Experimental study and theoretical modelling based on DFT of the elimination of ceftriaxone by an activated carbon synthesized from grape seeds waste Authors: Sirine Mabrouk, Hatem Dhaouadi Published: 2025 ABSTRACT AbstractAn activated carbon (AC) developed from grape seeds waste was characterized using various physico-chemical techniques to assess its structural and textural properties, then used for the removal of ceftriaxone (CFTX), an emerging pollutant, from aqueous solutions. Batch adsorption experiments were conducted, coupled with a theoretical study based on DFT (density functional theory) study , performed at the B3LYP level with a 6-311G (d,p) base , in order to study molecular geometries, electronic properties and molecular electrostatic potentials, to clarify the adsorption mechanism of CFTX on AC. The results revealed that the prepared AC had a well-established porosity offering a large exchange surface. Adsorption studies revealed that kinetic data aligned well with the PSO model, while equilibrium data were best described by the Freundlich isotherm and the multilayer STAT1 model. Thermodynamic analysis indicated endothermal and favorable adsorption, mainly controlled by chemical interactions. Moreover, the regeneration studies showed the difficult desorption of CFTX, supporting a chemisorption mechanism. The results of DFT calculations indicated that the adsorption of CFTX is favorable on AC functionalized with –CHO and –COOH–OH–CHO, with adsorption energies of − and − , respectively. AIM (atoms in molecules), NCI (non-covalent interactions) and RDG (reduced density gradient) analyses revealed that strong hydrogen-bond interactions mainly govern the adsorption mechanism. According to the obtained results, the produced AC is considered as a high-performance alternative material for the removal of CFTX from aqueous environments. In addition, a thorough understanding of the adsorption process resulted from integrating theoretical and experimental approaches.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-025-27184-6.Keywords: Activated carbon, Adsorption, Ceftriaxone, DFTSubject terms: Chemistry, Engineering Introduction In recent years, population growth and accelerated industrialization have led the emergence of new types of pollutants. These emerging contaminants are an environmental and health concern1. Among them, drug residues, especially antibiotics, are of increasing disquiet related to their persistence in the naturel environments, their ability to bioaccumulate in living organisms and their resistance to biodegradation. Antibiotics used in human2 and veterinary3 medicine have recently been founded in different aquatic environments, such as wastewater, surface water, groundwater and sometimes even in drinking water. This is mainly due to human and animal excretions, discharges from the pharmaceutical industries and hospitals, as well as runoff from agricultural activities4. Even at low concentrations, their presence represents a significant risk, particularly for global public health through the spread of infections that are difficult to treat5,6. Ceftriaxone (CFTX), a third-generation cephalosporin antibiotic, is one of the most widely used antimicrobial agents worldwide, thanks to its broad efficacy against several types of bacterial infections, including those affecting the respiratory and urinary tract7. Despite its clinical efficacy combined with its relatively low cost, its extensive use leads to worrying environmental impacts8. Recently, studies have shown the detection of CFTX in wastewater treatment plant effluents, at significant concentrations reaching up to 15 μg L-1 in urban and hospital wastewater, indicating its widespread diffusion and persistence9,10. CFTX is described as an emerging pollutant of concern1, due to its ability to promote bacterial resistance, which therefore requires effective treatment methods to minimize its impact on ecosystems and public health by reducing their quantity in the aquatic environment11. Due to the ineffectiveness of conventional treatments to completely remove antibiotics or other drugs from wastewater8, various sophisticated purification technologies have been developed, namely ultrafiltration12, advanced oxidation processes (AOP)13, adsorption14, and biodegradation11. However, these techniques have certain environmental and economic drawbacks, such as the generation of secondary pollutants, high cost and complexity of use15.Adsorption stands out among these techniques as a promotional method due to its simplicity of implementation, low operation cost, and its ability to effectively remove a wide variety of contaminants, including drugs, dyes, or even complex organic compounds16, from an aquatic environment17. Activated carbon (AC) is a highly efficient adsorbent widely employed in the purification and treatment of water and air by adsorption of various pollutants18. Its adsorption properties, such as well-developed porosity, a high specific surface area and strong adsorption capacity, enable it to can be recognized as high-performance adsorbent material, suitable for the elimination of different types of contaminants17. In addition, several studies have shown that AC can be modified or functionalized to improve its effectiveness in removing various organic and inorganic contaminants19. However, the elevated cost of commercially available ACs, often derived from renewable raw materials, emphasizes the necessity of exploring alternative sustainable and low-cost alternatives for the production of ACs. Recently, many attempts have been made to study the use of agricultural waste and renewable biomass as precursors for AC synthesis20. Among these unconventional sources, grape seeds, an abundant by-product of the wine industry, have attracted considerable attention because of their abundance, accessibility, and high carbon composition21. The production of AC from agro-industrial wastes, including lotus stalks22, wood and fishing kernels23, coconut shells24, pomegranate wood25, and clusters of empty palm fruits15, is a potentially attractive production route from an economic and environmental point of view. These wastes also constitute promoters for the production of AC due to their abundance and effectiveness in removing various pharmaceutical contaminants from water26. Several studies have demonstrated the effectiveness of AC derived from waste as a high-performance adsorbent for the elimination of pharmaceutical products present in water. In a study by Grisales-Cifuentes, Claudia et al.27 studied the adsorption of acetaminophen using activated carbon from palm oil, with an elimination efficiency reaching 7.3 mg g-1. In another study, Bagheri, Afrouz, et al.28 explored the elimination of ketoprofen by activated carbon derived from nutshell, obtaining an elimination capacity of 444 mg g-1. Chen, Yue, et al.29 developed activated carbon from corn cobs for the adsorption of levofoxacin from