Cleaner Approach for Atrazine Removal Using Recycling Biowaste/Waste in Permeable Barriers

: This work addresses the rehabilitation of water contaminated with atrazine, entrapping it in a permeable and sustainable barrier designed with waste materials (sepiolite) and with biomaterials (cork and pine bark). Atrazine adsorption was assessed by kinetics and equilibrium assays and desorption was tested with different extraction solvents. Adsorbed atrazine was 100% recovered from sepiolite using 20% acetonitrile solution, while 40% acetonitrile was needed to leach it from cork (98%) and pine bark (94%). Continuous ﬁxed-bed experiments using those sorbents as PRB were performed to evaluate atrazine removal for up-scale applications. The modiﬁed dose-response model properly described the breakthrough data. The highest adsorption capacity was achieved by sepiolite (23.3 ( ± 0.8) mg/g), followed by pine bark (14.8 ( ± 0.6) mg/g) and cork (13.0 ( ± 0.9) mg/g). Recyclability of sorbents was evaluated by adsorption-desorption cycles. After two regenerations, sepiolite achieved 81% of atrazine removal, followed by pine with 78% and cork with 54%. Sepiolite had the best performance in terms of adsorption capacity/stability. SEM and FTIR analyses conﬁrmed no signiﬁcant differences in material morphology and structure. This study demonstrates that recycling waste/biowaste is a sustainable option for wastewater treatment, with waste valorization and environmental protection. adsorption of atrazine (at 2 mg/L) onto the waste bioma-terials/materials. Atrazine concentration measured on a blank test remained unchanged during the experiment. Equilibrium is reached after 56 h for sepiolite, 72 h for cork and 8 days for pine bark, while for the powder-activated rubber char, this stage is attained after 14 h. The percentages of atrazine adsorbed at equilibrium time were approximately 22%, 36% and 50% for sepiolite, cork and pine bark at 5 g/L, respectively, and 100% for activated rubber char at 1 g/L. Results indicate that the adsorption of atrazine is initially fast (due to the availability of adsorption sites), and after a period of time the retention becomes slower, until equilibrium is reached. The most common kinetics models used to explain the sorption processes were used to evaluate the kinetic parameters, such as the PFO and


Introduction
Herbicides are a type of pesticides used to specifically target weeds and other unwanted plants. They are considered emerging contaminants that have become a threat to water supply network. Although they are found in effluents of wastewater treatment plants (WWTP) and in surface water at trace levels (normally ranging from ng/L to µg/L) [1], their fate, behavior and persistence in the environment became a threat to aquatic and terrestrial biodiversity and human health [2].
Atrazine (2-chloro-4-ethylamino-6-isopropylamino-1,3,5-triazine) is a recalcitrant compound and an organonitrogen herbicide extensively used to control broadleaf and grass weeds [3]. This herbicide does not tend to bioaccumulate, but its permanence in the environment and mobility in some soils can lead to surface and groundwater contamination [4]. In fact, atrazine is commonly found in drinking water at levels exceeding 0.1 µg/L [3][4][5][6] as a result of its massive use in agriculture. It is registered in more than 70 countries, making it the most commonly detected pesticide in surface water in the United States and also frequently detected in groundwater as well [7]. Its use has been banned among the European countries since 2003, due its potential threat to human health, its endocrine-disrupting nature and embryotoxic and embryolethal functions [3,4].
The Directive 2013/39 EU [8] states that atrazine is an environmental contaminant and a xenobiotic, considered one of the priority substances in the field of water policy justifying the urgency to develop cheaper and more effective water treatments to reduce its impact on the environment. The presence of herbicides in aquatic matrices is mainly

Preparation of Waste-Based Adsorbents/Biosorbents
Natural sepiolite was supplied by Tolsa S.A. located in Madrid, Spain. Firstly, the clay mineral was sieved to obtain a particle size with equivalent diameter between 1 and 2 mm and then it was washed with deionized water to remove dust. It was placed in an oven at 60 °C for 24 h and stored in a desiccator for further use.
Cork stoppers were supplied by a cork producing industry, Corticeira Amorim, SGPS, S.A. located in Aveiro, Portugal. The stoppers received were cut into small pieces and milled. The grains were sieved to a size with equivalent diameter between 1 and 2 mm and then washed with deionized water to remove dirt and particulate materials. The cleaned cork was placed in an oven at 60 °C for 24 h and then stored in a desiccator.
Pine bark was collected from a pine forest located in Braga, Portugal. It was washed with deionized water to remove mud, other materials and impurities and then placed in an oven at 60 °C to be dried. Then, it was ground into smaller particles, sieved in the range of 1 mm to 2 mm and stored in a desiccator.
Activated rubber char material was prepared using the pyrolysis of waste rubber granules provided from a truck tires and rubber waste recycling company, Eco-recycling LTD, located in Novi Sad, Serbia. Pyrolysis and activation processes were performed at the Department of Materials, Vinča Institute of Nuclear Sciences, Belgrade, Serbia. The waste rubber granules were pyrolyzed into a quartz tube furnace under an N2 flow at 800 °C for 4 h (heating rate was 5 °C/min). After this procedure, the char was washed with 2M HNO3 to clean the sample surface from metals. Carbonized granules and the solution of 2M HNO3 were mixed on a magnetic stirrer for 2 h at room temperature, filtered, washed with distilled water and dried in an oven at 110 °C. After that, the material was mixed with KOH pellets (sample/KOH = 1/4) and activated at 800 °C for 4 h under N2 atmosphere. In order to remove the impurities, the resulting activated carbon-based powder was washed with 0.5 M HCl solution (mixed in a magnetic stirrer at 85 °C for 30 min). Lastly, the sample was washed with deionized water and dried in an oven at 110 °C.

Materials Characterization
The pHzpc of materials was determined, as it helps to explain the electrostatic interactions between adsorbate and sorbent. A solution of 0.01 M NaCl was prepared, previously bubbled with nitrogen to avoid the dissolution of atmospheric CO2 and to prevent a change in the solution's pH. Solutions with pH values from 1 to 9 were prepared by adding 1.0 M H2SO4 or 1.0 M NaOH. For each pH value, 0.1 g of adsorbent was added to 25 mL of NaCl solution in 50 mL falcons. Flasks were sealed to avoid contact with air and were kept under moderate agitation in an incubator at 25 °C for 48 h. Afterwards, the adsorbent was filtered and the pH of the remaining solution was measured and plotted 215. 69

