Preparation, characterization, and application of modified magnetic biochar for the removal of benzotriazole: process optimization, isotherm and kinetic studies, and adsorbent regeneration

Benzotriazole (BTAs) is a polar conditioning chemical (CCs) that possesses a low log K_ow value of 1.23, which it presents in sediments and soil, surface water, and groundwater. BTA is commonly applied as a corrosion inhibitor in engine coolants, anti-freeze liquids, detergent additives, aircraft de-icing and anti-icing substances, hydraulic fluids, and a stabilizer for bronze commodities (Janna et al. 2011; Li et al. 2012; Mezzi et al. 2012; Grillo et al. 2013). However, it is classified as an emerging pollutant, and it is known to be toxic to vertebrates (LC₅₀ of 102 mg/L for water fleas Ceriodaphnia dubia); also, it has adverse effects such as anti-estrogenic and carcinogenesis effects (Harris et al. 2007; Pritchard et al. 2018; Richardson & Ternes 2018). BTA and its derivatives are regularly discharged in municipal/industry wastewater, resulting in high natural surface/groundwaters (Xu et al. 2014). They are often contaminated with BTA due to their high solubility in water, limited removal efficiency, and extensive applications (Giger et al. 2006). Nonetheless, limited information is available about the possible release of BTA into the environment, particularly in water resources (Richardson & Ternes 2018).

BTA removal is limited and challenging in wastewater treatment plants due to its non-biodegradability and resistance to disintegration (Rhodes-Dicker & Passeport 2019). Therefore, diverse methods including oxidation processes (Mawhinney et al. 2012), membrane bioreactors (Weiss & Reemtsma 2008), photoelectrocatalytic degradation (Ding et al. 2010), and adsorption (Xu et al. 2014) have been employed for the removal of BTA from aqueous solutions. However, the application of these technologies is limited by some factors such as low efficiency, detrimental byproducts, and high operation costs (Bernal-Martínez et al. 2013).

Among different BTA removal systems, adsorption is an efficient and low-cost process. To date, a number of adsorbents, such as carbonaceous materials (Bonvin et al. 2016), zeolites (Wang et al. 2017), and metal oxides have been used aiming at BTA adsorption from aqueous solutions (Sarker et al. 2017); however one of the most important aspects related to water and wastewater treatment is finding cost-effective adsorbents. Recently, the utilization of cheap carbon adsorbents (Biochar) from waste products and low-cost compounds has been prevalent (Dai et al. 2019). Biochar is a carbon-rich solid characterized by porous structure, large surface functional groups, and mineral components, making it a proper adsorbent for removing pollutants from aqueous solutions (Dai et al. 2019). Nonetheless, bare biochar has a low adsorptive capacity to adsorb contaminants in water and sewage treatment. Moreover, powdered adsorbents are tenacious to separate from effluents after adsorption. Separation via filtration and centrifugation methods may not completely isolate the adsorbent from the effluent; therefore, magnetized biochar has been considered a suitable solution to tackle this challenging problem (Khan et al. 2020). On the other hand, adsorption processes are mainly affected by various operational factors, including contact time, adsorbent dosage, initial concentration of contaminant, etc.; it is beneficial to utilize appropriate experimental methods. These mathematical and statistical methods can create a systematic path that permits the final goal (removal efficiency) to be based on influential parameters, which are the most dominant in the response factor, by the at least possible number of experiments for decline process expenditure and required time to developing of process installation (Haghighi et al. 2019). Response surface methodology (RSM) is one of the effective multivariate experimental design methods. This approach comprises several mathematical and statistical techniques that are employed to develop, improve and optimize various processes and analyze the relative significance of each operational condition even in the presence of a complex network of interactions (Najafpoor et al. 2019).

