The Development of Fe3O4-Monolithic Resorcinol-Formaldehyde Carbon Xerogels Using Ultrasonic-Assisted Synthesis for Arsenic Removal of Drinking Water

Inorganic arsenic in drinking water from groundwater sources is one of the potential causes of arsenic-contaminated environments, and it is highly toxic to human health even at low concentrations. The purpose of this study was to develop a magnetic adsorbent capable of removing arsenic from water. Fe3O4-monolithic resorcinol-formaldehyde carbon xerogels are a type of porous material that forms when resorcinol and formaldehyde (RF) react to form a polymer network, which is then cross-linked with magnetite. Sonication-assisted direct and indirect methods were investigated for loading Fe3O4 and achieving optimal mixing and dispersion of Fe3O4 in the RF solution. Variations of the molar ratios of the catalyst (R/C = 50, 100, 150, and 200), water (R/W = 0.04 and 0.05), and Fe3O4 (M/R = 0.01, 0.03, 0.05, 0.1, 0.15, and 0.2), and thermal treatment were applied to evaluate their textural properties and adsorption capacities. Magnetic carbon xerogel monoliths (MXRF600) using indirect sonication were pyrolyzed at 600 °C for 6 h with a nitrogen gas flow in the tube furnace. Nanoporous carbon xerogels with a high surface area (292 m2/g) and magnetic properties were obtained. The maximum monolayer adsorption capacity of As(III) and As(V) was 694.3 µg/g and 1720.3 µg/g, respectively. The incorporation of magnetite in the xerogel structure was physical, without participation in the polycondensation reaction, as confirmed by XRD, FTIR, and SEM analysis. Therefore, Fe3O4-monolithic resorcinol-formaldehyde carbon xerogels were developed as a potential adsorbent for the effective removal of arsenic with low and high ranges of As(III) and As(V) concentrations from groundwater.


Introduction
A current environmental and human health problem is the availability of water due to the increasing demand and contamination of drinking water sources. Most of the accessible drinking water is found in aquifers, which are underground reservoirs of water. However, the presence of contaminants such as arsenic, which naturally occur in the environment due to geological factors, can migrate into groundwater through weathering processes.
Inorganic arsenic (As) is a well-known carcinogenic element and one of the most significant chemical pollutants worldwide, found in several countries across the globe. The Agency for Toxic Substances and Disease Registry (ATSDR) has ranked arsenic as the top substance with potential risks to public health on a global scale [1]. The World Health Gels 2023, 9, 618 4 of 29 This research focuses on generating magnetic carbon xerogels with morphological, magnetic, textural, and physical-chemical properties in the form of pellets, which makes them reusable. These materials have the capability to adsorb arsenates and arsenites. The synthesis procedure was developed considering the effect of Fe 3 O 4 loading via ultrasonic methods, both direct and indirect, while varying the molar ratios of Fe, catalyst, and water. The arsenic adsorption test was conducted in batch, and the synthesized materials were characterized using various analytical techniques before and after the adsorption of arsenic. The intended purpose of this work is for the adsorbent materials produced to serve as viable alternatives within the technological advancements for water remediation. The magnetic properties of carbon xerogels facilitate the separation, reuse, regeneration, and recycling of the adsorbents so that their useful life is extended. The efficient separation of aged adsorbents also facilitates the recovery and final disposal of contaminants and strengthens the environmental sustainability of the water purification process.
The experimental reproducibility of Fe 3 O 4 -monolithic resorcinol-formaldehyde carbon xerogels involves several challenges, including the composition of the starting materials (variation in molar ratios), the homogeneity of dispersion (direct and indirect ultrasonication methods), the sonication conditions (power output, duration, and frequency), the gelation and curing conditions, and the post-synthesis treatments (pyrolysis). Controlling these parameters consistently across different experiments can be challenging, and this can affect the properties of the resulting monolithic carbon xerogels. Subsequently, the adsorption capacity of the final material was improved with increasing the initial concentration and the adsorption affinity for arsenic species.
The environmental sustainability of the arsenic adsorption process using Fe 3 O 4monolithic resorcinol-formaldehyde carbon xerogels can be evaluated through methodologies such as the lifecycle, planetary boundaries, and sustainable development goals [38]. However, applying these methodologies and their indicators is beyond the scope of this article.