Preparation of Waste-Based Adsorbents/Biosorbents
Natural sepiolite was supplied by Tolsa S.A. located in Madrid, Spain. Firstly, the clay mineral was sieved to obtain a particle size with equivalent diameter between 1 and 2 mm and then it was washed with deionized water to remove dust. It was placed in an oven at 60 • C for 24 h and stored in a desiccator for further use.
Cork stoppers were supplied by a cork producing industry, Corticeira Amorim, SGPS, S.A. located in Aveiro, Portugal. The stoppers received were cut into small pieces and milled. The grains were sieved to a size with equivalent diameter between 1 and 2 mm and then washed with deionized water to remove dirt and particulate materials. The cleaned cork was placed in an oven at 60 • C for 24 h and then stored in a desiccator.
Pine bark was collected from a pine forest located in Braga, Portugal. It was washed with deionized water to remove mud, other materials and impurities and then placed in an oven at 60 • C to be dried. Then, it was ground into smaller particles, sieved in the range of 1 mm to 2 mm and stored in a desiccator.
Activated rubber char material was prepared using the pyrolysis of waste rubber granules provided from a truck tires and rubber waste recycling company, Eco-recycling LTD, located in Novi Sad, Serbia. Pyrolysis and activation processes were performed at the Department of Materials, Vinča Institute of Nuclear Sciences, Belgrade, Serbia. The waste rubber granules were pyrolyzed into a quartz tube furnace under an N 2 flow at 800 • C for 4 h (heating rate was 5 • C/min). After this procedure, the char was washed with 2 M hNO 3 to clean the sample surface from metals. Carbonized granules and the solution of 2 M hNO 3 were mixed on a magnetic stirrer for 2 h at room temperature, filtered, washed with distilled water and dried in an oven at 110 • C. After that, the material was mixed with KOH pellets (sample/KOH = 1/4) and activated at 800 • C for 4 h under N 2 atmosphere. In order to remove the impurities, the resulting activated carbon-based powder was washed with 0.5 M hCl solution (mixed in a magnetic stirrer at 85 • C for 30 min). Lastly, the sample was washed with deionized water and dried in an oven at 110 • C.

Materials Characterization
The pH zpc of materials was determined, as it helps to explain the electrostatic interactions between adsorbate and sorbent. A solution of 0.01 M NaCl was prepared, previously bubbled with nitrogen to avoid the dissolution of atmospheric CO 2 and to prevent a change in the solution's pH. Solutions with pH values from 1 to 9 were prepared by adding 1.0 M h 2 SO 4 or 1.0 M NaOH. For each pH value, 0.1 g of adsorbent was added to 25 mL of NaCl solution in 50 mL falcons. Flasks were sealed to avoid contact with air and were kept under moderate agitation in an incubator at 25 • C for 48 h. Afterwards, the adsorbent was filtered and the pH of the remaining solution was measured and plotted against the initial pH. The pH zpc is defined as the pH at which the curve crosses the line pH initial = pH final .
The surface morphology of adsorbents (before atrazine adsorption and after adsorptiondesorption cycles) was examined by Scanning electron microscopy (SEM) by a Phenom ProX scanning electron microscope (Phenom-World BV, Eindhoven, The Netherlands). All data were processed using the ProSuite software (Thermo Fisher, Boston, MA, USA). The observation of samples was performed with gold coating.
The presence of functional groups in the surfaces (in their natural form and after the first and the second desorption steps) was evaluated by attenuated total reflection Fourier-transform infrared (ATR-FTIR) spectroscopy using an ALPHA II-Bruker spectrometer (Ettlingen, germany) with a diamond-composite cell. The measurements were recorded over the range from 4000 to 400 cm −1 , with a resolution of 4 cm −1 and 24 scans per sample.

Quantification of Atrazine
Atrazine quantification was performed by UHPLC-DAD, using a Shimadzu Nexera X2 and a Kinetex C18 column (Phenomenex, Torrance, CA, USA) (1.7 µm × 100 Å × 2.1 mm), operating in gradient mode. Eluents used were acetonitrile and ultrapure water (45:55 v/v %) at a flow rate of 0.2 mL/min. The autosampler and column temperature was kept at 25 • C, and an injection volume of 5 µL was used. The chromatograms were registered at atrazine maximum adsorption wavelength at 225 nm. The linearity of the method was assured by setting a calibration curve over the concentration range of 0.3-4 mg/L. Detection limit (0.1 mg/L) was determined as the minimum detectable concentration of the analyte in the test sample that can be reliably distinguished from zero. The quantification limit (0.3 mg/L) was set as the concentration above which the analytical method operates with precision. The average peak area of each standard was used for quantification. Data were processed by Lab Solutions software (version 5.71, Shimadzu Corporation, Kyoto, Japan).

Batch Experiments
Kinetics and equilibria experiments were performed to evaluate the atrazine adsorption capacity of different adsorbents/biosorbents. A predetermined mass of each material was placed in amber Erlenmeyer flasks, and a fixed volume of the atrazine stock solution with 2 mg/L was added. Desorption experiments were carried out to evaluate the desorption capacity of atrazine from sepiolite, cork and pine bark using different extraction solvents. To assure the reproducibility of the results, each assay was performed in triplicate and the average of the three assays was reported for each test. The experiments were conducted in the absence of light to avoid the photo degradation of the herbicide [47].
Experiments with activated rubber char were performed as a comparative term. Due to the high material production costs, reduced availability and different particle size, only batch kinetics and equilibria tests were performed with this material.

Adsorption Kinetics
Assays were performed to determine the equilibrium adsorption time and to evaluate the adsorption rate of atrazine onto the sorbents. Adsorbent doses of 1 g/L of activated rubber char, and 5 g/L for sepiolite, cork and pine bark were used for equilibria assessment. Each mass of adsorbent was placed in contact with 50 mL of 2 mg/L of atrazine solution in Erlenmeyer flasks and these were shacked in a temperature-controlled incubator at 140 rpm and 25 • C until equilibrium was reached (14 h for the activated rubber char, 56 h for sepiolite, 72 h for cork and 8 days for pine bark). For the kinetics, samples were taken at different time intervals from another set of erlenmeyers, filtered through 0.22 µm nylon syringe filters, and the filtrate was analyzed for atrazine concentration by UHPLC-DAD. A blank test, with 2 mg/L of atrazine without biosorbent/adsorbent material, was performed in an Erlenmeyer flask and analyzed during the experiment. To determine the adsorption capacity, q t (mg/g), the following Equation (1) was used: where C 0 (mg/L) is the initial concentration of atrazine in solution, C t (mg/L) is the remaining concentration of atrazine in solution at time t, V (L) is the initial volume of the solution, and m (g) is the mass of the adsorbent material.
The experimental data were fitted by Lagergren pseudo-first order (PFO) [48] (Equation (2)) and pseudo-second order (PSO) [49] (Equation (3)) models to describe atrazine q t is the adsorption capacity at time t (mg/g), q e is the adsorption capacity at equilibrium time (mg/g), and k 1 (min −1 ) and k 2 (g/mg.min) are the adsorption rate constants of PFO and PSO, respectively.