Several studies have employed modified magnetic biochar (MMBC) for the adsorption of a wide range of pollutants such as nitrate and phosphate (Yin et al. 2018), tetracycline (Zhou et al. 2019), methylene blue (Liu et al. 2019), 2,4-dichlorophynoxyacetic acid (Zhu et al. 2018), cadmium adsorption (Cui et al. 2019), and Remazol Brilliant Orange (Gokulan et al. 2019). However, to the best of our knowledge, few studies have focused on optimizing the BTA adsorption process using MMBC, modified with different modifiers and their efficiency compared with each other and its regeneration. In the present research, a porous MBC adsorbent was synthesized and characterized using various techniques (Najafpoor et al. 2019). The composite was modified with multiple chemicals, and then the best-modified composite was chosen for main experiments. The RSM based on central composite design (CCD) was adopted for modeling, analyzing, and optimizing the BTA removal process by MMBC as a function of three independent variables (reaction time (min), adsorbent dosage (g/L), and initial BTA concentration (mg/L)). The importance of each independent variable on the variation of the output response (BTA removal efficiency (%)) was also determined. Moreover, isotherm, kinetic, thermodynamic studies were conducted. Finally, MMBC adsorbent reusability study via different regenerators was investigated in detail.

BTA with the formula C₆H₅N₃ (Table 1), sodium hydroxide (NaOH), condensed sulfuric acid H₂SO₄, ZnCl₂, sodium dodecyl sulfate (SDS), and per acetic acid (PAA) were purchased from Merck Co. (Germany). All chemicals were of analytical grade. Ultrapure water was used for preparing solutions.

For biochar synthesis, Oakwood residues were rinsed, dried at 105 °C for 24 hours, crushed, and sieved (5-mm) to remove big lumps; then, it was stored in a sealed container an incubator for the experiment. Next, the biomass powder was heated from room temperature to 700 °C with a heating rate of 5 °C/min. The pyrolysis process was performed using a tube furnace with continuous nitrogen gas purging at 30 ml/min and the heating rate was constant at 15 °C/min. For promoting the functional group on biochar surface, the achieved biochar (2.5 g) was reacted with ZnCl₂ solution (the weight ratio of 2:1) at 100 °C for 2 h. Then, the reaction mixture was added to deionized water (100 mL), and precipitates were separated by filtration. The excess ZnCl₂ was washed by applying dilute hydrogen chloride. Subsequently, carbon was repeatedly washed to eliminate excess ZnCl_2, and was then dried at 60 °C for 12 h. Modification of BC with other chemicals obeys such a mentioned process and weight ratio.

An aqueous solution of NH₃ (15 mL, 25%) was added dropwise with powerful stirring under nitrogen flow for 1.5 h to a mixture of MBC (2.0 g), FeCl₃.6H₂O (2.7 g), FeCl₂.4H₂O (2.5 g), and deionized water (60 mL). Next, the black solid was gathered by an external magnet, washed with distilled water (three times), and dried at 60 °C for 5 h. The obtained ZnCl₂/MBC was utilized for the BTA adsorption process.

For evaluation of the second-order polynomial model reliability, normal plots, residual examination, and ANOVA were performed. In addition, adjusted R-square and F-test were applied to check the quality and statistical significance of the model, respectively. The F-test and Student's T-test were used to test the importance of the model at a 95% confidence level. Furthermore, the residuals (the differences between experimental and predicted values) were analyzed through residual and normal probability plots to examine the constant variance of errors; also normal distribution was assessed to detect possible systematic departures from the assumption.

This study used the Design-Expert (version 8.0.0) software package to design experiments, model-fitting, and data analysis.

Concerning the adsorptive remediation of water and wastewater containing contaminants, it is essential to know the removal rate for the design and the quantitative evaluation of the adsorbent. In addition, kinetics describes the adsorbate uptake rate, which controls the residence time of the adsorbate uptake at the adsorbent–solution interface. Therefore, it is vital for prediction of the BTA uptake removal rate from aqueous solutions in order to design an appropriate adsorption unit.

The BTA adsorption data were fitted to the non-linear kinetic and isotherm models using MATLAB^® 7.11.0 (R2015b), with subsequent interactions calculated by the Levenberg–Marquardt algorithm. Since the unwanted falsification of error distribution occurs due to data transformation to a linear form, the non-linear method is superior to the linear one in determining the parameters of the isotherm and kinetic models. In this study, five widely used adsorption isotherm models (Langmuir, Freundlich, Redlich–Peterson, Temkin, and Liu) and five general adsorption kinetic models (pseudo-first-order equation of Lagergren, pseudo-second-order equation of Ho, Elovich, Avrami fractional order, and the intraparticle diffusion model) were used to describe the adsorption equilibrium and the adsorption kinetics of BTA onto MMBC, respectively (Vilardi et al. 2018; Zhang et al. 2018; Jeon 2019; Rezakazemi & Shirazian 2019; Syafiuddin et al. 2019). All these mathematical isotherms and kinetic models are summarized in Supplementary Table 1.

