One-pot fabrication of zero-valent iron-embedded activated carbon from rosemary distillation residues for malachite green removal

The current research proposes an innovative strategy for the facile preparation of magnetic activated carbon (MAC) from rosemary distillation residues (RDR). As a magnetic precursor, FeCl3 was impregnated into RDR before KOH was added as an activating agent. One-pot pyrolysis was then conducted to produce zero-valent iron nanoparticles (14.4 wt%) embedded in the activated carbon matrix. Moreover, KOH activation yielded MAC with a large total pore volume of 0.27 cm3 g−1, a high specific surface area of 459 m2 g−1, and hierarchical porosity. With a large porous system and different polar functional groups, MAC was subsequently investigated for malachite green (MG) removal in aqueous media. At pH 6.0, the adsorption process was consistent with the pseudo-second-order kinetic model and the Langmuir isotherm, with a maximum adsorption capacity of 82.6 mg g−1. Additionally, MAC demonstrated effective reusability after five consecutive cycles, when MG removal slightly decreased from 96.4 ± 0.6 to 91.8 ± 2.3%. Notably, MAC with a strong saturation magnetization of 18.4 emu g−1 could be conveniently recovered from treated media through magnetic fields. Overall, rosemary distillation residue-derived magnetic activated carbon can be a potential adsorbent for malachite green remediation thanks to its cost-effectiveness, eco-friendliness, and magnetic separability.


Introduction
Throughout the preceding decades, the global economy has experienced extraordinary growth. Industrialization, agricultural expansion, and urbanization have caused massive contamination of the human living environment [1,2]. Among these, water pollution is a pervasive issue. Wastewater contains a variety of harmful substances, such as pathogens, excess nutrients, heavy metals, and organic pollutants [3,4]. Hazardous dyes, the largest category of organic contaminants, are one of the major contributors to the rise in environmental pollution [5,6]. Their voluminous quantities are released from textiles, leather, printing, papermaking, food processing, pharmaceuticals, and cosmetics [7][8][9]. These untreated toxic substances create chemical and biological changes, endangering aquatic species. The dye molecules deplete the dissolved oxygen concentration, preventing sunlight from reaching water sources [10,11]. In addition, synthetic dyes are neither biodegradable nor resistant to environmental conditions [12,13]. Other than that, the bioaccumulation of dyes in aquatic environments might bring these harmful compounds into the food chain [14,15]. Hence, they can pose significant risks to human health, such as skin irritation, respiratory problems, cancer risks, reproductive health effects, and neurological effects [16,17]. As a result, the adverse effects of synthetic dyes in wastewater highlight the importance of their proper treatment and management. essential oils [53]. The oils are highly concentrated extracts that are obtained through steam distillation or hydrodistillation [54,55]. However, this production releases a large amount of rosemary distillation residues (RDR), which can include stems, leaves, and other plant parts [56]. Instead of discarding RDR and potentially polluting the surrounding environment [57], further research and development could help unlock its potential. Although numerous biomass resources have been explored for the preparation of MBC or MAC, RDR is rarely mentioned in published literature. Hence, RDR valorization is necessary, leading to a more sustainable and ecofriendly approach to essential oil production. Herein, MAC was fabricated via one-pot pyrolysis and activation of RDR. Fe 0 particles were formed easily in the carbon framework, and the potential use of MAC was explored for the adsorption of malachite green.

Materials
Rosemary distillation residues were taken from a factory for the extraction of rosemary essential oil by steam distillation located in Ho Chi Minh City, Vietnam. RDR was first cleaned with distilled water and oven-dried at 105°C for 18 h. Following this, the biomass underwent a milling process to achieve a finely powdered form. To protect the powder from moisture, it was placed in a hermetic container for use afterwards. FeCl 3 .6H 2 O ( 99.0%), KOH ( 85.0%), and KH 2 PO 4 ( 99.5%) were purchased from Guangdong Guanghua Sci-Tech Co., Ltd. Malachite green ( 94.5%), HCl (36%-38%), NaOH ( 96.0%), Na 2 HPO 4 .12H 2 O ( 99.0%), and absolute ethanol ( 99.7%) were obtained from Xilong Scientific Co., Ltd. All chemicals were used in their original form without any additional purification.