Adsorption Equilibrium
Once the equilibrium time for each sorbent was established, equilibrium experiments were performed to build up the adsorption isotherms and to determine the maximum uptake value for each material. Activated rubber char concentrations tested ranged between 0.05 and 0.75 g/L, while for sepiolite, cork and pine bark they ranged from 1 to 15 g/L. Different amounts of each adsorbent were placed in contact with 2 mg/L of atrazine solution in amber Erlenmeyer flasks. The flasks were kept under continuous stirring at 140 rpm and 25 • C, during the period of time needed to attain the equilibrium. Samples were then taken, filtered and analyzed by UHPLC to determine the remaining atrazine concentration. The adsorption capacity of sorbents at equilibrium q e (mg/g) was determined using Equation (1), where C 0 is substituted by C e (mg/L), equilibrium concentration of atrazine in solution.
Freundlich [50] and Sips [51] isotherm models (Equations (4) and (5), respectively) were applied to equilibrium results to evaluate adsorption behavior: q e = q max ·K s ·C e 1 ms K S is the affinity constant related to the adsorption energy (L/g) and m s is a parameter related with the heterogeneity of the system.

Desorption Behavior
Batch assays were performed using 0.75 g of sorbents in 50 mL of 2 mg/L atrazine solution (solid/liquid ratio of 15 g/L) at 25 • C, during the time needed to reach equilibrium for each material (56 h for sepiolite, 72 h for cork and 8 days for pine bark). Then, samples were taken, filtered and analyzed by UHPLC to quantify the atrazine remaining in solution. The suspensions were filtered, the materials were recovered and dried at 30 • C. Atrazine desorption assessment was then conducted by using the following extraction solvents: acetate buffer at pH 5, potassium phosphate buffer at pH 7, aqueous solutions with 5%, 20% and 40% (except for sepiolite) acetonitrile at 25 • C and deionized water at 25 • C, 40 • C and 60 • C. In detail, sorbents loaded with atrazine were mixed with 50 mL of each extraction solvent (using 15 g/L of sorbent concentration) under continuous agitation at 140 rpm for 24 h. Liquid samples were collected at regular intervals and filtered, and the efficiency of atrazine desorption was evaluated through UHPLC analysis as described previously. The desorption percentage, D (%) was calculated by using Equation (6): Here, C des is the concentration of desorbed atrazine along the time and C ads the initial concentration of adsorbed atrazine on adsorbent.

Column Adsorption Experiments
Column adsorption using a PRB at lab scale in continuous flow to treat 2 mg/L aqueous solution of atrazine allowed to determine operational parameters as breakthrough and exhaustion time, useful for process design and upscale. Regeneration and reusability assays and PRB simulation with sepiolite, cork and pine bark as supports were performed in column systems at lab scale. Consecutive adsorption of atrazine from 2 mg/L solution and desorption processes (using the extraction solvent with the best desorption capacity obtained for each material) were carried out to establish the adsorption capacity profile of each material after several adsorption-desorption cycles. To assure the reproducibility of the results, each assay was performed in duplicate and the average of both assays is reported for each test. The experiments were conducted in the absence of light to avoid the photo-degradation of the herbicide.

Regeneration and Reusability
The viability of adsorbents depends on their effective reuse, so the evaluation of adsorbent recyclability becomes necessary. The operation of a column with a fixed barrier for cyclic adsorption/desorption processes is an effective technique for the purpose, as it makes the best use of the concentration difference driving force for adsorption and a more efficient utilization of the adsorbent capacity. The repetitive adsorption-desorption cycles were performed to determine the loss in the uptake capacity of the adsorbents. 20 g of each material was placed inside an acrylic column (height: 30 cm, internal ø 4 cm) and a predetermined volume of 2 mg/L atrazine solution (using a solid/liquid ratio 15 g/L) was added. The pollutant influent was introduced into the column from the bottom to the top in closed-loop mode at a controlled flow rate of 1 mL/min for 96 h, 72 h and 83 h for sepiolite, cork and pine bark, respectively. When equilibria were reached, column operation stopped and the excess atrazine solution inside the columns was withdrawn. The regeneration of the adsorbent by desorption processes initiated in continuous mode over 48 h with the selected extraction solvents, aqueous solution with 20% acetonitrile for sepiolite and 40% acetonitrile solution for cork and pine bark. Materials were regenerated twice and reused another 2 times. During adsorption and desorption cycles, samples were collected periodically from the outflow of the column to determine atrazine concentration by UHPLC. To assure the reproducibility of the results, each assay was performed in duplicate and the average of two assays was reported for each test.

PRB Simulation at Lab Scale
The use of PRB in a continuous flow column is a practical and efficient technique to remove pollutants from real wastewater. The ability of the waste-based sepiolite, cork and pine bark to function as effective PRB for the removal of atrazine at 2 mg/L was evaluated in open system experiments (continuous flow) at a laboratory scale. The continuous column experiments were operated using three individual acrylic columns (height: 30 cm, internal ø 4 cm), each one filled with 20 g of each material. Atrazine solution was fluxed upwards through the columns using a peristaltic pump at 1 mL/min for 76 h without recirculation. Periodically, samples were collected in the outflow to determine atrazine concentration by UHPLC until exhaustion time was reached. The assays were conducted in duplicate and the results are an average of the duplicates. Breakthrough curves were obtained by plotting the normalized concentration values (C t /C 0 ) versus time.
Adsorption dynamics acquaintance and modelling are essential because they provide valuable information on some practical aspects such as sorbent capacity and prediction of the time necessary for the effective operation of a PRB column. Among the different mathematical models that can be used to describe the removal behavior of a certain pollutant, the modified dose-response [52] (Equation (7)) and Yoon-Nelson [53] (Equation (8)) models were selected to predict breakthrough curves and to determine the characteristic parameters of the column that are useful for process design, with nonlinear regression fitting. In this model, α is the parameter that is associated with the shape of the breakthrough curve. When α ≤ 1, the model displays a parabola-like curve, and for α > 1 the curve assumes a sigmoidal shape [28].
Q is the flow rate (mL/min), q 0 is the adsorption capacity of the bed (mg/g), m B is the mass of the adsorbent (g) and α is a dimensionless empirical constant.
is the time required to retain 50% of the initial adsorbate and t is the operating time (min).

Statistical Analysis
GraphPad Prism ® software (version 8.0; graphPad Software, Inc., San Diego, CA, USA) was used for statistical analyses. The level of significance was determined by two-ways ANOVA followed by Tukey's test for multiple comparisons. Significance was accepted at p < 0.05.