It is essential to evaluate their validity to select the most suitable kinetic and isotherm model. Here, the validity of the kinetic and isotherm models at different temperatures (283, 298, and 313 K) was assessed by criteria such as the determination coefficient (R²), the adjusted determination coefficient (R²adj), the sum squared error (SSE), and the root mean square error (RMSE) (Rangabhashiyam et al. 2014; Suganya 2019). These criteria describe the goodness of fit between the experimental and predicted data. The best model was chosen based on the lowest RMSE, SSE, and R²adj and R² as close as possible to 1. R², R²adj, SSE, and RMSE were calculated according to Supplementary Table 2.

To evaluate the possibility of MMBC regeneration and reuse, methanol, as a desorbing solution, was utilized to extract the BTA adsorbed on MMBC. The reusability of the adsorbents was determined using five adsorption–regeneration cycles. A sample of 1.0 g of MMBC was shaken with 1 L solution of 50 mg/L BTA for 1 h at 25 ± 1 °C and pH = 7.0. The BTA-loaded MMBC was magnetically collected, washed, and shaken at 200 rpm for 24 h with 10 mL of desorbing solutions methanol at 25 ± 1 °C. After desorption, regenerated adsorbents were dried in an oven at 80.0 °C for 100 min and used for the next adsorption-regeneration cycle.

Table 3 represents some physicochemical characteristics of the adsorbent. The elemental analysis demonstrated that C (49%) is the dominant element in the composite structure; also, the presence of O (18.43%) and Fe (9.42%) indicated that the magnetization agent had been loaded correctly on biochar. Moreover, results confirmed the existence of Zn (9.07%) and Cl (13.85%) as a modifier in the MMBC structure. The specific surface area of MMBC (383.5 g/m²) was determined by BET analysis; this high value can be attributed to the increased surface area of loaded iron oxide nanoparticle and modification by ZnCl₂, which is beneficial to the increasing adsorption capacity of biochar in BTA adsorption; whereas, the specific surface areas for BC and MBC were 210 and 297 g/m², respectively. Moreover, the morphology of the samples was examined with an FEI Quanta 400 FEG scanning electron microscope (SEM), and the SEM image of MMBC is illustrated in Figure 1(a) and 1(b). Based on Figure 1(a), it is evident that the BC surface was relatively smooth and had a porous structure, without significant development on its surface. However, based on the SEM images (Figure 1(b)), the produced modified biochar had a relatively rough, much more porous structure than BC.

Moreover, Fe₃O₄ nanoparticles tended to aggregate on the BC surface, and a vigorous aggregation is evident after magnetization. This phenomenon can create a rough and uneven surface followed by increased adsorption sites for BTA adsorption. It is worth noting that the white parts in the SEM may be attributed to the zinc salt residues. It seems that the cavities on the surface of biochar resulted from the ZnCl₂ evaporation during carbonization, leaving the space previously occupied by zinc chloride. The pore-forming effect of ZnCl₂ decomposition during heat treatment could promote the organic compound breakage in the precursor of BC (Oakwood residues) residues and gradually recombine the solid matrix to form an extended porous structure, which could accelerate the reaction rate, raise the surface area, and provide more active sites for the adsorption of BTA (Yan et al. 2020). Energy dispersive X-ray (EDX) analysis, including spectrum and elemental composition (Figure 3), illustrating BC/Fe₃O₄ nanocomposite contains peaks responding to the C, Fe, O, Cl, and Zn atoms. The peaks of Fe (9.42%) and O (18.43%) are related to Fe₃O₄, while the C (49.23%) atom is associated with BC. In addition, Cl (13.85%) and Zn (9.07%) confirmed the modification of MBC by ZnCl₂.