Preparation of magnetic activated carbon from rosemary distillation residues
Magnetic activated carbon was prepared from rosemary distillation residues using KOH as an activation agent and FeCl 3 as a magnetic precursor. Initially, 10.00 g of RDR powder and 2.00 g of FeCl 3 were blended into 100 ml of distilled water, which was then agitated continuously for 2.0 h. Next, 6.00 g of KOH was added to the mixture. After another 2.0 h of stirring, the mixture was desiccated at 100°C for 24 h. The solid composite was reground after drying. In a vertical furnace, 4.00 g of the mixture was put into a reaction tube for pyrolysis. Its inert atmosphere was maintained by a continuous flow of 0.25 l min −1 of nitrogen. The ambient temperature of the tube was increased to 600°C at an average rate of 5°C min −1 . The final temperature was then kept for 60 min. The resulting material was repeatedly rinsed to eliminate all water-soluble substances and dried at 80°C for 12 h to obtain MAC. Comparatively, biochar (BC) and magnetic biochar (MBC), which were served as reference samples, were produced by pyrolyzing RDR and FeCl 3 -loaded RDR without the addition of KOH under the same conditions.

2.3.
Characterization of magnetic activated carbon X-ray diffraction (XRD) of MBC and MAC was studied on a Bruker D2 diffractometer in the 2θ = 10°-80°range using CuKα radiation (λ = 1.5418 Å). Fe contents in MAC and MBC were analyzed by the ferrozine method [58]. The metal was extracted from MBC and MAC using HCl (6 M) at 60°C for 60 min. Porous properties of BC, MBC, and MAC were explored by nitrogen adsorption and desorption isotherms using a Micromeritics ® TriStar II Plus at 77 K. These samples were outgassed at 250°C for 6 h. Total pore volume (V total ) was determined at P/P o = 0.992. Specific surface area (S BET ) was computed from the Brunauer-Emmett-Teller (BET) equation. Micropore volume (V micro ) and external surface area (S ext ) were calculated by the t-plot method. Mesopore volume (V meso ) was obtained by subtracting V micro from V total . Micropore surface area (S micro ) was determined from the difference between S BET and S ext . Average pore size (d average ) was given by 4V total /S BET . Pore size distribution was calculated by the Barrett-Joyner-Halenda (BJH) method. Fourier transform infrared (FTIR) spectra of BC, MBC, and MAC were obtained using a PerkinElmer Spectrum spectrometer. The analysis of scanning electron microscope (SEM) images, energy dispersive x-ray (EDX) spectra, and elemental mappings was conducted using a FE-SEM Hitachi S-4800 system. Transmission electron microscopy (TEM) images were acquired using a JEOL JEM-1010 instrument. The magnetic properties of MAC and MBC were investigated with a vibrating sample magnetometer (VSM) at ambient temperature (30°C).

Removal of malachite green using magnetic activated carbon
The as-prepared BC, MBC, and MAC samples were investigated for the removal of malachite green at room temperature. In each 200-mL glass flask, 2.00 g l −1 of an adsorbent and a certain concentration of MG were mixed to make a total volume of 100 ml. The initial pH of the medium was adjusted with HCl (0.1 M) and NaOH (0.1 M) solutions. The mixture was shaken continuously during the adsorption process. At appropriate time intervals, 1.00 ml of sample was taken, and the existing adsorbent was immediately separated using a magnet.
The solution was subsequently poured into a phosphate buffer solution (pH 7.0). MG concentrations were quantitatively measured using a Lovibond PC Spectro UV-Vis spectrophotometer. The adsorption capacity (q t ) and MG removal, therefore, were calculated using the following equations: is the adsorbent dosage, C 0 and C t (mg/l) are MG concentrations at the beginning and after t (min) of adsorption, respectively.
In order to evaluate the reusability of MAC for MG removal, consecutive experiments were conducted with the same sample in duplicate. The used adsorbent was recovered by a magnet, cleaned with distilled water and ethanol, and dried in an oven at 105°C. The recovered material was weighed for use in the subsequent experiment. Adsorption capacity and MG removal were calculated at 120-min adsorption.  [51].  3 . Subsequently, sequential reductions in the pyrolysis step were accelerated to produce Fe 0 rather than the intermediate Fe 3 O 4 . As opposed to the previous case of MBC, these reactions were not H 2 O-limited.    Figure 2(a) depicts the nitrogen adsorption & desorption isotherms of BC, MBC, and MAC samples. A similar trend is observed for the three samples. As P/P o increased from 0.0004 to 0.01, the adsorbed volume increased significantly. All samples contained micropores, according to these results. Next, the adsorbed volume rose gradually in the P/P o range of 0.01-0.99. These results reveal that the samples contained different pore sizes. Notably, hysteresis loops at P/P o from around 0.40 to 0.99 might belong to type IV, according to the IUPAC classification [60]. Additionally, the shape of the hysteresis loops might be associated with a particular porous structure. These ones might be of the H1 type, which possibly relates to capillary condensation in slit-shaped pores [61]. The BJH pore size distribution, in fact, shows that all samples had a hierarchical microporous and mesoporous structure ( figure 2(b)).