Fe 3 O 4 -Monolithic Resorcinol-Formaldehyde Xerogels: Effect of Loading of Magnetite with Indirect and Direct Sonication, and Modification of Catalyst
This study focuses on the development and initial preparation of monolithic xerogels with magnetic properties using magnetite (Fe 3 O 4 , Lanxess) as an adsorbent material for water treatment. The magnetic xerogel monoliths (MCs and MXs) were synthesized through the sol-gel polymerization of resorcinol and formaldehyde (RF) with sodium carbonate (C) as a catalyst, employing indirect and direct sonication, respectively. The effect of varying the molar ratios of resorcinol/catalyst was evaluated to obtain the high adsorption capacity in the arsenate adsorption.

Characterization of MCs and MXs
To identify the phases in the xerogels, XRD analysis was carried out. XRD pattern of RFX revealed the presence of both crystalline and amorphous phases as shown in Figure 1, which is similar to the pattern reported by [39,40].
The XRD patterns of magnetic xerogels prepared using direct and indirect sonication methods and different R/C ratios ( Figure 1) were found to be similar, with diffraction peaks at 2 θ values of 18 • , 30 • , 35.5 • , 43 • , 57 • , and 62 • corresponding to the crystallographic planes of magnetite 111, 220, 311, 400, 511, and 440, as reported in the ICCD card number 00-01900629. These findings align with the research of [41]. The percentage of crystalline phase for RFX, MX1, and MX2 was 10.54%, 12.45%, and 10.51%, respectively. Meanwhile, the percentage of crystalline phase for MC1-MC4 ranged from 10.51% to 12.64%. The percentage of crystalline phase for magnetic xerogels prepared using direct and indirect methods (MX1 and MC1) at the same molar ratios and gelation process was approximately 12%. card number 00-01900629. These findings align with the research of [41]. The percentage of crystalline phase for RFX, MX1, and MX2 was 10.54%, 12.45%, and 10.51%, respectively. Meanwhile, the percentage of crystalline phase for MC1-MC4 ranged from 10.51% to 12.64%. The percentage of crystalline phase for magnetic xerogels prepared using direct and indirect methods (MX1 and MC1) at the same molar ratios and gelation process was approximately 12%. The crystal size of the magnetic xerogels was determined by calculating the X-ray diffraction peak widths using Bragg's law and Debye Scherrer equation, as described by [42].
where D is the crystalline size, K denotes represents the Scherrer constant (0.98), λ represents the X-rays wavelength (1.54178 Å), β denotes the full width at half maximum (FWHM) and θ is the Bragg diffraction angle (radians). Table 1 shows the average crystalline sizes of MC50, MC100, and MC200 were in the range of 22.94-25.88 nm. Additionally, their values of β (0.32-0.36 radians) and θ (35.52-35.53 radians) were similar. However, the crystallinity of particle of MC200 (R/C = 200) was higher than MC50 (R/C = 50) and MC100 (R/C = 100). This indicates that increasing the R/C ratio can result in increased crystallinity, which is similar to the results obtained by [43].  The crystal size of the magnetic xerogels was determined by calculating the X-ray diffraction peak widths using Bragg's law and Debye Scherrer equation, as described by [42].
where D is the crystalline size, K denotes represents the Scherrer constant (0.98), λ represents the X-rays wavelength (1.54178 Å), β denotes the full width at half maximum (FWHM) and θ is the Bragg diffraction angle (radians). Table 1 shows the average crystalline sizes of MC50, MC100, and MC200 were in the range of 22.94-25.88 nm. Additionally, their values of β (0.32-0.36 radians) and θ (35.52-35.53 radians) were similar. However, the crystallinity of particle of MC200 (R/C = 200) was higher than MC50 (R/C = 50) and MC100 (R/C = 100). This indicates that increasing the R/C ratio can result in increased crystallinity, which is similar to the results obtained by [43]. The morphology of RF xerogels (RFX) was observed using scanning electron microscopy (SEM). Figure 2a shows that RFX is composed of a large number of microclusters that are uniformly distributed. These microclusters contain the resorcinol-formaldehyde polymer. The RF gel network is formed with nearly spherical particles, showing similar results to those obtained by [44]. Furthermore, the interconnects between the microclusters were observed to form porous materials. This porosity is likely due to the gelation process used in the synthesis of RFX, which involves the formation of a three-dimensional network of interconnected polymer chains [39,45,46]. Therefore, RFX is a highly porous material with a complex microstructure. Gels 2023, 9, 618 6 of 29 microclusters in the outer region appear more compact than those in the inner region. This difference in microstructure is likely due to the contact of the outer region with the glass tube during the gelation process. During gelation, the RF solution is typically poured into a mold or tube and allowed to solidify. The contact of the outer region with the glass surface may have caused the microclusters to pack more tightly together, resulting in a more compact