Raman spectroscopy (PerkinElmer Raman Station 400F dispersive Raman micro-spectrometer with a CCD detector) was applied to investigate the chemical functionality and mineralogy of biochar. Raman was conducted at the wavelength of 500–300 cm⁻¹ (Figure 1(d)). The highest peak was observed at 1,596 cm⁻¹ and 1,325 cm⁻¹. The peaks at 799 and 1,017 cm⁻¹ were assigned as a D-band, as well as the peaks at 1,332 cm⁻¹ and 1,586 cm⁻¹ are related to G-band peaks in the Raman spectra. Different biomass-derived biochar has typical characters associated with the structural disordered aromatic ring in sp² carbon atoms or sample defective graphite structure (D-band), in contrast, G-band presents completeness of the degree of sp² carbon atoms with less defective graphene structures (Liu et al. 2021). Additionally, Fourier transform infrared (FTIR) spectra of the MMBC before and after reaction with BTA are illustrated in Figure 1(e) and 1(f). The FTIR spectra of biochar were recorded in 4,000–500 cm⁻¹. The peak at 3,569 cm⁻¹ is a firm indicator of the OH phenol functional group onto the surfaces of the adsorbent. The two sharp peaks at 1,402 cm⁻¹ and 1,617 cm⁻¹ are related to the amine functional group and the symmetric bending of CH₃, respectively. Based on Figure 1, among the surface functional groups, –OH groups, secondary amine group, C = O stretching of ether group, aliphatic C–H group, C–N stretch of aliphatic amines, and symmetric bending of CH₃ had a considerable influence on the adsorption of BTA.

The effects of various parameters (modifier types, biochar size, ZnCl₂ concentration, and aqueous solution pH) on BTA removal by prepared composite are presented in Figure 2(a)–2(d). The obtained results from the evaluation of several modifier efficiencies showed that ZnCl₂ had the highest performance in BC modifying for BTA adsorption. Chemical activation by ZnCl₂ presumably provides much more porous and higher surface area in activated carbons, thereby providing a high adsorption capacity. The reaction between carbon atoms and ZnCl₂ (dehydrating agent) promotes carbonaceous materials’ decomposition in the extended carbon interlayers (Şahin et al. 2015). The use of ZnCl₂ in chemical activation improves the carbon content through the aromatic graphitic structure formation. The obtained results are consistent with findings in similar studies (Xia et al. 2016). As shown in Figure 2(b), MMBC particle size plays a vital role in BTA adsorption. As with decreasing the ZnCl₂-MBC size, the BTA removal increases, maximum BTA removal was attained at the size of <150 μm. Reducing adsorbent particle size increases specific surface area, followed by the surface energy and enhanced BTA removal efficiency. In addition, this trend is also related to the adsorbent surface area and the BTA diffusion rate. Generally, assuming that the adsorption rate depends only on the surface area (the surface area of the smaller particle is high); hence, the dispersion way is shortened in the small adsorbent particle, and it provides a better opportunity to make the adsorbed BTA penetrate all internal pore structure. Similar results have been reported previously (Liu et al. 2021).

In addition, increasing the ZnCl₂ concentration from 5 to 100 g/L produced BTA removal percentage growth from 76 to 94%, but it remained constant after increasing the concentration up to 200 g/L. This was possible because increase in ZnCl₂ concentration reaction may cause a part of the micro-pore structure to be destroyed due to the pores collapsing (Şahin et al. 2015). A similar study was conducted by Şahin et al. (2015). According to Figure 2(d), the highest BTA adsorption efficiency was obtained at neutral pH 7–8. pH played a significant role in the organics’ adsorption from aqueous solutions because of the protonation/deprotonation of adsorbates and the alteration in the surface charge of adsorbents by pH changes (Sarker et al. 2017). BTA is a heterocyclic aromatic compound; therefore, π–π interactions between the aromatic ring of BTA and adsorbent might be possible. The hydrophobic and π–π interactions are not highly dependent on pH. The declined adsorption of BTA with increasing pH might be explained by a partial contribution of electro-static repulsions between the negative surface charge of adsorbent and the negative charge of the deprotonated BTA. Similarly, the decrease of q_t at an acidic pH might be because of repulsion between the positive adsorbent and positive or protonated BTA (Sarker et al. 2017).

Subjected to −2 X_i + 2

The experimental and predicted values of BTA removal at various experimental conditions are illustrated in Table 4. The highest removal percentage was >99.9% in the 1st, 10th, 15th, 26th and 29th runs. Moreover, the results of ANOVA (Table 5) approved the ability of the employed model to design the optimization experiments. The relationship between predicted values of BTA adsorption efficiency against the experimental ones was also evaluated. Logically, since the coefficient of determination (R²) is close to 1, the compatibility of the applied model is apparent. In addition, the high R² (0.987) confirms the robust correlation between the experimental and predicted data of BTA removal efficiency. It can be postulated that the model mentioned above could not predict only 1.13% of data. If the F-value of the model is more remarkable than tabulated F for a specific domain of degree of freedom, the proposed model is quite valid. As can be seen from Table 5, the F-value of the model (298.51), which is the proportion of model mean square and residual error, is higher than the tabulated F-value that approves the statistical significance of the model. F-value for Lack of Fit was 22.08 (significant at <0.0001).