Porous properties of BC, MBC and MAC
RDR-derived BC had a large porous system with a S BET of 249 m 2 g −1 and a V total of 0.20 cm 3 g −1 (table 1). Interestingly, the BJH pore size distribution of BC shows that this material possessed both micropores (V micro of 0.05 cm 3 g −1 ) and mesopores (V meso of 0.15 cm 3 g −1 ) with a typical pore diameter of 2.0 nm. Unlike other natural biomass resources, this RDR was the waste from the steam distillation of rosemary. Hot steam with high dynamics might severely destroy the cell walls of rosemary to extract essential oil [62]. Therefore, it appears that the porous structure of RDR was expanded before further use. This distinctive feature might be advantageous for the porous carbon systems of RDR-derived carbon materials.
As compared with BC, MBC had similar S BET (277 m 2 g −1 ) and V total (0.20 cm 3 g −1 ). With the low amount of FeCl 3 loading, the activation was still limited. However, this loading resulted in a higher V micro (0.08 cm 3 g −1 ) and a lower V meso (0.12 cm 3 g −1 ). In fact, BJH pore size distribution revealed that micropores were expanded, whereas mesopores were reduced. It seems that FeCl 3 activation mainly yielded micropores. Conversely, the presence of Fe 3 O 4 products may affect mesopores. Due to the limitation of FeCl 3 activation, additional KOH activation is needed to expand the porous carbon system.
Despite higher Fe loading, S BET (459 m 2 g −1 ) and V total (0.27 cm 3 g −1 ) of MAC were higher than those of MBC and BC. Furthermore, BJH pore size distribution shows that new micropores and mesopores were developed in MAC. A typical pore diameter of 1.4 nm was observed. These results show that the addition of KOH activated the RDR-derived porous carbon structure. When KOH was introduced to RDR containing FeCl 3 , Fe(OH) 3 and KCl were produced. As such, the mixture prior to pyrolysis consisted of RDR, Fe(OH) 3 , extra KOH, and KCl. KOH could serve as the most important component in the activation of porous carbon. Furthermore, reductions (6) indicate that Fe(OH) 3 could contribute to the improvement of the porous carbon system. Gong et al [63] prove that iron oxides can accelerate the development of micropores inside a carbon framework. Lastly, the involvement of KCl in the activation process should be recognized. According to Gómez et al [64], the introduction of KCl can result in increased surface area and pore volume. Nevertheless, KCl does not play an activating role. Rather, it facilitates the interaction of KOH with porous carbon, thereby increasing the activation efficiency. The presence of KCl may lower the amount of KOH required for the activation process  Prior studies showed that FeCl 3 can act as a catalyst for faster biomass carbonization [38,43]. In addition, oxygen-containing functional groups appear to be augmented, which may strengthen the polarity of the MBC surface. Notable discovery was the vibration peak of the Fe-O bonds at 574 and 431 cm −1 [68][69][70], which was ascribed to the presence of Fe 3 O 4 in MBC.
Regarding MAC, different absorption bands were recorded, including 3158 cm −1 (O-H stretching vibration), 1584 cm −1 (C=C stretching vibration in the aromatic rings), 1203 cm −1 (C-O stretching vibration), and 607 cm −1 (O-H bending vibration) [71,72]. Compared with MBC, MAC had similar oxygen-containing functional groups but lacked Fe-O bonds. On the other hand, the C-H groups in MAC were weak, indicating that KOH activation could result in a more robust RDR carbonization than FeCl 3 activation alone.

SEM images of MAC
The surface morphology of MAC is described in figure 4. Generally, the material had a smooth surface, and cavities were found. Depending on the porous properties, the activation process might strongly affect micropores and mesopores rather than macropores. Therefore, those micrometer-sized pores might come from the natural structure of RDR. In addition, aggregated Fe-based particles were not identified clearly on the MAC surface. Several reports indicated that nanoscale Fe-based particles generated by FeCl 3 activation of biomass could be kept inside the carbon matrix instead of on its surface [39,73]. To clarify these diffusions, TEM images of MAC were presented in the later part.