water, with a removal efficiency of 56.6 mg g-1. In another study by Chen, Jiaqi, et al.30 adsorbents derived from lotus leaves were used to remove norfloxacin, the result shows an elimination efficiency reaching 91.7 mg g-1. The chemical composition, structural properties of the AC material31, physicochemical characteristics of the adsorbed molecules, and the nature of molecular interactions formed with the surface of the adsorbent32, have a major impact on the adsorption mechanism and capacity. In this context, the evaluation of adsorption mechanisms uses advanced quantum chemistry methods. These approaches are particularly effective for analyzing interactions at the molecular level33. In recent years, simulations relying on density functional theory (DFT) have been studied to investigate intermolecular interactions and their effects on molecular properties. These simulations provide precise information on molecular configurations, adsorption sites and also the energies of the bonds formed34,35. Understanding the adsorption of CFTX on AC from examining the impact of various function groups exists on the AC surface is the objective of this study. Theoretical calculations based on (DFT) applying the B3LYP approach coupled to the 6–311G (d, p) basis set were used to study the interactions formed between CFTX in these functional groups by estimating the adsorption energies, while determining the most stable molecular configurations. This study mainly introduces two key novelties. First, it presents the development of AC derived from grape seeds, a new and sustainable biomaterial for the elimination of CFTX in water, thus contributing to green chemistry and waste valorization. Secondly, this research combines the experimental analysis of adsorption with density functional theory (DFT) based modeling, offering a theoretical framework for a thorough understanding of the adsorption mechanisms of CFTX on functionalized activated carbon surfaces. This dual approach is relatively unexplored in the context of CFTX adsorption, providing new insights into interactions at the molecular level. Considering the above, this study focuses on the evaluation of CFTX removal, by the use of AC produced from grape seed waste, present in aqueous solution. The structural and textural proprieties of the developed adsorbents were determined using various analytical methods. An equilibrium, kinetic and thermodynamic study was performed in order to examine the adsorption process. A desorption study was executed to evaluate the regeneration of these materials. In addition to the experimental tests, quantum chemistry calculations based on DFT were conducted. This combination of experimental and theoretical approaches not only validates the results obtained but also allows for a better understanding of the retention mechanisms. Materials and methods Materials Grape seeds waste from the AGRILAND® agro-food industry, based in Mitbasta, Kairouan, Tunisia, were employed as the raw material for AC production. Phosphoric acid H3PO4 (85% (W)) was used as the chemical activation agent. Ceftriaxone (CFTX), an injectable solution with a purity of more than 97% (UNIMED®), was selected for the adsorption experiments. The characteristics and chemical structure of CFTX are shown in Table 136. Table 1CFTX characteristics and chemical structure.StructureChemical formulaC18H18N8O7S3Molecular weight (g mol-1)554.6Water solubility (g L-1)0.105pka3, 3.2 and 4.1 CFTX characteristics and chemical structure. Activated carbon preparation Approximately 200 g of grape seeds waste were soaked in a diluted H3PO4 solution for 24 h at room temperature. The impregnation was carried out according to different impregnation rates defined as the mass ratio between the activating agent and the precursor, ranging from 1 to 2. The pyrolysis of the impregnated and dried material was then performed in a controlled tubular furnace, under a continuous nitrogen flow (3 L min-1), at temperatures between 400 and 800 °C, for a duration varying from 60 to 180 min. Once cooled, the obtained AC underwent a washing with hot distilled water until neutrality to remove activation agent residues and unwanted impurities. Finally, the material was dried at 100 °C for 4 h. In this study, the preparation of AC was performed under optimized conditions, namely the impregnation rate (X1), temperature (X2, in °C), and pyrolysis time (X3, in minutes), obtained from the analysis of the experimental results using the Box – Behnken Design (BBD), based on the Response Surface Methodology (RSM). Figure S1 visually illustrates the final optimization diagram resulting from this analysis, indicating the optimal operating conditions for the AC preparation: temperature = 800 °C, time = 180 min and impregnation rate = 2. The iodine index II2 (Y1, expressed in mg g-1) and methylene blue index MBI (Y2, expressed in mg g-1) taken as process responses were maximized. The experimental design was performed using Minitab 20.3 statistical software. The Table 2 presents the experimental design with the corresponding results. Table 2BBD experimental design matrix with response results.TEST NoSettingsAnswersX1X2X3Y1Y2Impregnation rateTemperature (°C)Time (min)Iodine Index II2 (mg g-1)Methylene blue index MBI (mg g-1)11.5600120467.010.6021.5400180614.220.2332400120583.760.2841.540060578.680.2451400120609.141.206160060558.380.517260060573.610.8181.5600120548.230.5792600180609.141.07101.5600120548.230.85111.580060659.901.11122800120634.521.19131.5800180736.041.21141600180507.620.47151800120710.661.05 BBD experimental design matrix with response results. Physical and chemical characterization of AC In a previous study37, the textural properties of the prepared AC were characterized by analysis of N2 adsorption/desorption at 77 K, measured with a surface analyzer (Micromeritics ASAP 2020). The specific surface area was evaluated by the BET method (Brunauer–Emmett–Teller)38. The BJH approach (Barrett-Joyner-Halenda) was used to analyze the pore size distribution39. Scanning electron microscopy (SEM) coupled with energy-dispersive X-ray spectroscopy (SEM–EDX) (Thermo Scientific Apreo S) was employed to analyze the surface morphology and elemental composition of AC. In addition, high-resolution X-ray diffraction (XRD) was used to study the crystal structure of AC. The analysis of acid and basic functional groups on the AC surface was evaluated by Boehm titration40. 