The normal probability plot of the data is shown in Figure 3(a). The apparent trend in the mentioned graph apparently has a normal distribution form which resembles a straight line. The plot of residuals versus predicted values (Figure 3(b)) revealed a randomized scattering trend in the adjacency of the centerline.

The surface and contour plots as a promising method were drawn to illustrate BTA adsorption onto the MMBC surfaces at different chosen experimental variables. In Figure 4, the adsorption efficiency of BTA as a function of the combined effects of adsorbent dosage and reaction time is shown. Based on Figure 4, an intensive increase in BTA adsorption occurred by increasing the reaction time from 2 to 60 min. It can be concluded that at the beginning of the experiment, due to the abundance of reactive sites, a large amount of BTA can be adsorbed, while the tendency of the adsorbent to sorb the adsorbate eventually declined due to saturation of the reactive sites. Additionally, the formation of insoluble silicates on the MMBC prevented the BTA from entering the internal sorption layer. The intraparticle diffusion in this phase may also be corresponding to a slower adsorption rate. A similar trend was observed previously in another work (Gwenzi et al. 2018; Zhu et al. 2020).

The interaction effects of reaction time and initial BTA concentration on the adsorption efficiency of BTA are depicted in Figure 5. According to Figure 5, at the constant MMBC concentration, increasing initial BTA concentration caused a significant decrease in the BTA adsorptive removal. The high concentration of BTA provides a thick layer of it onto the surfaces of modified biochar which certainly prevented BTA adsorption by the adsorbent (Srivastava et al. 2006). This trend can be explained by binding all BTA molecules with a composite surface at low concentration, while at high concentration, the available adsorption positions are reduced with build-up of BTA molecules on the surface of MMBC; this prevents the diffusion of more BTA molecules into the adsorbent pores. The low diffusion rate is related to pores, which are similar to the diffusing molecules. This phenomenon has previously been reported in the literature (Hameed et al. 2008).

The interaction effect of adsorbent dose and initial BTA concentration at a constant reaction time on BTA adsorption was also examined, and the results are given in Figure 6. By increasing the adsorbent dosage from 0.1 to 1 g/L, the adsorption efficiency increased from 35.5% to 94.4%. The most probable proof of our observation can be attributed to the increased effective specific surface area or exchangeable sites, followed by raising free active sites for adsorbate removal, which results in more interactions between BTA and adsorbent (Oliveira et al. 2008; Tan et al. 2015; Pourzamani et al. 2017; Dai et al. 2019).

For optimization of independent variables, the Derringer's desirability function method was utilized. In the mentioned method, the function scale operated between 0 and 1, in which 0 shows an entirely undesirable response and 1 represents a fully desired response (Haghighi et al. 2017).

The optimization results for BTA adsorption are given in Table 6. Accordingly, the optimum experimental conditions for removing BTA from aqueous solution were 35 min reaction time, 0.55 g/L adsorbent dose, and 50 mg/L initial BTA concentration. The predicted BTA adsorption efficiency was 92.6% at these optimal conditions, with an overall desirability value of 0.992. To further validate the model and to ensure that the model is representative of the actual system, six additional experiments were conducted under optimized conditions. Experimental responses were plotted versus the responses predicted by the model, and the results are illustrated in Figure 7. The predicted responses had correlations with the experimental ones with a high coefficient of determination of R² = 0.992 and R² = 0.974 for internal and external validation, respectively. It could be counted as undeniable proof of the suitability of the model.