EDX spectra and elemental mappings of MBC and MAC
EDX analysis was used to explore the elemental types and amounts on the surfaces of MBC and MAC (figure 5). According to the data, MBC included 80.69 wt% C, 4.31 wt% Fe, and 13.04 wt% O. Remarkably, the surface Fe content as determined by the EDX method is comparable to the bulk Fe content (6.1 wt%) as quantified by the ferrozine method. According to Do et al [39], the identical contents of Fe on the surface and in the bulk imply that this element was distributed throughout the carbon framework. Intriguingly, the high oxygen content found by EDX analysis corroborated that this element was present not only in Fe 3 O 4 but also in the oxygen functional groups described by FTIR spectroscopy. Regarding trace elements, 0.26 wt% S and 1.70 wt% Cl were detected. The source of these elements might come from RDR or chemical impurities. For Cl, numerous washings with distilled water painstakingly eliminated any remaining FeCl 3 . Ergo, any residual Cl in MBC may be bound firmly in place by strong mechanical or chemical interactions. In fact, Xu et al [38] demonstrated that the activation of biomass with FeCl 3 can result in the formation of strong C-Cl linkages.
Unlike the activation by FeCl 3 alone, the addition of KOH contributed to a significant change in the elemental composition appearing on the MAC surface ( figure 5(b)). The results revealed that the main elements were C, Fe, and O at 65.38, 15.27, and 16.79 wt%, respectively. Compared with MBC, MAC had a higher Fe content but a lower C content. The presence of KOH, as previously stated, not only loaded all Fe(OH) 3 in RDR, but it also aided in the etching of the porous carbon during the activation process. Despite the higher Fe surface content, this value is close to the bulk Fe content (14.4 wt%). Similar to MBC, this comparison revealed that Fe 0 particles were well embedded in the carbon framework. A high oxygen content in the MAC could also result from functional groups. MAC did, however, contain more trace elements (< 1.00 wt%) of Mg, Al, Si, P, S, K, and Ca. Perhaps the addition of KOH might play a certain role in the immobilization of these elements in the carbon framework. Moreover, the impurities in chemicals might introduce trace elements to MAC. Regarding K, this element might come from KOH. In addition, elemental mapping demonstrates that all available elements were distributed consistently throughout the surfaces of both MBC and MAC samples. This may be the benefit of the molecularly well-impregnated iron precursor over the naturally porous system of RDR feedstock. Overall, trapping magnetic Fe-based particles inside the carbon frameworks of MBC and MAC materials is expected to improve their long-term stability.

TEM images of BC, MBC and MAC
The inner structures of BC, MBC, and MAC were examined by TEM images, as illustrated in figure 6. At the nanoscale, BC had a naked surface with relatively uniform brightness over a vast area. After pyrolysis, covalent bonds were formed between the carbon framework and the functional groups. This made the BC structure fairly uniform. In contrast, the erratic and darker regions in MBC revealed the morphology of Fe 3 O 4 . It appears that the carbon framework contained dust-like Fe 3 O 4 nanoparticles. During the impregnating process, the iron precursor could be introduced into the narrow pores of RDR. As a result, the Fe-based particles produced were confined within the carbon framework. Such studies described the distribution of Fe 3 O 4 nanoparticles in MBC [74][75][76]. In a similar manner, Fe 0 nanoparticles resembling dust were well spread over the carbon framework of MAC. In addition, because the carbon framework was composed of covalent bonds, ionic Fe 3 O 4 or metallic Fe 0 particles might be entrapped via mechanical linkages as opposed to chemical bonds. More importantly, it is expected that the rigid immobilization of magnetic nanoparticles could improve the stability of MBC and MAC composites. Figure 7 demonstrates that the MAC sample was efficiently attracted by a magnet, indicating that it is a strong magnetic material. To investigate the magnetic characteristics of MBC and MAC in more detail, VSM was used. As a result, similar curves are observed for both materials. Magnetic hysteresis loops with extremely low coercivity exhibited superparamagnetic behavior. Hence, these soft magnetic materials were easily magnetizable and demagnetizeable. The saturation magnetization (M S ) of MAC with Fe 0 nanoparticles (18.4 emu g −1 ) was 6.3 times greater than that of MBC with Fe 3 O 4 nanoparticles (2.9 emu g −1 ). However, the Fe content of MAC (14.4 wt%) was only 2.4 times greater than that of MBC (6.1 wt%). Therefore, the magnetic properties of various Fe-based materials may be the main explanation for this disparity. Feng et al [77] reported that when Fe 3 O 4 was reduced to Fe 0 , the magnetic properties of the resultant material increased. Consequently, Fe 0 can provide stronger magnetic properties than Fe 3 O 4 . Compared with MBC, MAC possessed not only a higher Fe content but also stronger magnetic particles. Equally significant, other features, like the size, shape, magnetic anisotropy, and crystallinity of Fe 0 and Fe 3 O 4 particles, which are highly dependent on synthesis conditions, could affect their magnetic properties [78,79]. Because the magnetic properties of Fe-based crystals are size-dependent, their dust-like nanoparticles in MBC and MAC could play a certain role. Altogether, the addition of KOH resulted in a significant enhancement in the magnetic properties of MAC, which is beneficial for magnetic separation.  groups may be beneficial for MG adsorption on the MBC surface. Notably, the adsorption capacity of MAC was much higher than that of BC and MBC. The advantages of MAC could come from its expanded porous system and oxygen-containing functional groups. Thus, additional KOH activation contributed significantly to the enhancement of MG removal.