1.5 g of AC was mixed with 0.05 M solutions of NaOH, Na2CO3, NaHCO3, and HCl. After stirring for 24 h, the remaining excess acid or base was titrated with NaOH or HCl (0.05 M), respectively. The quantification of functional groups is based on neutralization reactions. The pHpzc was carried out following the method described by Lopez-Ramon et al.41. A 0.01 M NaCl solution was mixed with 0.15 g of AC whose initial pH was adjusted to values between 1 and 12. After stirring for 48 h, the change in pH values (ΔpH), was plotted as a function of the initial pH. The pHpzc corresponds to the point where pHinitial is equal to pHfinal. Adsorption study The study of CFTX adsorption on AC was carried out by a batch adsorption method. For the experiments, an optimal mass of 0.2 g AC (Fig. S2) was mixed with CFTX solutions of 20 mL volume and at concentrations varying from 4 and , at room temperature and constant pH (Fig. S3). All the experiments were performed under constant stirring at 150 rpm, with samples taken at intervals of 3 to 360 min until equilibrium was achieved. The thermodynamic study was conducted between 30 °C and 40 °C, by determining the main parameters, notably the standard Gibbs free energy (ΔG°), standard enthalpy (ΔH°), and entropy (ΔS°). After each experimental run, the residual CFTX concentrations were determined by UV–vis spectrophotometer (CamSpec M501) at 240 nm. The following equations, including removal percentage (Eq. 1)42, as well as the retention capacity measured at equilibrium (Eq. 2)42, and at any given time in the process (Eq. 3)42, were used to determine the amount of CFTX retained by AC: Thermodynamic (Langmuir, Freundlich, BETn-finite, and BETn-infinite) and statistic (STAT1, STAT2, and STAT3) models were used to adjust the equilibrium isotherm data. Table 3 summarizes the fundamental equations and parameters associated with these models employed to analyze CFTX adsorption on AC. Adsorption kinetic were evaluated using the pseudo-first-order (PFO)43, pseudo-second-order (PSO)44, and intraparticle diffusion45 models. Table 3Equations and parameter significance for thermodynamic and statistic isotherm models.ExpressionParameter significanceReferencesThermodynamic models Langmuirqmax (mg g-1): maximum adsorption capacity of CFTXKL (L mg-1): Langmuir constant associated with binding energies48 FreundlichKF (mg g-1) (L mg-1)1/nF): Freundlich constant related to the adsorption capacity nF: degree of surface irregularity49 BETn-finiteqe = qmax,1 qmax,1 (mg g-1): maximum adsorption capacity of CFTX on the first layern: maximum number of adsorbed layers (L mg-1): adsorption equilibrium constant for the first layer (L mg-1): adsorption equilibrium constant for the upper layers, whereEL and ES (J mol-1): adsorption heat on the first and upper layersa1 and b1: rate constants for adsorption and desorption on the first layera2 and b2: rate constants for adsorption and desorption on the upper layers50 BETn-infiniteStatistic models STAT1qmax,1 (mg g-1): maximum adsorption capacity of CFTX on the first layerm, steric parameter where:mCFTX + S < = = > CFTXmSwith S representing the adsorption site: half-saturation concentration of the first layer: half-saturation concentration for the entire isotherm and adsorption energies for the first and upper layers, respectively50 STAT2 STAT3 Equations and parameter significance for thermodynamic and statistic isotherm models. qmax (mg g-1): maximum adsorption capacity of CFTX KL (L mg-1): Langmuir constant associated with binding energies qmax,1 (mg g-1): maximum adsorption capacity of CFTX on the first layer n: maximum number of adsorbed layers (L mg-1): adsorption equilibrium constant for the first layer (L mg-1): adsorption equilibrium constant for the upper layers, where EL and ES (J mol-1): adsorption heat on the first and upper layers a1 and b1: rate constants for adsorption and desorption on the first layer a2 and b2: rate constants for adsorption and desorption on the upper layers qmax,1 (mg g-1): maximum adsorption capacity of CFTX on the first layer m, steric parameter where: mCFTX + S < = = > CFTXmS with S representing the adsorption site : half-saturation concentration of the first layer : half-saturation concentration for the entire isotherm and adsorption energies for the first and upper layers, respectively The selection of the best fit for the kinetic and isotherm models was generally performed using the linear correlation coefficient R2. However, because of potential bias in the results, the Chi-square (χ2) and sample-adjusted Akaike information criterion (AICc) were also used in addition to the correlation coefficients. The AICc is expressed as follows (Eq. 4)46: n represents the number of data points, k is the number of parameters in the model and SSR refers to the residuals sum of squares, which is also used to calculate R2. The formulation of the Chi-square (χ2) is expressed as follows (Eq. 5)47: n corresponds to the number of data points used for calculation. Desorption study Desorption tests were carried out to evaluate the efficiency of adsorbent recovery. For this purpose, 0.2 g of the adsorbent charged with CFTX, obtained from the isothermal study at different concentrations, was agitated in 20 mL of distilled water. The desorption efficiency (%) was determined using the following equation (Eq. 6)51: qdes and qads represent the quantities of desorbed and adsorbed CFTX (mg g-1), respectively. DFT modelling Molecular modeling was performed to study the interactions and chemical reactivity of CFTX during its adsorption onto the AC surface. Optimization of molecular geometries was performed using Density Functional Theory (DFT) with the exchange–correlation function B3LYP, associated with the bases 6-311G (d, p). The B3LYP functional, is commonly used as it offers a favorable compromise between accuracy and computational efficiency, which then allows it to be adapted to a variety of molecular systems52. This hybrid functional is a combination between the three-parameter Becke exchange functional and the gradient-corrected Lee–Yang–Parr correlation functional. The occuracy of calculations, for non-covalent interactions was improved by applying the Grimme D3BJ dispersion correction53. Due to the limitation of standard DFT calculations in describing the various interactions formed between adsorbate molecules and the adsorbent surface, this correction is important to integrate long-range Van der Waals forces. The influence of solvents was simulated using the CPCM (Conductor-like Polarizable Continuum Model)54. The optimized geometries were validated as local minima by the absence of imaginary frequencies in the corresponding Hessian matrices. To deepen the understanding of the adsorption mechanisms, key quantum chemistry descriptors, such as the highest energy occupied molecular orbital (HOMO, EH) and the lowest unoccupied molecular orbital (LUMO, EL), as well as the energy gap (Egap = EL-EH), were analyzed on the basis of DFT calculations. Using the HOMO and LUMO values, some reactivity descriptors for all molecules (adsorbate and adsorbent) were determined, in particular the electron chemical potential (μ), global hardness (η), and electrophilic index (ω) using equations (Eq. 7)55, (Eq. 8)55 and (Eq. 9)55 respectively: The adsorption energies (Eads), which reflect adsorption that is more favorable as their values are high, were calculated for all adsorbate/adsorbent complexes using the following equation (Eq. 10)56,57: where ECFTX is the total energy of CFTX; EAC represents the total energy of AC and ECFTX+AC corresponds to the total energy of the CFTX/AC complex. The theoretical simulations were conducted using the Gaussian 09 software. The modelling and visualization of the structures were performed using VMD58 and Gauss View59 software, respectively. The interaction types between CFTX and AC are best described using the Reduced Density Gradient and Non-Covalent Interactions (NCI and RDG)60 and the theory of Atoms In Molecules (AIM)61. The Multiwfn software62, a multifunctional wave function analysis program, was used to perform AIM NCI, andRDG topological analyses. Experimental results and interpretation Structural and physicochemical characterization of AC Morphological and textural characteristics The prepared AC had a large exchange surface of 543 m2 g-1, supported by the presence of micropores representing 87% of the total pore volume (0.240 cm3 g-1)37 (Fig. S4). A well-established porosity offering a large exchange surface have a fundamental importance in assessing the effectiveness of adsorbents in water treatment, because they promote the existence of active sites on the AC surface and allow a better penetration and diffusion of contaminants through pores58. The morphological characteristics of the AC surface were studied using a SEM analysis. As illustrated in Fig. 1, the micrographs reveal a rough surface dotted with cavities presenting a varied morphology with irregular shape and size. The formation of these cavities results from the vaporization of the volatile compounds generated by the activation agent H3PO4 during carbonization, which leads to the expansion of existing pores and/or the formation of new pores63. These cavities constitute channels facilitating the access of adsorbed molecules to microporous and mesoporous regions of carbon. Similar results have been reported in previous researches64, which also highlighted surfaces with comparable characteristics. SEM micrographs of the prepared AC, at magnifications of 1000x (a) and 5000x (b). The elemental composition of AC, determined by energy diffusion X-ray (EDX) analysis and presented in Table 4, and Fig. 2, shows that carbon and oxygen are the main constituents of AC, with percentages of 69.41% and 19.25%, respectively. These elements indicate the presence of organic functional groups which can act as adsorption sites. Adsorbents with high carbon content are well known for their efficiency in removing various contaminants. In addition, phosphorus detection can be attributed to the use of phosphoric acid (H3PO4) during activation65. Other elements, such as magnesium, aluminum, potassium, and calcium, have also been identified. Table 4EDX elemental analysis of prepared AC.CompoundCOMgAlSiPKCaFeCuWeight (%)69.4119.250.520.280.336.710.442.870.100.08 EDX elemental analysis of prepared AC. EDX spectrum of the prepared AC. The carbon X-ray diffraction (XRD) diagram shown in Fig. 3 shows strong and wide peaks at approximately 25° and 45°, attributed to the Bragg reflections (002) and (100) of amorphous carbon66,67. These drawings reveal the partially graphitic structure of AC. The presence of similar peaks indicates a high degree of disorder, which is typical of carbon materials. Comparable results have been obtained in other studies working on AC derived from agro-industrial waste68. XRD diagram of the AC. Surface chemical characteristics For surface chemistry, pH at the zero charge point (pHPZC), is a key parameter in adsorption mechanisms. The pHPZC was measured at about 2.5, indicating a predominance of acidic character on the AC surface. Below this pH value, the AC surface is acidic, promoting the adsorption of negatively charged particles. In contrast, when the pH exceeds the pHPZC, the surface moves to a basic nature, facilitating interactions with cationic species. In the case of CFTX, its behavior in solution changes according to its three acid dissociation constants (pKa = 3.0, 3.2, and 4.1). Thus it can exist in different stable ionic forms depending on the pH of the solution. These ion-shaped transitions cause changes in the electron configuration, allowing them to establish specific interactions with the AC surface. The quantification results of the sites potentially responsible for drug residue adsorption are presented in Table 5. These data show that the surface has a slight acidity, mainly attributed to the existence of carboxylic, lactonic, and phenolic functional groups. In addition, the total basicity value (2.37 meq g-1) was lower than the total acidity value (2.99 meq g-1), this confirms the acid character of the AC. Table 5Acid and basic functional groups on the prepared AC surface determined by Boehm titration and pHPZC value.Acid functionsTotal basicitypHpzcCarboxylicLactonicPhenolicTotal acidity0.950.551.492.99 (± 0.02)2.37 (± 0.03)2.5 Acid and basic functional groups on the prepared AC surface determined by Boehm titration and pHPZC value. The analysis of the IR spectrum (Fig. 4) of the raw processed activated carbon (before adsorption) showed mainly two bands. A band around 1600 cm-1 corresponds to the elongation vibration of the C=O group (carbonyls) or aromatic type vibrations C=C, and a wide band in the region 1000–1200 cm-1, related to the elongation vibrations of the C–O group (carboxylic and phenolic). FTIR spectra of AC before and after adsorption of CFTX. This observation corroborates the results of the Boehm assay, confirming the presence of polar functional groups likely to interact with CFTX molecules through hydrogen bonds or electrostatic interactions. Adsorption studies Adsorption kinetics Adsorption kinetics experiments were performed to estimate the CFTX elimination rate of the aqueous solution and to better understand the adsorption mechanism of the adsorbents. Figure S5 illustrates the adsorption capacities of AC plotted against the contact time for varying initial concentrations of CFTX. The results showed that adsorption occurred rapidly during the first 30 min, then gradually slowed down to equilibrium after approximately 180 min with maximum adsorbed amounts ranging from 0.37 mg g-1 to 1.36 mg g-1 for initial CFTX concentrations of 4 and 20 mg L-1, respectively. This can be explained by the abundance of unoccupied active sites at the beginning of the process, whose availability decreases over time, suggesting a saturation of the adsorbent surface67. Furthermore, this relatively low adsorption capacity achieved can also be explained by the complexity of the CFTX molecule, comprising several functional groups, thus limiting its interaction with the carbon surface of the elaborated AC. Thus, its size and high spatial configuration may not be fully adapted with the pores of AC, despite its developed porosity and high specific surface area. However, although other materials69,70 have revealed better adsorption performances due to their very porous structures and