The adsorption isotherms are always considered feasible tools for determining the removal mechanism of pollutants in all adsorption systems. In order to obtain the isotherm parameters of BTA adsorption onto the MMBC, in general, the non-linear approach was adopted. The graphs of non-linear isotherms were prepared using MATLAB^® 7.11.0 (R2015b). Five widely used isotherm models, namely Langmuir, Freundlich, Redlich–Peterson, Temkin, and Liu were applied. The results of non-linear isotherm studies of BTA adsorption at different temperatures (283 K, 298 K, and 313 K) are presented in Table 7. Accordingly, higher adsorption capacities of Langmuir isotherm (q_e) were observed at higher temperatures (q_m = 563.1 at 313 K, determination coefficient = 0.9799). Based on Table 7, at all studied temperatures, the Langmuir model, in comparison with the others, yielded the best fit with the experimental data of BTA adsorption with high coefficients of determination (R² > 0.97) using the non-linear approach. Such observation clearly indicates the homogeneous identity of BTA adsorption onto the monolayer surfaces of MMBC. The Langmuir isotherm is applied to explain single-solute systems, and it is assumed that there are particular homogenous sites in the adsorbent and no significant interaction exists between diverse types of adsorbed substances. On the other hand, the Freundlich isotherm is used to explain heterogeneous systems (Suganya 2019). The adsorption capacities of BTA removal using various adsorbents are reported in Table 8, which confirms the high capability of ZnCl₂-activated magnetic biochar in eliminating BTA from aqueous solutions.

Kinetics describes solute uptake rates and defines the residence time of the adsorbate at the solid-liquid interface. Additionally, valuable insights into the reaction pathways and the adsorption mechanisms can be achieved via kinetic studies. Kinetic experiments were performed using the MMBC dosage, solution pH, and the initial BTA concentration of 1.0 g L⁻¹, 7.0, and 100 mg L⁻¹, respectively, at 283 K, 298 K, and 313 K (Figure 8).

As shown in Figure 8, BTA uptake seems to have occurred in two steps. The first step involved swift uptake within the first 10 min of the contact followed by the subsequent removal of BTA, which continued for a relatively short period until adsorption equilibrium was obtained. The kinetic parameters acquired from non-linear fitting results are presented in Table 9. Based on the higher values of the adjusted determination coefficient (R²adj >0.99) and the lower values of SSE and the RMSE, it can be concluded that the kinetic models of the Avrami fractional-order kinetic model were the most appropriate for representing BTA adsorption onto MMBC.

The reusability of the adsorbents is an essential criterion for their commercial applications. Simple methanol washing was performed to recover MMBC for reuse in other adsorption cycles since BTA is readily soluble in methanol. As represented in Figure 9, the adsorption percentage of BTA by MMBC decreased non-significantly from 99.2 to 93.9% after the fifth cycle. This proves that the synthesized MMBC can be recycled and reused for at least five successive cycles with an adsorption efficiency of >90%. It was also observed that desorption efficiencies did not noticeably change by increasing desorption cycles. More than 93.5% of the adsorbed BTA could be desorbed and recovered from the MMBC surface in the presence of methanol in the fifth cycle. These results suggest that the MMBC has a good potential for regeneration and reusability. Therefore, it can serve as a cost-effective and robust adsorbent for BTA removal from aqueous solutions in industrial applications based on simple and easy renewal with solvent treatment.

In the present research, the MMBC composite was synthesized and characterized by various techniques for BTA adsorption from aqueous solutions. Therefore, CCD was applied to investigate and optimize BTA adsorption variables such as BTA initial concentration, pH value, reaction time, and composite dosage. Based on the obtained results, the following conclusions are reached:

• SEM and EDX analysis announced that ZnCl₂ successfully modified Fe₃O₄ particles uniformly distributed on BC and prepared composed.

• Among various chemicals, ZnCl₂ was the best modifier for modification of MBC, and ZnCl₂/MBC had the highest performance in BTA adsorption.

• The primary operational parameter evaluation indicated that the BTA removal rate increases with increasing MMBC dosage and reaction time, while it declines with increasing initial BTA concentration.

• According to Derringer's desirability function method, the optimum condition was found to be 35 min reaction time, 0.55 g/L adsorbent dose, and 50 mg/L initial BTA concentration.

• The kinetic studies illustrated that the Avrami fractional-order model could describe BTA adsorption behavior well. Also, the adsorption isotherm was fitted with a non-linear Langmuir model with a maximum adsorption capacity of 563.1 mg/g. Moreover, thermodynamic parameters indicated a feasible, spontaneous, and endothermic physisorption.

• The regeneration assessments approved that the MMBC composite had appealing features for field application.

The authors would like to thank the Research and Technology Deputy of Ahvaz Jundishapur University of Medical Sciences for financial support (grant no: ETRC-9505).

All relevant data are included in the paper or its Supplementary Information.