Characterization of the Adsorbents/Biosorbents
The pH zpc is of fundamental importance in surface science applied to environmental rehabilitation, since it allows to determine the ability of a substrate to retain the molecules of interest. This parameter is relevant in adsorption processes since it influences the solution pH and the electrostatic interactions between adsorbate and adsorbent. This parameter is described as the pH at which the net charge of the adsorbent surface is equal to zero. At a pH lower than pH zpc , the surface of the adsorbent is positively charged, favoring anions species to be adsorbed. On the opposite, when the pH is higher than pH zpc , the surface of the adsorbent is negatively charged and cations will be preferentially adsorbed. In Figure 1, the results obtained for pH zpc determination by the pH drift method are shown. The pH zpc is identified as the point where the line pH initial = pH final crosses the experimental pH curve. Sepiolite presents a pH zpc value around eight, while for cork and pine bark the obtained pH zpc values are three and five, respectively. Atrazine is considered a very weak base, with a pKa value of 1.68 and an isoelectric point (pI) of 9.23. Under the conditions of the study, the pH of atrazine solution at 2 mg/L has a value of 5.7. When atrazine molecules are in contact with the sorbent materials, their pH changes because of the ionic interactions between adsorbate and adsorbent. Atrazine is considered a very weak base, with a pKa value of 1.68 and an isoelectric point (pI) of 9.23. Under the conditions of the study, the pH of atrazine solution at 2 mg/L has a value of 5.7. When atrazine molecules are in contact with the sorbent materials, their pH changes because of the ionic interactions between adsorbate and adsorbent. Figure 2 shows the charge distribution of atrazine at different pH levels. Increasing pH led to progressive deprotonation of the functional groups from the surface. Atrazine molecules are in the neutral form in the range of pH between 6.5 and 11.5. Atrazine is positively charged for pH values lower than 6.5, and negatively charged at pH higher than 11.5. Atrazine is considered a very weak base, with a pKa value of 1.68 and an isoelectric point (pI) of 9.23. Under the conditions of the study, the pH of atrazine solution at 2 mg/L has a value of 5.7. When atrazine molecules are in contact with the sorbent materials, their pH changes because of the ionic interactions between adsorbate and adsorbent. Figure 2 shows the charge distribution of atrazine at different pH levels. Increasing pH led to progressive deprotonation of the functional groups from the surface. Atrazine molecules are in the neutral form in the range of pH between 6.5 and 11.5. Atrazine is positively charged for pH values lower than 6.5, and negatively charged at pH higher than 11.5. According to pHzpc study, cork and sepiolite surfaces assume a negative charge when in contact with atrazine solution, once the solution pH is higher than their pHzpc. On the other hand, pine bark surface becomes neutral, once solution pH is at the same pHzpc value. By ion-exchange mechanism, the atrazine molecule is favorably adsorbed by cork. Table 2 summarizes the charge distribution interactions of sorbent materials, when in contact with atrazine solution. According to pH zpc study, cork and sepiolite surfaces assume a negative charge when in contact with atrazine solution, once the solution pH is higher than their pH zpc . On the other hand, pine bark surface becomes neutral, once solution pH is at the same pH zpc value. By ion-exchange mechanism, the atrazine molecule is favorably adsorbed by cork. Table 2 summarizes the charge distribution interactions of sorbent materials, when in contact with atrazine solution. SEM has been used as a primary tool to characterize the surface morphology and the essential physical properties of the adsorbents, such as porosity and structure. Sepiolite, pine and cork (before adsorption experiments and after two regeneration cycles), as well as activated rubber char samples were analyzed by SEM (Figure 3). Raw sepiolite ( Figure 3A) exhibits a micro-fibrous texture, with inter-fibers/bundles spaces. Its textural properties were reported in previous works by this research group: surface area (S BET ) of 181.35 m 2 /g, pore volume of 0.30 cm 3 /g and average pore size of 65.8 Å [28]. Figure 3C shows a radial view of a ray of Pinus pinaster bark constituted of small pits. Quercus suber ( Figure 3E) shows a porous surface formed from adjacent hollow polyhedral prismatic cells, similar to a honeycomb structure. Cork and pine bark present a very low specific surface area (<4 m 2 /g [15]). Figure 3G exhibits a particle of rubber char (dark area) activated with KOH (white area). This activated char is essentially mesoporous, with a high S BET value and pore volume. For adsorbents with small micropores, the adsorption mechanism is mainly chemical, once a stronger binding of the adsorbate to the adsorbent pore wall occurs. For larger molecules that cannot access to pores, this enhanced adsorption does not occur. As the pores increase in size, adsorption becomes continuously a physical (or multi-layer) process [46]. The morphology of adsorbent/biosorbent after sorbate second leaching was also observed in order to evaluate the effect of those processes on it. Sepiolite ( Figure 3B) becomes more fibrous, with a rougher surface, meaning that the exposure to the extraction solvent (aqueous solution with 20% of acetonitrile) may have worn out its surface, becoming more polished. For pine bark ( Figure 3D) and cork waste ( Figure 3F) differences in morphology were not observed. The ATR-FTIR spectra of sepiolite, cork and pine bark in their natural form, as well as after the first and second desorption steps, were recorded in the region of 400-4000 cm −1 and are shown in Figures 4-6, respectively. Analyses were performed to evaluate possible The ATR-FTIR spectra of sepiolite, cork and pine bark in their natural form, as well as after the first and second desorption steps, were recorded in the region of 400-4000 cm −1 and are shown in Figures 4-6, respectively. Analyses were performed to evaluate possible changes on surface functional groups after the regeneration process, and therefore to clarify the atrazine adsorption mechanism. In Figure 4, the spectrum of raw sepiolite is shown and the characteristic stretching vibration of -OH groups attached to octahedral Mg 2+ located in the interior blocks are evidenced, at 3690 cm −1 . At 3581 cm −1 , the Si-O band is verified. From 3000 to 4000 cm −1 , the broad band is due to h-O-H vibrations. In addition, the band at 1658 cm −1 is related to the -OH bond and at 1197 cm −1 , as well as at 940 cm −1 , vibrations can be attributed to the Si-O combination. The 1024 cm −1 band corresponds to the Si-O-Si vibrations, and the one at 450 cm −1 to the Si-O-Mg link, and, at 653 cm −1 , there is the bending vibration of Mg-OH. Sepiolite impurities derived from dolomite are also detected at 1443 cm −1 [54,55]. Comparing the bands between the spectra of the raw sepiolite with the ones after exposure to the desorbing solvent, for both leaching steps, no substantial differences were detected (the 1443 cm −1 band from dolomite is the one that does not appear). So, it can be considered that the chemical structure of sepiolite is not affected after two regeneration cycles. Natural cork spectra ( Figure 5) display several absorption peaks, indicating the complex nature of this biomaterial. The band at 3400 cm −1 represents -OH groups, and the bands observed at 2923 cm −1 to 2853 cm −1 could be assigned to the C-H stretch. The peaks around 1760 cm −1 and 1638 cm −1 correspond to the C=O stretching. The absorption peak resulting from the C-O bond is observed at 1036 cm −1 [56]. The principal constituents of Portuguese Quercus suber L. can be identified in these cork spectra as suberin, at 2923, 2853 and 1760 cm −1 , guaiacyl-type lignin at 1513, 858, and 816 cm −1 , polysaccharides at 1100 and 1036 cm −1 and extractives at 1603 and from 1457 to 1300 cm −1 [57]. Comparing the bands between the spectra of the cork in its natural form and after two desorption processes, the initial peaks are still observed, meaning that cork material after 2 regeneration cycles remains chemically unalterable, not affected by the desorption processes. As it can be seen in Figure 6, raw Pinus pinaster bark spectrum evidences at 3400 cm −1 the stretching vibration of the bond between oxygen and hydrogen. The bands at 2920 cm −1 and 1450 cm −1 can be assigned to the stretching of the C-H bond in the aromatic and aliphatic structures. The bands at 1610 cm −1 and 1510 cm −1 are attributed to the aromatic C-C skeletal vibrations, and the band around 1733 cm −1 results from the stretching vibrations in C-O carbonyl structure. At 1160 cm −1 and 1210 cm −1 , the stretching of C-O-C in the cellulose and hemicellulose is detected [58,59]. The phenolic -OH and aliphatic -CH stretch vibrations can be identified at 1364 cm −1 . Vibrations of -CO and aromatic -CH are detected at 1100 cm −1 . The 860, 820 and 773 cm −1 bands indicate the -CH bending vibrations from aromatic rings, related to phenolic compounds [60]. In this spectrum, the distinct signal at 1036 cm −1 is identified, produced by the aliphatic C-O stretching [58]. When comparing raw pine spectra with the ones obtained after two desorption steps, no substantial differences occur, in contrast to the other materials.