large specific surfaces, the AC developed remains advantageous in terms of its low-cost synthesis from agro-industrial waste and their relatively simple preparation process. To identify the most suitable kinetic model for describing CFTX retention on AC, the experimental adsorption data were adjusted in a non-linear manner using PFO, PSO, and intra-particle diffusion models. The parameters listed in Table 6 indicate that adsorption is most accurately represented by the PSO model, for all concentrations investigated. Indeed, it is distinguished by the highest values of the correlation coefficient (R2) and the smallest values of chi-square (χ2) and corrected Akaike Information Criterion (AICc). In addition, the qe values calculated using this model are presents a great similarity with experimental data, thus confirming that the kinetics mainly follows a pseudo-second-order model. This suggests a chemisorption process characterized by strong interaction forces68. These results are consistent with another study71, indicating as well that the PSO model provides the best suited for describing antibiotics removal on AC derived from waste. Table 6Parameters of kinetic models for various initial concentrations of CFTX.Ceftriaxone concentration (mg L−1)46810121620qe (mg g−1)0.370.550.610.770.921.221.36ModelsPseudo-first orderqt = qe [1-exp (-K1t)]qe (mg g−1)0.370.560.660.830.971.311.44k1 (min−1)0.050.040.030.020.020.010.01R20.920.910.920.910.930.960.91AICc− 51.40− 36.74− 32.50− 23.57− 24.61− 24.39− 8.31χ20.230.540.621.120.741.082.83Pseudo-second orderqt = qe (mg g−1)0.390.610.730.921.101.591.66k2 (g mg−1. min−1)0.200.110.060.030.020.000.01R20.950.940.950.940.960.970.93AICc− 63.19− 46.82− 42.12− 32.61− 35.72− 30.54− 13.88χ20.070.200.250.490.300.621.55Intra-particle diffusionC0.080.130.120.140.110.050.16Kdif (mg g−1 min0.5)0.020.030.040.040.050.070.08R20.820.850.900.930.960.970.95AICc− 38.99− 29.63− 32.11− 33.51− 36.55− 31.43− 24.06χ20.200.260.230.210.180.140.27 Parameters of kinetic models for various initial concentrations of CFTX. Pseudo-first order qt = qe [1-exp (-K1t)] Pseudo-second order qt = Intra-particle diffusion Although the analysis of FTIR spectra before and after adsorption does not reveal significant modifications, it nevertheless suggests the involvement of the C–O group in the adsorption mechanism, through the observation of a slight increase in the intensity of the characteristic band of the C–O, located in the region 1000–1200 cm-1. Adsorption isotherm Adsorption isotherms represent the equilibrium correlation between the quantity of adsorbate retained by the adsorbent (qe) and its concentration in the solution (Ce), at a fixed temperature. These isotherms offer information about the surface characteristics of the adsorbent and the affinity between the adsorbent and adsorbate, helping to understand the retention mechanisms of pharmaceutical compounds in aqueous environments. The experimental equilibrium data for CFTX adsorption on AC were fitted to various models, including thermodynamic (Langmuir, Freundlich, BETn-finite, and BETn-infinite) and statistical (STAT1, STAT2, and STAT3) models. The fitting curves of these models are illustrated in Fig. 5 and the corresponding adjusted parameters are listed in Table 7, along with statistical analysis criteria (nonlinear regression coefficients (R2), Akaike information criterion corrected for sample size (AICc), and chi-square (χ2)) to assess the adequacy of the model. Isotherms of experimental and thermodynamic adsorption (a) versus statistical (b) modelled CFTX on AC. Table 7Thermodynamic and statistic model parameters for CFTX adsorption isotherms.Model parametersCFTXR2AICcχ2Thermodynamic models Langmuirqmax (mg g-1)2.150.89− 1.350.92KL (L mg-1)0.31 FreundlichnF2.230.94− 8.150.14KF ((mg g-1)/ (mg L-1)1/n0.61 BETn-finiten51.410.93− 1.760.10qmax (mg g-1)0.39KL (L mg-1)7.20KS (L mg-1)0.07 BETn-infiniteqmax (mg g-1)0.780.93− 5.010.10KL (L mg-1)3.63KS (L mg-1)0.07Statistic models STAT1m0.150.95− 7.330.08qmax (mg g-1)0.44C1 (mg L-1)/ΔE01(kJ mol-1)15.67/0.59C2 (mg L-1)/ΔE02(kJ mol-1)35.02/− 1.96 STAT2m0.450.94− 5.400.14qmax (mg g-1)2.47C1 (mg L-1)/ΔE01(kJ mol-1)21.13/− 0.13C2 (mg L-1)/ΔE02(kJ mol-1)11.29/1.52 STAT3m0.410.94− 5.620.13qmax (mg g-1)13.17C1 (mg L-1)/ΔE01(kJ mol-11797.31/− 10.96C2 (mg L-1)/ΔE02(kJ mol-1)1275.33/0.83 Thermodynamic and statistic model parameters for CFTX adsorption isotherms. The results of the modelling show that the Freundlich thermodynamic model and the multilayer statistical model STAT1 are most appropriate for describing the adsorption of CFTX on the prepared AC compared to other models tested. Indeed, these two models have the highest regression coefficients (R2 = 0.94 for Freundlich and R2 = 0.95 for STAT1) as well as the lowest values of chi-square and the corrected Akaike information criterion for small samples (χ2 = 0.14 and AICc = − 8.15 for Freundlich and χ2 = 0.08 and AICc = − 7.33 for STAT1). These indicators (R2, χ2, AICc) confirm their better adequacy to experimental data. Freundlich’s model describes multilayer adsorption on heterogeneous surfaces with variable binding energies72. The intensity parameter, 1/n, is essential for interpreting the adsorption behavior: If n = 1 the adsorption is linear, indicating homogeneous sites with no interaction between adsorbed molecules. 1/n < 1, the adsorption is favorable, with increased adsorption capacity and the formation of new adsorption sites. In contrast, when 1/n > 1, the adsorption becomes unfavorable due to weaker adsorption bonds and reduced adsorption capacity. By applying the nonlinear Freundlich equation, 1/n value of 0.44 was obtained, suggesting that the CFTX retention on the produced AC is favorable. This result is consistent with observations of the AC’s SEM images, which clearly reveal the heterogeneity of its structure. In addition, the Freundlich curve shows that the adsorption has not reached saturation of AC, suggesting a multi-layer adsorption process73. The STAT1 statistical model, founded on a statistical physics approach applied to type II isotherms (according to the Brunauer, Emmett, and Teller classification), also assumes multilayer adsorption on the AC surface. It allows us to determine a stoichiometric coefficient (m = 0.15), representing the average number of CFTX molecules adsorbed as aggregates per adsorption site. This value reflects significant adsorbate–adsorbate interactions, as well as heterogeneity in the distribution of adsorbed molecules74,75. In addition, the STAT1 model provides energy parameters for monolayer (for n = 1) and multilayer (for n > 1) adsorption. Indeed, it allows to estimate the adsorption energies ΔE01 (monolayer) and ΔE02 (multilayer). For a CFTX saturation concentration of 20 mg L-1, the calculated adsorption energies are ΔE01 = 0.59 kJ mol-1 and ΔE02 = − 1.96 kJ mol-1. These results suggest that monolayer adsorption (CFTX-AC interactions) is endothermic in nature (ΔE01 > 0), while the multilayer adsorption process (CFTX–CFTX interactions) is exothermic (ΔE02 < 0). Adsorption thermodynamics The study of the thermodynamic behavior of CFTX retention on AC allows the evaluation of key thermodynamic parameters, such as ΔG°, ΔH°, and ΔS°, thus providing information on the energetic variations during adsorption, as well as the feasibility and spontaneity of this process. These thermodynamic parameters were determined by applying the following equations (Eq. 11–13)76: where R (8.314 J mol-1 K-1) is the universal gas constant, T (K) is the absolute temperature and KL is the thermodynamic equilibrium constant determined from the Langmuir isotherm. ΔG° (kJ mol-1) represents the Gibbs free energy change, ΔH° (kJ mol-1) is the enthalpy change, and ΔS° (kJ mol-1 K-1) is the entropy change. According to the Van’t Hoff equation (Eq. 13), ΔS° and ΔH° were determined from the slope and the intercept, respectively, of the linear plot of ln (KL) as a function of (1/T) (Fig. 6). The measured thermodynamic parameters are summarized in Table 8. Ln KL versus 1/T diagram. Table 8Thermodynamic parameters (ΔG°, ΔH°, and ΔS°) of CFTX retention on AC.Temperature (°C)ΔG° (kJ mol-1)ΔH° (kJ mol-1)ΔS° (kJ mol-1 K-1)30− 30.2250.110.2635− 31.5540− 32.87 Thermodynamic parameters (ΔG°, ΔH°, and ΔS°) of CFTX retention on AC. The negative ΔG° values indicate that the retention of CFTX on AC is thermodynamically favorable and can occur spontaneously, without requiring external energy input77. Moreover, the increasingly negative ΔG° values with rising temperature suggest that the retention process becomes more favorable at higher temperatures. The positive ΔH° value indicates that the adsorption process is endothermic, implying that higher temperatures enhance the adsorption of CFTX on AC 78. This endothermic behavior is explained by the increased mobility of the adsorbed molecules across the external boundary layer and into the internal pores of the adsorbent. A similar endothermic behavior was also observed in a previous study on the antibiotics adsorption on AC 79. The nature of the adsorption process can be interpreted according the magnitude of ΔH°. A ΔH° value between 0 and 40 kJ·mol⁻1 indicates physisorption, while a value between 40 and 800 kJ·mol⁻1 is indicative of chemisorption 80. Based on this and considering the obtained ΔH° value (50.11 kJ·mol⁻1), the CFTX adsorption on AC can be classified as a chemisorption process. These results corroborate the conclusions of the kinetic study. The positive ΔS° value (0.26 kJ mol-1 K-1) indicates an increase in the randomness at the adsorbent surface during CFTX retention, reflecting a strong affinity of AC for CFTX 81. Study of desorption The stability and reusability of an adsorbent are key parameters for its practical use in various fields. The durability of AC prepared from grape seeds waste was evaluated through an adsorption–desorption regeneration test which is shown in Fig. 7. The results show that the desorbed quantity stabilizes at approximately 0.11 mg g-1 from an initial concentration of , suggesting that only a limited fraction of the adsorbed CFTX can be desorbed. In addition, the desorption rate significantly decreases from 25% to 6.46% when the CFTX concentration increases from 4 to 20 mg L-1. This behavior indicates that the adsorption becomes less reversible at higher concentrations, likely due to the establishment of stronger interactions between CFTX and the prepared AC, which makes desorption more difficult and reduces the efficiency of the recovery process. These observations corroborate the results achieved by kinetic and thermodynamic studies, which suggest a chemisorption mechanism. These findings also highlight the importance of retention mechanisms when designing adsorbent materials for remediation applications, while highlighting the challenges associated with the regeneration of AC saturated with a high amount of adsorbate. Study of CFTX desorption. DFT modelling Theoretical calculations of the interactions that occur in the presence of AC require a precise understanding of its atomic structure. In this study, four cycles of fused benzene (the chrysene) were used to model the AC surface. To assess the impact of functional groups, –OH, –CHO, and –COOH groups were incorporated into the active sites, and selected based on the Boehm titration results82,83. Study of global reactivity of molecules The electronic properties of Ceftriaxone (CFTX) and functionalized activated carbon (AC) were studied to investigate molecular interactions and affinity between adsorbate and adsorbent. Figure 8 shows the molecular electrostatic potential (MEP) and the positions of the HOMO and LUMO orbitals of the optimized AC models. These models are designated as follows: AC (unfunctionalized AC), AC–COOH (AC functionalized with –COOH), AC–OH (AC functionalized with –OH), AC–CHO (AC functionalized with –CHO) and AC–COOH–OH–CHO (AC functionalized with –COOH, –OH, and –CHO), as well as Ceftriaxone (CFTX) molecule. MEP (e/au3 scale) and HOMO/LUMO molecular orbital mapping of optimized geometric structures of: (a) CFTX, (b) AC, (c) AC–CHO, (d) AC–COOH, (e) AC–OH and (f) AC–COOH–OH–CHO, calculated by DFT at the level of B3LYP/6–311G (d, p). MEP is a well-established calculation method for studying chemical reactivity and variations in electron density within a molecule. It allows us to predict the reactive sites that are favorable to nucleophilic and electrophilic attacks84. The MEP map represents, through a color gradient, the regions where the electrostatic potential is negative, positive, and neutral; where the red color indicates the areas of most negative potential (electron-rich) favorable to electrophilic attacks while the blue color corresponds to the most positive potential (electron-poor), associated with nucleophilic reactivity. Regions with neutral potential are shown in green85. On the MEP map (Fig. 8a), it is observed that the electron-rich zones of CFTX are concentrated mainly around carbonyl groups (–C=O), which are likely to interact with the functional groups of AC, acting as preferred adsorption sites. On the other hand, electron-poor areas are located near hydrophobic groups. For functionalized ACs, the –CHO, –COOH, and –OH groups have electron-rich zones, promoting electrostatic interactions and hydrogen bonds with CFTX. The analysis of the frontier molecular orbitals, namely the HOMO and LUMO, is fundamental for a better understanding of chemical reactivity at the atomic level. It allows to study the charge transfer mechanisms during adsorption, by identifying regions of a molecule that can give or accept an electron, respectively 86. The HOMO and LUMO energy gap (Egap) influences the reactivity and chemical stability of a molecule, a narrower gap usually corresponds to a higher reactivity. As shown in Fig. 8, CFTX has an energy gap of 2.14 eV, indicating a moderate reactivity, with HOMO orbitals