Effect of pH on adsorption of MG by MAC
pH can influence the ionization of adsorbates, the surface charge of adsorbents, and the adsorption processes all together [80]. Therefore, different pH values from 2.0 to 10.0 were investigated for the adsorption of MG by MAC ( figure 9). By increasing pH from 2.0 to 6.0, MG removal and adsorption capacity at 240 min remarkably rose from 43.6 to 95.1% and from 10.9 to 23.8 mg g −1 , respectively. MG is a water-soluble cationic dye with a pKa of 6.9 [81]. In acidic environments, MG becomes cationic. In addition, such studies reported that the  surface charge of biomass-derived porous carbon with oxygen-containing functional groups could be altered by pH [82][83][84]. As pH decreased, more positive charges on MAC were formed. Consequently, the decrease in MG adsorption at low pH levels was due mostly to the increasing repulsion between positively charged MG species and the positively charged surface of MAC. Additionally, H + ions might compete with MG for the adsorption sites [85]. At pH 6.0, the MAC surface might become negative because of the deprotonation of oxygencontaining groups. Then, it might be favorable for MG adsorption. However, when the pH was higher than 6.0, MG removal and adsorption capacity slightly decreased. In basic environments, MG might become less cationic and form dye dimers, restricting MG adsorption [86]. Thus, MAC removed MG at a broad pH range, but the appropriate one was 6.0.

Effect of MG concentration on the adsorption capacity of MAC
It is important to explore the adsorption capacity of MAC at different initial MG concentrations. It was obvious that when the initial MG concentration increased, MG adsorption capacity improved gradually (figure 10). Depending on the adsorption equilibrium, higher concentrations of dye could promote an equilibrium shift in the adsorption direction. Moreover, high MG concentrations could be a key driving factor for faster mass transfer from the bulk environment to the MAC surface. In general, the adsorption processes were nearly complete within 120 min. These results were subsequently discussed in later parts of the thermodynamic and kinetic analyses.

Kinetics of MG adsorption onto MAC
The kinetics of MG adsorption onto MAC were studied with the pseudo-first-order, pseudo-second-order, and intraparticle diffusion models. Equations, parameters, and calculated results are presented in table 2. The plots of the linearized forms of the three models are shown in figure 11. q e and q t are the equilibrium adsorption capacity and the adsorption capacity at time t, respectively. k 1 , k 2 , and K p are the rate constants of the pseudofirst-order, pseudo-second-order, and intraparticle diffusion models, respectively. As a result, the R 2 values of the pseudo-second-order model, which were found to be close to 1, were much better than those of the pseudofirst-order and intraparticle diffusion models. In addition, the calculated MG adsorption capacities from the pseudo-second-order kinetic model at equilibrium were all close to the experimental ones. Hence, that model could be more appropriate to describe the adsorption behavior of MG onto MAC. The pseudo-second-order kinetic model for MG adsorption was fitted for zero-valent iron/biochar composites [82,86], SrFe 12 O 19 /biochar [85], and other adsorbents [81].