Adsorption Kinetics Modelling
The assessment of adsorption kinetics can support the definition of the adsorption pathway and probable mechanism involved. Figure 7 shows the experimental data and the fitting kinetics models for the adsorption of atrazine (at 2 mg/L) onto the waste biomaterials/materials. Atrazine concentration measured on a blank test remained unchanged during the experiment. Equilibrium is reached after 56 h for sepiolite, 72 h for cork and 8 days for pine bark, while for the powder-activated rubber char, this stage is attained after 14 h. The percentages of atrazine adsorbed at equilibrium time were approximately 22%, 36% and 50% for sepiolite, cork and pine bark at 5 g/L, respectively, and 100% for activated rubber char at 1 g/L. Results indicate that the adsorption of atrazine is initially fast (due to the availability of adsorption sites), and after a period of time the retention becomes slower, until equilibrium is reached. The most common kinetics models used to explain the sorption processes were used to evaluate the kinetic parameters, such as the PFO and PSO. For both models, the experimental q e values are in agreement with the calculated ones. Although both models fit experimental data with high correlation coefficient values (R 2 ), the PSO model showed the best fit for the generality of the sorbents. Comparing the k 2 value (Table 3) for the studied materials, the adsorption rate of atrazine follows this sequence: sepiolite > cork > rubber char > pine bark. The waste-based materials without pretreatment, sepiolite, cork and pine bark have the capacity to retain atrazine effectively, in their natural form, without resorting to chemical pretreatments.

Adsorption Kinetics Modelling
The assessment of adsorption kinetics can support the definition of the adsorp pathway and probable mechanism involved. Figure 7 shows the experimental data the fitting kinetics models for the adsorption of atrazine (at 2 mg/L) onto the waste materials/materials. Atrazine concentration measured on a blank test remained changed during the experiment. Equilibrium is reached after 56 h for sepiolite, 72 h cork and 8 days for pine bark, while for the powder-activated rubber char, this stag attained after 14 h. The percentages of atrazine adsorbed at equilibrium time were app imately 22%, 36% and 50% for sepiolite, cork and pine bark at 5 g/L, respectively, 100% for activated rubber char at 1 g/L. Results indicate that the adsorption of atrazin initially fast (due to the availability of adsorption sites), and after a period of time retention becomes slower, until equilibrium is reached. The most common kinetics mo used to explain the sorption processes were used to evaluate the kinetic parameters, s as the PFO and PSO. For both models, the experimental qe values are in agreement w the calculated ones. Although both models fit experimental data with high correla coefficient values (R 2 ), the PSO model showed the best fit for the generality of the sorbe Comparing the k2 value (Table 3) for the studied materials, the adsorption rate of atra follows this sequence: sepiolite > cork > rubber char > pine bark. The waste-based ma als without pretreatment, sepiolite, cork and pine bark have the capacity to retain atra effectively, in their natural form, without resorting to chemical pretreatments.
The development of materials with high carbon content, such as the rubber char, w mesoporosity, is important to achieve higher adsorption capacities, as well as faster sorption kinetics. In addition, the particle size of an adsorbent also plays a significant in adsorption capacity and rate. This rate tends to increase when particle size decrea by increasing mesopore volume and SBET [61] and reducing interior diffusional rate l tations. In this study, the rubber char particle has a substantially smaller size than other materials and is, in fact, the one that achieves an equilibrium stage faster and w a higher uptake value, suggesting that atrazine can be easily adsorbed on the outer sur of the material [62].

Adsorption Equilibrium Modelling
Adsorption isotherms are useful for establishing the adsorption capacity of sorbents The development of materials with high carbon content, such as the rubber char, with mesoporosity, is important to achieve higher adsorption capacities, as well as faster adsorption kinetics. In addition, the particle size of an adsorbent also plays a significant role in adsorption capacity and rate. This rate tends to increase when particle size decreases, by increasing mesopore volume and S BET [61] and reducing interior diffusional rate limitations. In this study, the rubber char particle has a substantially smaller size than the other materials and is, in fact, the one that achieves an equilibrium stage faster and with a higher uptake value, suggesting that atrazine can be easily adsorbed on the outer surface of the material [62].

