferous residues [80,82,83,84].

In practical applications, the stability and recyclability of adsorbents are critical factors that determine their long-term viability. Among the materials studied, pumice demonstrated significant stability in adsorption capacity, even under high ionic competition, making it suitable for repeated use in agricultural water treatment. Almond shell and eggshell, while effective in sodium removal, may face challenges related to structural integrity and regeneration efficiency after multiple cycles. Lignocellulosic materials like almond shell are prone to degradation in aqueous environments, while calciferous materials such as eggshell can undergo partial dissolution, reducing their effectiveness over time.

Recycling adsorbents often involves desorption processes using chemical or thermal treatments, which can be cost-prohibitive and environmentally taxing. These treatments may also alter the physicochemical properties of the adsorbents, affecting their performance in subsequent cycles. For pumice, its chemical inertness and robust structure suggest better recyclability compared to almond shell and eggshell. However, the efficiency of these materials in real agricultural settings would depend on developing cost-effective regeneration techniques and addressing the variability of water composition, including organic matter and competing ions. Further research is needed to optimize these processes and evaluate the economic feasibility of large scale implementation.

5. Conclusions

The efficiency observed in this study for almond shell and pumice, even under high ionic competition, underscores their viability for Mediterranean agriculture, where water quality is a constant concern. Unlike studies conducted under controlled conditions, the ability of these biowastes to maintain efficiency in real field scenarios confirms their potential to enhance agricultural water quality. This sustainable and cost-effective approach to water resource management could significantly contribute to the agricultural sector’s resilience in the face of climate uncertainties and increasing pressure on water resources.