Adsorption isotherms of MG onto MAC and adsorption mechanism
The Langmuir, Freundlich, and Temkin isotherms were used to study MG adsorption onto MAC. Equations, parameters, and calculated results are presented in table 3. The plots of the linearized forms of the three isotherms are presented in figure 12. The equilibrium concentration (C e ) and the equilibrium adsorption capacity (q e ) were determined after 240 min of adsorption. q m is the maximum adsorption capacity, and k L is the Langmuir constant. K F and n are the Freundlich constants. Lastly, K T is the equilibrium bond constant, and RT/ b is the Temkin constant.     As shown in figure 12, the Langmuir isotherm was best fitted with experimental adsorption data. The R 2 value of the Langmuir model (0.998) was better than that of the Freundlich model (0.935) and Temkin (0.987). The MG adsorption onto MAC appears as a monolayer, according to the Langmuir model. The Temkin model was also relatively suitable for the experimental data. Both the Langmuir and Temkin models are considered for chemisorption. The negatively charged surface of MAC may well attract positively charged MG species at pH 6.0, resulting in monolayer adsorption. As reviewed by Raval et al [81], the Langmuir isotherm is frequently used for MG adsorption on biomass-derived porous carbon. Furthermore, the maximum adsorption capacity of MAC on MG was 82.6 mg g −1 , which could be comparable with that of other adsorbents. More importantly, the introduction of magnetic properties to MAC was a great benefit for convenient separation.
Malachite green is a cationic dye with the molecular formula [C 6 H 5 C(C 6 H 4 N(CH 3 ) 2 ) 2 ]Cl. It is constructed of three phenyl rings and two amine groups. The positive charges of MG might interact well with the negative charges of the MAC surface through electrostatic interactions [87]. The phenyl rings have the potential to form π-π interactions with the graphitic planes of MAC [88]. The amine groups are capable of forming hydrogen bonds with hydroxyl and carboxyl groups on the MAC surface [89]. Moreover, the hydrophobicity of both MG and MAC might encourage their hydrophobic interactions in aqueous media. Lastly, the porous structure of MAC might allow MG molecules to diffuse into the pores and be trapped inside, leading to an additional adsorption mechanism known as pore filling [90]. As a summary, the plausible mechanism of MG adsorption on MAC might include electrostatic interactions, π-π stackings, hydrogen bonding, hydrophobic interactions, and pore-filling effects.

Reusability of MAC adsorbent
The reusability of adsorbents is a crucial characteristic for determining their commercial viability for industrial applications. Five cycles with the same adsorbent were performed to explore the reusability of MAC for MG removal. The utilized MAC was collected using a magnet, cleaned with water and ethanol, and dried out. These investigations were performed in duplicate. As depicted in figure 13, the adsorption performance of MAC remained reliable after five cycles. MG removal marginally decreased from 96.4 ± 0.6% in the first cycle to 91.8 ± 2.3% in the fifth. Consequently, the adsorption capacity decreased from 24.1 ± 0.2 to 22.9 ± 0.6 mg g −1 . It appears that water and ethanol are unable to completely remove MG from the surface of MAC. In comparison to activated carbon, the magnetic separability of MAC significantly facilitates its recovery and reuse. With regard to clean technology and green chemistry, the high adsorption efficiency, low cost, and exceptional reusability of MAC make it a promising self-regenerating adsorbent for MG treatment.

Conclusion
In this study, magnetic activated carbon was successfully prepared via a one-pot strategy using rosemary distillation residues as a carbon source, FeCl 3 as a magnetic precursor, and KOH as an activating agent. The results showed that zero-valent iron nanoparticles were firmly encapsulated within the porous carbon matrix. MAC had a high specific surface area of 459 m 2 g −1 , a large total pore volume of 0.27 cm 3 g −1 , hierarchical porosity, and a powerful saturation magnetization of 18.4 emu g −1 . MAC was subsequently explored for malachite green removal. As a result, the adsorption process was well fitted with pseudo-second-order kinetics and Langmuir isotherms, with a maximum adsorption capacity of 82.6 mg g −1 . Moreover, the high reusability of MAC was proven after five consecutive cycles, when MG removal slightly declined from 96.4 ± 0.6 to 91.8 ± 2.3%. In conclusion, magnetic activated carbon derived from rosemary distillation residues can serve as a potential adsorbent for malachite green remediation due to its low cost, eco-friendly nature, and magnetic separability.