Adsorption Equilibrium Modelling
Adsorption isotherms are useful for establishing the adsorption capacity of sorbents and for describing their interaction with the adsorbate. In this study, the relation between q e and C e was fitted by several different isotherms for water and wastewater treatment, and just the best fitting ones for the selected adsorbents/biosorbents are presented in Table 4, namely Freundlich and three-parameter Langmuir-Freundlich (Sips). The nonlinear regression was used in this analysis to avoid the errors associated with the commonly used linearized models. Figure 8 shows the results obtained from the equilibrium adsorption experiments fitted with the isotherm models. According to Table 4, the Freundlich model is the one that best fits the isothermal adsorption onto cork, sepiolite and activated rubber char, while the Sips model is the one that more adequately fits the isothermal adsorption onto pine bark. The Freundlich model is suitable for heterogeneous adsorbent surfaces, and it can be applied to non-ideal or multilayer adsorption by surfaces of varied affinities. n F and K F are Freundlich constants which denote the adsorption intensity and capacity, respectively. When n F < 1 the adsorption process is chemical, but if n F > 1, adsorption becomes a favorable physical process [62]. According to Table 4, sepiolite and rubber char present n F > 1, suggesting a physical adsorption process, while cork (with n F < 1) retains atrazine chemically. The highest K F value was attained for rubber char, demonstrating its higher adsorption capacity. The Sips model results from the combination of Langmuir and Freundlich, and predicts a heterogeneous adsorption system, thereby avoiding the limitations associated with the increased adsorbate concentration. When the concentration of the adsorbate is low, the model reduces to Freundlich isotherm, while for high concentrations, it predicts a monolayer adsorption process on a homogeneous surface (Langmuir model) [63]. The m S parameter gives information about the degree of surface non-homogeneity. heterogeneous adsorbents present m S values close to 0, while m S values closer to 1 indicate a material with relatively homogeneous binding sites [64]. Pine bark presents an m S value closer to 1, suggesting that its binding sites have the same adsorption affinity, homogeneously distributed over its surface. In terms of maximum adsorption capacities, q max , activated rubber char, is the one that presents the highest value, followed by cork, sepiolite and pine bark, sequentially. Rubber char is the adsorbent with the highest atrazine adsorption capacity since it has enhanced physical properties, as described in the previous section (higher surface area and micropore volume). It is reported that adsorbents/biosorbents without any kind of physical or chemical treatment normally present lower adsorption capacities when compared with the commercial/synthetic ones [15]. Among the tested biomaterials/materials without pretreatment, cork was the one that showed the best performance in terms of atrazine retention; 0.456 ± 0.034 mg/g. Adsorption of atrazine on cork has a strong concurrency with adsorption of water on this biomaterial. On cork, water is firstly adsorbed on hydrophilic sites constituted by hydroxyl and methoxyl groups and then water adsorption continues by clusters formation around the hydrophilic sites. Water adsorption isotherms in cork present the same profile as atrazine [65,66]. Statistical analysis shows no significant differences between biowaste/waste performance. When comparing the adsorption capacity of biowaste/waste (without pretreatment) with activated rubber char, a significant difference was found (p < 0.0001). The adsorption capacity of these waste materials can be improved through surface modification treatments, which results in a higher affinity for the target pollutant. The methods used to modify the materials can include physical, chemical and biological methods.
is low, the model reduces to Freundlich isotherm, while for high concentrations, it predicts a monolayer adsorption process on a homogeneous surface (Langmuir model) [63]. The mS parameter gives information about the degree of surface non-homogeneity. Heterogeneous adsorbents present mS values close to 0, while mS values closer to 1 indicate a material with relatively homogeneous binding sites [64]. Pine bark presents an mS value closer to 1, suggesting that its binding sites have the same adsorption affinity, homogeneously distributed over its surface. In terms of maximum adsorption capacities, qmax, activated rubber char, is the one that presents the highest value, followed by cork, sepiolite and pine bark, sequentially. Rubber char is the adsorbent with the highest atrazine adsorption capacity since it has enhanced physical properties, as described in the previous section (higher surface area and micropore volume). It is reported that adsorbents/biosorbents without any kind of physical or chemical treatment normally present lower adsorption capacities when compared with the commercial/synthetic ones [15]. Among the tested biomaterials/materials without pretreatment, cork was the one that showed the best performance in terms of atrazine retention; 0.456 ± 0.034 mg/g. Adsorption of atrazine on cork has a strong concurrency with adsorption of water on this biomaterial. On cork, water is firstly adsorbed on hydrophilic sites constituted by hydroxyl and methoxyl groups and then water adsorption continues by clusters formation around the hydrophilic sites. Water adsorption isotherms in cork present the same profile as atrazine [65,66]. Statistical analysis shows no significant differences between biowaste/waste performance. When comparing the adsorption capacity of biowaste/waste (without pretreatment) with activated rubber char, a significant difference was found (p < 0.0001). The adsorption capacity of these waste materials can be improved through surface modification treatments, which results in a higher affinity for the target pollutant. The methods used to modify the materials can include physical, chemical and biological methods.

Desorption Evaluation
The selection of a suitable desorption eluent, with a high affinity for the sorbate and able to remove it from the surface of the sorbent, without substantially changing its morphology and efficiency, while being competitive and environmentally friendly, is seen as challenge in the field of biosorption [31].
The desorption of atrazine from sepiolite, cork and pine bark was evaluated, aiming to assess their regeneration/recycling potential. After one adsorption step, each material loaded with atrazine was placed in contact with different leaching agents. The increase in atrazine concentration in the solution was measured over time for 24 h. Figure 9 shows the atrazine desorption efficiency for waste biomaterials/materials using different eluents and different operation conditions (temperature and percentage of acetonitrile in aqueous solution). After 24 h of operation, around 100% of atrazine was recovered from sepiolite using an aqueous solution with 20% of acetonitrile, while for cork and pine bark, recoveries of 94% and 98%, respectively, were attained using an aqueous solution with 40% of acetonitrile. These results evidence the reversible character of atrazine adsorption onto the materials. An increase in acetonitrile content was needed in the case of cork and pine bark to reach similar desorption values to those attained for sepiolite.
Recycling 2021, 6, x 18 of 28 Figure 9. Atrazine desorption efficiency for sepiolite, pine bark and cork, using different solvents and operation conditions, after 24 h of experiment. Different letters show significant differences (p < 0.05) between extraction solvents for the same waste biomaterial/material. Sepiolite (a-g), pine bark (h-j) and cork (m-r).

PRB Breakthrough Modelling
Assays with continuous flow in fixed-bed columns were carried out as a simulation of an up-scale PRB to evaluate the dynamic removal of atrazine. The overall performance of fixed-bed columns and some operational parameters such as breakthrough time, saturation time, adsorption capacity and breakthrough curve shape were evaluated. These parameters can be determined by plotting the experimental results in normalized concentration, C/C0 over time, t. Figure 10 shows the breakthrough curve obtained with the column set-up experiments with a fixed mass of 20 g of cork, pine bark or sepiolite, when exposed to a continuous flow of a solution with 2 mg/L of atrazine.
Mathematical modelling allows the up-scale design and optimized operation, avoiding unnecessary investments and high operational costs by eventual underutilization or oversaturation of beds. Yoon-Nelson and a modified dose-response model were applied to describe the fixed-bed column behavior and the fitting results are shown in Table 5. The modified dose-response was the model that best fitted experimental data (R 2 > 0.96 for all the materials), represented in Figure 10, suggesting its suitability to be used for the scale-up purpose. The bed adsorption capacity, q0, is a critical indicator of its performance and may be calculated from the breakthrough curve. According to the modified doseresponse model, the highest adsorption capacity was attained for sepiolite (23.3 ±0.8 mg/g), followed by pine bark (14.8 ± 0.6 mg/g) and cork (13.0 ± 0.9 mg/g). Using the Yoon-Nelson model, it was possible to predict the time required to retain 50% of the initial atrazine concentration (τ). Sepiolite achieved the highest value of τ, followed by pine bark and cork, which is in agreement with the adsorption capacity performance predicted by the modified dose-response model. Sepiolite showed the best adsorption performance and, since it is the densest material, a smaller bed length is occupied, which becomes an advantage in terms of experimental operation. Continuous treatments become more feasible than batch processes, allowing an intra-particle diffusion of a bigger amount of atrazine An organic solvent such as acetonitrile increases atrazine solubility in water, justifying its selection for this desorption study. The use of methanol as the most adequate extraction solvent for atrazine desorption from synthetic polymers was demonstrated in a recent study, achieving 73.5% as maximum atrazine desorption percentage [67]. In the same study, it was concluded that higher efficiency values are reached for all materials when using a lower content of organic solvent, which minimizes the environmental impact of this process. Thus, atrazine recovery from these biomaterials/materials can be easily and successfully obtained through extraction with aqueous acetonitrile solutions, which is a rapidly and highly biodegradable solvent.
For sepiolite, a considerably higher desorption efficiency was reached using water at 60 • C, in comparison to the recoveries attained at 25 • C and 40 • C. These results confirm the reversible nature of the exothermic adsorption process. however, this tendency was not as evident for cork and pine bark as for sepiolite.
The desorption efficiency increased for all materials with the increase in the pH of the buffer from 5 to 7. With pH 5 buffer as acidic solvent, atrazine desorption reached 30%, 20% and 5% of the original load from sepiolite, cork and pine bark, respectively. Statistical analysis of desorption data is expressed in Figure 9. In the case of sepiolite, all the extraction solvents used for desorption differ from each other (p < 0.05). For pine bark and cork, the aqueous solution with 40% of acetonitrile revealed to be the one with the highest difference compared to the other studied solvents (p < 0.001).