microstructure. In this study, the SEM analysis revealed the effect of direct and indirect ultrasonication on the preparation of magnetic xerogels. Figure 3 depicts the morphology of magnetic gels, namely MX1 and MC4, synthesized with the same molar ratios. It can be observed that the morphology of MC4 (Figure 3b,d) characterizes nearly spherical particles that partially overlap, resulting in the formation of large pores. This morphology is likely attributed to the incorporation of magnetite particles into the RF gel during synthesis. Comparing MX1 and MC4 at the same magnification range of 15,000 and 25,000, it is evident that MX1 has smaller particle and pore sizes compared to MC4. Additionally, MX1 shows a more compact RF gel structure than MC4. Both techniques involve delivering energy to the RF solution with magnetite particles through probe sonication. However, the resulting particle sizes and mesoporosity of RF gels differ between the two methods. Indirect sonication generates cavitation in the water bath using high-intensity ultrasound through a water bath, while direct sonication involves the probe causing cavitation during sample processing. It can be explained that the particle size of MX1 decreases after ultrasonication, as observed by [37,47,48].  Figure 2a shows the difference between the outside and inside of the RF gel. The microclusters in the outer region appear more compact than those in the inner region. This difference in microstructure is likely due to the contact of the outer region with the glass tube during the gelation process. During gelation, the RF solution is typically poured into a mold or tube and allowed to solidify. The contact of the outer region with the glass surface may have caused the microclusters to pack more tightly together, resulting in a more compact microstructure.
In this study, the SEM analysis revealed the effect of direct and indirect ultrasonication on the preparation of magnetic xerogels. Figure 3 depicts the morphology of magnetic gels, namely MX1 and MC4, synthesized with the same molar ratios. It can be observed that the morphology of MC4 (Figure 3b,d) characterizes nearly spherical particles that partially overlap, resulting in the formation of large pores. This morphology is likely attributed to the incorporation of magnetite particles into the RF gel during synthesis. Comparing MX1 and MC4 at the same magnification range of 15,000 and 25,000, it is evident that MX1 has smaller particle and pore sizes compared to MC4. Additionally, MX1 shows a more compact RF gel structure than MC4. Both techniques involve delivering energy to the RF solution with magnetite particles through probe sonication. However, the resulting particle sizes and mesoporosity of RF gels differ between the two methods. Indirect sonication generates cavitation in the water bath using high-intensity ultrasound through a water bath, while direct sonication involves the probe causing cavitation during sample processing. It can be explained that the particle size of MX1 decreases after ultrasonication, as observed by [37,47,48].
Energy-dispersive X-ray spectroscopy (EDX) is a technique used to determine the elemental composition of a material. In this study, EDX analysis was used to determine the Fe content, confirming the incorporation of Fe content in the structure of magnetic RF gels. Both magnetic gels demonstrate the physical incorporation of magnetite into the structure of RF gel without participating in the polycondensation reaction of RF, as stated by [47]. MC4 shows a more uniform distribution of magnetite contents in the structure of RF gel than MX1.
During the synthesis of MX1 with direct sonication, the RF solution was mixed, and the temperature increased dramatically from 45 • C to 79 • C within 5 min, resulting in the formation of a black gel. On the other hand, MC4 was prepared using indirect sonication. The mixed solution of MC1-MC4 allowed the dispersion of magnetite into the RF gel, and the temperature of the solution continuously increased from 33 • C to 85 • C, and finally leading to the formation of a black gel within the water bath for 60 min. Energy-dispersive X-ray spectroscopy (EDX) is a technique used to determine elemental composition of a material. In this study, EDX analysis was used to determ the Fe content, confirming the incorporation of Fe content in the structure of magn RF gels. Both magnetic gels demonstrate the physical incorporation of magnetite into structure of RF gel without participating in the polycondensation reaction of RF, as s ed by [47]. MC4 shows a more uniform distribution of magnetite contents in the st ture of RF gel than MX1.
During the synthesis of MX1 with direct sonication, the RF solution was mixed, the temperature increased dramatically from 45 °C to 79 °C within 5 min, resulting in formation of a black gel. On the other hand, MC4 was prepared using indirect s cation. The mixed solution of MC1-MC4 allowed the dispersion of magnetite into the gel, and the temperature of the solution continuously increased from 33 °C to 85 °C, finally leading to the formation of a black gel within the water bath for 60 min.