located on the electron donor amine groups and LUMO orbitals on the carbonyl acceptors. In contrast, the unfunctionalized AC shows a higher energy gap (2.71 eV), reflecting a lower intrinsic reactivity. However, the functionalization of carbon gradually reduces the energy gap, reaching 2.40 eV, for AC_COOH–OH–CHO, suggesting an increased reactivity due to the synergy between the functional groups. Chemical descriptors, such as chemical potential (μ), global hardness (η), and electrophilic index (ω) provide additional information on the stability and reactivity of molecules. A high overall hardness (η) indicates lower reactivity, while a high electrophilic index (ω) suggests a greater ability to accept electrons87. Gap energies (Egap) and the chemical descriptors for CFTX and functionalized carbons are calculated and summarized in Table 9. Table 9Quantum chemical descriptor values for monomers (CFTX molecule and functionalized AC models before adsorption) and for dimers (CFTX/AC complexes representing the molecular situation after adsorption).Quantum chemical parametersEnergy gap(Egap, eV)Chemical potential(μ, eV)Global hardness(η, eV)Electrophilicity index(ω, eV)MONOMERS CFTX2.146.391.0719.03 AC2.717.741.3522.11 AC–CHO2.577.821.2823.76 AC–COOH2.657.771.3222.77 AC–OH2.537.611.2622.84 AC–COOH–OH–CHO2.407.721.2024.84DIMERS CFTX_AC1.226.900.6138.95 CFTX_AC–CHO0.887.030.4455.77 CFTX_AC–COOH1.036.920.5146.25 CFTX_AC–OH1.256.860.6237.56 CFTX_AC–COOH–OH–CHO0.997.000.4949.48 Quantum chemical descriptor values for monomers (CFTX molecule and functionalized AC models before adsorption) and for dimers (CFTX/AC complexes representing the molecular situation after adsorption). Adsorption mechanisms Figure 9 presents the optimized complexes formed by the coordination of CFTX with both unfunctionalized and functionalized AC, along with their corresponding interaction energies. The energy gap values and related chemical descriptors for each complex are summarized in Table 9. Optimized structures of CFTX/AC complexes and their adsorption energies: (a) CFTX_AC, (b) CFTX_AC–CHO, (c) CFTX_AC–COOH, (d) CFTX_AC–OH and (e) CFTX_AC–COOH–OH–CHO. After adsorption, a significant decrease in the energy gap is observed for all complexes, with the CFTX_AC–CHO complex exhibiting the lowest value at 0.88 eV. This reduction suggests enhanced stability of the adsorbed systems, attributed to efficient charge transfer from the HOMO of CFTX to the LUMO of the functionalized carbon surface. The calculated chemical potential (7.03 eV) and global hardness (0.44 eV) further confirm this stabilization, indicating increased reactivity and improved thermodynamic stability of the complexes post-adsorption. The most negative adsorption energy recorded for the CFTX_AC–CHO complex () confirms that this functional group promotes stronger interactions with CFTX, mainly via electrostatic interactions and hydrogen bonds. This result suggests that the functionalization of AC by the –CHO group significantly improves the adsorption process, compared to the –OH and –COOH groups located on the AC surface. The CFTX_AC–COOH–OH–CHO complex (Fig. 9e) also shows particularly favorable adsorption energy (− 4.47 kcal mol-1), highlighting strong and synergistic interactions between the –COOH, –OH, and –CHO groups and the polar sites of the CFTX. This synergy results from the combination of the specific effects of each functional group, which increase the local electron density and thus facilitate interactions with CFTX. AIM, NCI, and RDG analyses were employed to examine the intensity and nature of molecular interactions involved in adsorption within CFTX_AC–CHO and CFTX_AC–COOH–OH–CHO complexes. These methods provide a better understanding of interactions, particularly non-covalent bonds, thus contributing to the interpretation of adsorption mechanism and the stability of complexes formed 88. Non-covalent interactions can be visualized on the NCI and RDG maps. Green regions represent weakly attractive interactions, mainly of the Van der Waals type, while blue areas indicate stronger attractive interactions, such as hydrogen bonds, and the red zones reveal repulsive interactions, often attributed to steric hindrance. The NCI and RDG graphs of the two studied complexes, CFTX_AC–CHO and CFTX_AC–COOH–OH–COOH, are shown in Fig. 10. NCI and RDG molecular graphs of CFTX_AC–CHO (a) and CFTX_AC–COOH–OH–CHO (b). In the CFTX_AC–CHO complex (Fig. 10a), small red spots indicating steric repulsion, are primarily observed at the center of the phenolic cycles of AC. A blue spot appears at the oxygen atom of the –CHO group in AC and the hydrogen atom of the CFTX, indicating the presence of a hydrogen bond in this region. In addition, low Van der Waals-type interactions, shown in green, are observed between CFTX and AC. In the CFTX_AC–COOH–OH–CHO complex (Fig. 10b) there are also red spots, indicating steric repulsions, as well as blue zones corresponding to hydrogen bonds. The figure clearly shows green areas representing weak Van der Waals interactions between CFTX and AC. AIM analysis corroborates these observations by identifying critical binding points (BCP, Bond Critical Points), which clearly confirms the presence of chemical interactions and bonds between the CFTX and the functional groups in the AC. Topological properties of intermolecular BCP, such as electron density ρ(r), the Laplacian of electron density ∇2ρ(r), total electronic energy density H(r), Lagrangian kinetic energy G(r) and potential energy density V(r), have been determined to assess the nature and intensity of binding interactions between molecules. In addition, the intermolecular bond energies at BCP (Eint) and the hydrogen bond interaction energies (EHB) were calculated using equations (Eq. 14)89 and (Eq. 15)90 respectively: Figure 11 illustrates the AIM molecular graph, highlighting the different BCP identified in the two complexes. The measured topological parameters associated with these BCP are listed in Table 10. AIM molecular graphs of the complexes CFTX_AC–CHO (a) and CFTX_AC–COOH–OH–CHO (b). Table 10Topology parameters at BCP, as well as interaction energies and hydrogen bonding energies in the CFTX_AC–CHO and CFTX_AC–COOH–OH–CHO complexes.BCPConnected atomsρ(r) (a·u)∇2ρ(r) (a·u)H(r) (a·u)G(r) (a·u)V(r) (a·u)Eint(kcal mol-1)EHB(kcal mol-1)CFTX_AC–CHO 114H38–O854.78 .10–21.04 .10–1− 1.52 .10–24.13 .10–2− 5.65 .10–2− 17.73− 9.91 141H86–O81.22 .10–25.69 .10–22.48 .10–31.17 .10–2− 9.26 .10–3− 2.91− 1.99CFTX_AC–COOH–OH–CHO 136C7–O796.18 .10–32.96 .10–21.66 .10–35.75 .10–3− 4.09 .10–3− 1.28– 138H38–O793.51 .10–21.07 .10–1− 6.98 .10–33.38 .10–2− 4.08 .10–2− 12.79− 7.10 141O8–O821.19 .10–26.86 .10–22.44 .10–31.47 .10–2− 1.23 .10–2− 3.85– 151N2–H884.63 .10–32.09 .10–21.20 .10–34.03 .10–3− 2.83 .10–3− 0.89− 0.29 157H40–C674.83 .10–31.71 .10–27.74 .10–43.49 .10–3− 2.72 .10–3− 0.85− 0.33 166O27–C844.42 .10–32.28 .10–21.45 .10–34.26 .10–3− 2.81 .10–3− 0.88– 183H47–C601.33 .10–25.10 .10–21.86 .10–31.09 .10–2− 9.03 .10–3− 2.83− 2.23 199O30–C703.27 .10–31.91 .10–21.31 .10–33.46 .10–3− 2.15 .10–3− 0.670.01 222H49–H751.98 .10–38.16 .10–35.87 .10–41.45 .10–3− 8.68 .10–4− 0.270.30 Topology parameters at BCP, as well as interaction energies and hydrogen bonding energies in the CFTX_AC–CHO and CFTX_AC–COOH–OH–CHO complexes. The results of the AIM reveal that, in the two studied complexes, the adsorption of CFTX on AC results from non-covalent interactions, characterized by low values of ρ(r) at the level of the BCP and positive values of ∇2ρ (r). These characteristics confirm the non-covalent nature of the interactions, mainly of hydrogen and Van der Waals bond type 91. In the CFTX_AC–CHO complex, two significant BCP (114 and 141) were identified, presenting low values of ρ(r) and positive values of ∇2ρ (r), indicating non-covalent hydrogen bond interactions. According to the classification proposed by Rozas et al.92, the strength of hydrogen bonds can be classified into three main types: Strong hydrogen bond definite by ∇2ρ (r) < 0 and H(r) < 0; Strong hydrogen bond definite by ∇2ρ (r) < 0 and H(r) < 0; Mean hydrogen bond characterized by ∇2ρ (r) > 0 and H(r) < 0; Mean hydrogen bond characterized by ∇2ρ (r) > 0 and H(r) < 0; Weak hydrogen bond when ∇2ρ (r) > 0 and H(r) > 0. Weak hydrogen bond when ∇2ρ (r) > 0 and H(r) > 0. As shown in Table 10, the CFTX_AC–CHO complex is mainly stabilized by a moderate hydrogen bond between the H38 atom of the CFTX and the O85 atom of the -CHO group of the AC, as well as by a weaker interaction between H86 and O8. The H38—O85 bond, characterized by a higher hydrogen bond energy EHB (EHB = − 9.91 kcal mol-1), is significantly stronger than the H86–O8 bond (EHB = − 1.99 kcal mol-1). These observations are corroborated by the total energy of the complex, estimated at − 11.90 kcal mol-1, which is almost equivalent to the adjusted adsorption energy found (− 13.07 kcal mol-1), confirming that the CFTX retention is primarily stabilized by hydrogen bonds. In the CFTX_AC–COOH–OH–CHO complex, several BCP have been identified, revealing a diversity of interactions. The interactions at BCP 136, 141, 166, and 199 show moderate values of ρ(r) and positive values of ∇2ρ (r), suggesting non-covalent Van der Waals-like interactions or a very weak hydrogen bond, with low interaction energies between − 0.67 kcal mol-1 and . The H38–O79 interaction at BCP138 is distinguished by an EHB energy of − 7.10 kcal mol-1, representing the strongest hydrogen bond and significantly contributing to the stabilization of the complex. In both complexes, the adsorption process is mainly governed by hydrogen bonds between a hydrogen atom of the CFTX (H38) and an oxygen atom of the AC (O85 or O79), presenting the highest bond energies. These results show that adsorption is optimized by diversification of non-covalent interactions, where hydrogen bonds play an important role, while Van der Waals interactions, although weaker, also contribute to overall stabilization, particularly in the CFTX_AC–COOH–OH–CHO complex. The combined results of the AIM, NCI, and RDG analyses reveal that the retention mechanism of CFTX on AC is driven by hydrogen bonds and Van der Waals interactions. These interactions are essential for the stabilization of the formed complexes. Conclusion The present research focused on the elimination of Ceftriaxone (CFTX), an antibiotic considered as an emerging contaminant in aqueous solutions, through the employment of novel AC synthesized from grape seeds waste. The synthesized carbon was characterized using various physico-chemical methods, which demonstrated well-established porosity offering a large exchange surface of 543 m2 g-1. Batch adsorption studies, coupled with theoretical models, allowed us to elucidate the mechanisms of adsorption. Kinetic and equilibrium data demonstrated strong agreement with the PSO model and Freundlich’s thermodynamic model, as well as with the multilayer statistical model STAT1, respectively. Thermodynamic parameters suggest that the adsorption process is promoted, endothermic, and mainly controlled by chemical interactions. Desorption studies confirmed a chemisorption mechanism between CFTX and AC, characterized by desorption, thus opening the possibility of exploring an alternative approach to the adsorbent regeneration process. Theoretical calculations based on the DFT at B3LYP/6-311G (d, p) level of theory, were applied to identify the sites responsible for the adsorption of CFTX and to further understand the mechanisms of their elimination by synthesized AC, by estimating the electronic properties of the molecules involved. The results highlighted that the –CHO functional group presents a stronger interaction with CFTX compared to other functional groups on the AC surface, with an adsorption energy of − 13.07 kcal mol-1, reflecting favorable adsorption. Moreover, the DFT simulations also revealed that the functionalization of AC models with –COOH–OH–CHO groups leads to an interaction with CFTX characterized by a favorable adsorption energy of -4.47 kcal mol-1. These moderate adsorption energies obtained show stable but not very strong interactions, which explains the low value of the adsorption capacity qe of the CFTX by the developed AC. The AIM NCI and RDG analysis of two complexes CFTX_AC–CHO and CFTX_AC–COOH–OH–CHO showed that adsorption is mainly governed by strong hydrogen bonds. These hydrogen bonds observed justify the limited desorption observed experimentally. In summary, this study demonstrates that using an AC synthesized from grape seeds waste can be a powerful and an ecological material for the elimination of CFTX in water. These results highlight the combination of experimental and theoretical approaches for a deep understanding of the adsorption process, which may lead to new applications in the treatment of other emerging pollutants. These results also highlight the value of using agricultural waste to produce low-cost adsorbent materials, which offer a sustainable and eco-friendly alternative for water depollution. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements The authors would like to thank the Tunisian Higher Education and Scientific Research Ministry for the financial support. Author contributions S M: Writing—review and editing, Writing—original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization, Software. HD: Writing—review and editing, Resources, Project administration, Methodology, Software, Investigation, Formal analysis, Conceptualization, Validation. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Data availability The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request and will be made publicly available in a suitable data repository upon acceptance of the manuscript. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request and will be made publicly available in a suitable data repository upon acceptance of the manuscript.