PRB Breakthrough Modelling
Assays with continuous flow in fixed-bed columns were carried out as a simulation of an up-scale PRB to evaluate the dynamic removal of atrazine. The overall performance of fixed-bed columns and some operational parameters such as breakthrough time, saturation time, adsorption capacity and breakthrough curve shape were evaluated. These parameters can be determined by plotting the experimental results in normalized concentration, C/C 0 over time, t. Figure 10 shows the breakthrough curve obtained with the column set-up experiments with a fixed mass of 20 g of cork, pine bark or sepiolite, when exposed to a continuous flow of a solution with 2 mg/L of atrazine.
Recycling 2021, 6, x 19 of 28 molecules into the pores of the materials. Although batch processes are often used in adsorption studies, the collected data from those experiments are not necessarily applicable to a continuous adsorptive system. Continuous adsorption studies with PRB columns, at lab and pilot scale, are still needed to give real information for a trustworthy scale up.

Atrazine Retention and Regeneration/Recycling
After the selection of the best desorbing eluent, it is essential to evaluate the number of cycles of adsorption/regeneration that selected materials can cope with, as they may reduce the overall operational costs. These processes can be performed either in batch or continuous procedures; however, when the pollutant retention is conducted in PRB columns, it is better to perform the regeneration step in continuous mode [31]. When the PRB column becomes saturated, the adsorption process is stopped and switched to the desorption/regeneration step. In order to assess the recycling potential of sepiolite, pine bark and cork, successive cycles of adsorption and desorption were performed. Experiments were conducted in columns with the materials placed as PRB. Based on the results of Section 2.4, aqueous solutions with 20% acetonitrile (for sepiolite) and 40% acetonitrile (for pine bark and cork) were used for sorbent regeneration. The assays were initiated by 1 step of adsorption from 2 mg/L of atrazine solution in close-loop (until equilibrium was reached for each material), and then followed by one desorption step in continuous mode over 48 h. The equilibrium of the adsorption process using sepiolite as PRB was reached in 96 h, with pine in 82 h and with cork in 48 h. Three adsorption and two desorption steps were performed, and the values of atrazine removal (%) and respective standard deviations for each step and each material are presented in Table 6. The percentage of atrazine removal Mathematical modelling allows the up-scale design and optimized operation, avoiding unnecessary investments and high operational costs by eventual underutilization or oversaturation of beds. Yoon-Nelson and a modified dose-response model were applied to describe the fixed-bed column behavior and the fitting results are shown in Table 5. The modified dose-response was the model that best fitted experimental data (R 2 > 0.96 for all the materials), represented in Figure 10, suggesting its suitability to be used for the scale-up purpose. The bed adsorption capacity, q 0 , is a critical indicator of its performance and may be calculated from the breakthrough curve. According to the modified dose-response model, the highest adsorption capacity was attained for sepiolite (23.3 ± 0.8 mg/g), followed by pine bark (14.8 ± 0.6 mg/g) and cork (13.0 ± 0.9 mg/g). Using the Yoon-Nelson model, it was possible to predict the time required to retain 50% of the initial atrazine concentration (τ). Sepiolite achieved the highest value of τ, followed by pine bark and cork, which is in agreement with the adsorption capacity performance predicted by the modified dose-response model. Sepiolite showed the best adsorption performance and, since it is the densest material, a smaller bed length is occupied, which becomes an advantage in terms of experimental operation. Continuous treatments become more feasible than batch processes, allowing an intra-particle diffusion of a bigger amount of atrazine molecules into the pores of the materials. Although batch processes are often used in adsorption studies, the collected data from those experiments are not necessarily applicable to a continuous adsorptive system. Continuous adsorption studies with PRB columns, at lab and pilot scale, are still needed to give real information for a trustworthy scale up.

Atrazine Retention and Regeneration/Recycling
After the selection of the best desorbing eluent, it is essential to evaluate the number of cycles of adsorption/regeneration that selected materials can cope with, as they may reduce the overall operational costs. These processes can be performed either in batch or continuous procedures; however, when the pollutant retention is conducted in PRB columns, it is better to perform the regeneration step in continuous mode [31]. When the PRB column becomes saturated, the adsorption process is stopped and switched to the desorption/regeneration step. In order to assess the recycling potential of sepiolite, pine bark and cork, successive cycles of adsorption and desorption were performed. Experiments were conducted in columns with the materials placed as PRB. Based on the results of Section 2.4, aqueous solutions with 20% acetonitrile (for sepiolite) and 40% acetonitrile (for pine bark and cork) were used for sorbent regeneration. The assays were initiated by 1 step of adsorption from 2 mg/L of atrazine solution in close-loop (until equilibrium was reached for each material), and then followed by one desorption step in continuous mode over 48 h. The equilibrium of the adsorption process using sepiolite as PRB was reached in 96 h, with pine in 82 h and with cork in 48 h. Three adsorption and two desorption steps were performed, and the values of atrazine removal (%) and respective standard deviations for each step and each material are presented in Table 6. The percentage of atrazine removal after the adsorption processes is calculated based on the difference between the influent atrazine concentration and its outflow concentration, at equilibrium stage. Additionally, atrazine desorption is calculated based on the concentration of atrazine that was retained in the PRB column on each process and its concentration in the desorbing eluent at the end of 48 h of desorption. Figures 11-13 show the output concentration (C out ) profile of atrazine over time during adsorption using sepiolite, pine bark and cork in PRB, respectively. Figures 14-16 exhibit the concentration profile of atrazine during desorption on the outflow of the PRB of sepiolite, pine bark and cork, respectively. The results presented on Table 6 show that desorption was performed effectively after the two cycles for sepiolite and pine bark, with performance percentages varying from 85% to 96%. Cork achieved 73% and 55% of desorption capacity after the first and the second process, respectively. The results clearly show that the use of the selected desorption solvents, as well as promoting an effective extraction of the retained pollutant, also promote, in a general way, an improvement of the adsorption capacity during the second adsorption step. After two regeneration cycles, removal percentages increased 30% for sepiolite, 15% for pine bark and remained practically unchanged for cork. This can be explained by the incomplete desorption of atrazine from cork after 48 h, with some adsorption sites remaining unavailable for the second adsorption step. Sepiolite presented the best performance in terms of desorption efficiency, as well as good adsorption stability after regeneration, with the advantage of a smaller environmental impact of the selected elution solvent.

Sustainability Assessment and Lifecycle Perspective
It is common knowledge that AC is the most adopted adsorbent in the field of wastewater treatment, since it has the capability to efficiently retain a wide range of pollutants. However, its widespread use often becomes disadvantageous due to its high production cost (unaffordable to underdeveloped countries), difficult regeneration and endof-life disposal options [70]. Additionally, the use of AC induces high environmental im-

Sustainability Assessment and Lifecycle Perspective
It is common knowledge that AC is the most adopted adsorbent in the field of wastewater treatment, since it has the capability to efficiently retain a wide range of pollutants. However, its widespread use often becomes disadvantageous due to its high production cost (unaffordable to underdeveloped countries), difficult regeneration and endof-life disposal options [70]. Additionally, the use of AC induces high environmental im- Previous studies have demonstrated the performance of different solvents on the regeneration of adsorbent materials. Mashile et al. [68] demonstrated the efficient use of methanol (up to 95%) to desorb methylparaben and propylparaben and to regenerate waste tire activated carbon-chitosan composite without affecting adsorption capacity. garcía et al. [69] selected pure acetonitrile as an organic desorption agent to regenerate spent biochar, almost achieving a complete desorption of the pollutants, 96.8% and 98.0%, for methyl paraben and sulfamethoxazole, respectively. In the present study, high desorption capacities are also reached, but lower environmental impact solvents are used while maintaining the adsorption capacity of the materials. On this basis, it is possible to conclude that sepiolite, cork and pine bark can be regenerated and reused at least three times, without any loss in their adsorption efficiency, while acquiring an added value and favoring circular economy.

Sustainability Assessment and Lifecycle Perspective
It is common knowledge that AC is the most adopted adsorbent in the field of wastewater treatment, since it has the capability to efficiently retain a wide range of pollutants. however, its widespread use often becomes disadvantageous due to its high production cost (unaffordable to underdeveloped countries), difficult regeneration and end-of-life disposal options [70]. Additionally, the use of AC induces high environmental impacts such as ecotoxicity, acidification, ozone depletion and fossil fuel depletion [71]. Biosorbents have been recognized as an environmentally sustainable solution for wastewater treatment. In addition to their capacity to retain different pollutants, and the fact that they are available on a sustainable basis, their low cost, biodegradability and reduced toxicity character are important advantages to consider. The use of waste-based biomaterials/materials contributes to environmental sustainable growth, since it reduces waste, as well as promotes the recovery and reuse of materials [15]. The reuse of residues allows their life cycle to increase and energy requirements to decrease, resources and emissions [72]. Moreover, it allows the waste disposal problem and the costs for producing new adsorbents for water treatment to be reduced [31]. Through waste valorization, a circular and sustainable economy is promoted, a prevailing priority in the EU in order to achieve a continuous decrease in the negative environmental effects of services and products throughout their life cycle [73]. Bioeconomy is meant to reduce the dependency on natural resources by converting biomass and organic waste into diversified end-products and materials, and providing competitive goods in an environmentally friendly way [74]. This study is focused on the recycling of biological and natural resources waste, which is in line with the expected environmental policies. The sorbent materials, sepiolite, cork and pine bark were tested without any previous chemical treatment, with reduced waste generation (only the one produced by milling and washing steps) so they are good choices to be used as sorbents. As well as being non-toxic materials, they are selective towards the target pollutant, are easy to operate in PRB due to their particle size and are regenerable and reusable. Exhausted sorbents can be also regenerated by successive adsorption/desorption cycles using acetonitrile aqueous solutions, which allows the further transference of atrazine into new solutions, for possible reuse. An economic and environmental evaluation should be taken in consideration to select the material to perform an adsorption process.

Conclusions
This study shows the adsorption potential of waste-based biomaterials/materials without resorting to chemicals for the removal of atrazine from water. Process kinetics were satisfactorily described by the PSO model for sepiolite, cork and pine bark. The Freundlich model fits the adsorption isotherms for cork and sepiolite, while the Sips model fits adsorption on pine bark more adequately. Kinetics and equilibrium of atrazine adsorption onto activated rubber char from waste rubber were also evaluated as a comparative term for the other materials without treatment.
Atrazine desorption feasibility was evaluated in batch systems with different extraction solvents. In 24 h, 100% of atrazine was recovered from sepiolite using an aqueous solution of 20% acetonitrile, while from cork and pine bark the highest values of 98% and 94% were reached, respectively, using solutions of 40% acetonitrile. These desorbing agents were selected to regenerate sorbents in column recyclability assays.
Continuous adsorption of 2 mg/L of atrazine using sepiolite, cork and pine bark as PRB was performed in columns at lab scale. Breakthrough curves were best fitted by the modified dose-response model for all the tested materials (R 2 > 0.96), indicating its suitability for process design and upscaling. The maximum adsorption capacity for sepiolite was 23.3 (±0.8) mg/g, 14.8 (±0.6) mg/g for pine bark and 13.0 (±0.9) mg/g for cork, and these values were determined by mathematical modelling. Sepiolite was revealed to be the best material in terms of adsorption capacity, with the longest breakthrough time of 11 h, followed by 5 h for cork and 4 h for pine bark.
The recycling potential of sorbents in PRB was assessed by consecutive cycles of atrazine adsorption-desorption in column assays. After two regeneration cycles, the best performance in terms of atrazine recovery was measured for sepiolite with 81%, followed by pine bark and cork with values of 78% and 54%, respectively. The desorption efficiency obtained in the second step of 48 h, reached 96% for sepiolite and pine bark, and 55% for cork. Results show that sepiolite is the best material in terms of atrazine adsorption capacity and stability after regeneration, with high potential for reuse and depending on a leaching agent with smaller environmental impact. The characteristics of the selected sorbents, before and after two cycles of sorption/regeneration, were evaluated by SEM and FTIR analyses and no substantial differences were found in structure or morphology.
Adsorption technology employing waste-based materials/biomaterials can be a sustainable option for wastewater treatment, once it promotes waste valorization, pollution mitigation and environmental protection by reducing natural resources dependency. Through waste valorization, a circular and sustainable bioeconomy is promoted by the increased waste life cycle, a prevailing priority in the EU.
The application of sorbents able to compete with commercial ones is still seen as an emerging research challenge. The present study is in line with the growing concern on herbicide removal, and uses reusable waste with recovery capacity and recycling ability, expected by environmental policies.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.