It can be explained that direct sonication involves the use of a sonication probe rectly immersed in the reaction mixture. The probe emits ultrasonic waves that dire interact with the sample, resulting in more localized and precise energy transfer. H ever, direct sonication generally requires shorter processing times compared to indi sonication, as the energy efficiently reaches the desired regions, accelerating the requ reactions. Consequently, the mixed solution of MX1 with ultrasonication experience significant increase in temperature, leading to reduced gelation times [49]. In this stu magnetite was added to the RF solution, and due to its natural behavior, magne tends to agglomerate within a short mixing time. Considering the variables involve the solution, gelation, and curing processes, high temperatures during synthesis lea porosity shrinkage [49].
After ultrasonication of magnetite into an aqueous RF solution, the particle siz magnetite was decreased, which can be clearly observed with MC4. The EDX anal revealed that MC4 incorporated 1.19% Fe content (Figure 4d). The M/R ratio used in synthesis was 0.01, indicating a low concentration of magnetite compared to the RF ymer. This suggests that even with a low M/R ratio, the incorporation of Fe in the RF It can be explained that direct sonication involves the use of a sonication probe directly immersed in the reaction mixture. The probe emits ultrasonic waves that directly interact with the sample, resulting in more localized and precise energy transfer. However, direct sonication generally requires shorter processing times compared to indirect sonication, as the energy efficiently reaches the desired regions, accelerating the required reactions. Consequently, the mixed solution of MX1 with ultrasonication experienced a significant increase in temperature, leading to reduced gelation times [49]. In this study, magnetite was added to the RF solution, and due to its natural behavior, magnetite tends to agglomerate within a short mixing time. Considering the variables involved in the solution, gelation, and curing processes, high temperatures during synthesis lead to porosity shrinkage [49].
After ultrasonication of magnetite into an aqueous RF solution, the particle size of magnetite was decreased, which can be clearly observed with MC4. The EDX analysis revealed that MC4 incorporated 1.19% Fe content (Figure 4d). The M/R ratio used in the synthesis was 0.01, indicating a low concentration of magnetite compared to the RF polymer. This suggests that even with a low M/R ratio, the incorporation of Fe in the RF gel was successful, due to the use of magnetite particles in the synthesis process.
The morphology of the MCs was studied by SEM as shown in Figure 5. A threedimensional RF gel network was formed with nearly spherical particles [46,50]. MC1 and MC4 prepared with different Na 2 CO 3 concentrations, the morphology and pore size distribution can be observed that MC1 with lower R/C molar and high initial pH solution exhibit smaller particles and pore sizes than other materials. pH variations can alter the nucleation and growth of the gel network, leading to changes in the average pore size, pore connectivity, and surface area of the xerogel. Higher pH values can promote the formation of smaller pores, while lower pH values may result in larger pores [50].  The morphology of the MCs was studied by SEM as shown in Figure 5. A threedimensional RF gel network was formed with nearly spherical particles [46,50]. MC1  the nucleation and growth of the gel network, leading to changes in the average pore size, pore connectivity, and surface area of the xerogel. Higher pH values can promote the formation of smaller pores, while lower pH values may result in larger pores [50].
The mesoporosity of RF gels increases with an increase in the R/C ratio, as reported in previous studies by [39,45,51]. This indicates that the porosity of RF gels can be controlled by adjusting the R/C ratio in the synthesis process. Mesopores are pores with diameters between 2 and 50 nm and are desirable for various applications such as adsorption.
(a) (b)  Table 2 shows the textural properties of magnetic xerogels, the effect of direct, and indirect sonication on the textural properties of materials, specifically MX1 and MC4, respectively. The surface area of the magnetic xerogels for both MX1 and MC4 increased significantly compared to the xerogel. MC4 exhibited a higher surface area of 529.47 m 2 /g, whereas MX1 had a surface area of 472.41 m 2 /g. Additionally, the total pore volume and average pore diameter of MC4 were lower than those of MX1. This can be explained by the fact that MX1, prepared through direct sonication with a shorter sonication time for gelation, resulted in a lower surface area but higher total pore volume and larger average pore size.
The effect of pore structure in magnetic xerogels using the ultrasound-assisted solgel method was investigated through N2 physisorption analysis. It can be explained that direct sonication promotes the formation of smaller and more uniform pores within the xerogel structure, as the energy can be precisely targeted to specific regions. On the other hand, indirect sonication may result in the generation of larger or more irregularly shaped pores in the xerogel due to less controlled and localized energy transfer. This observation is consistent with the findings in Figure 3 of SEM images and Figure 4 depicting particle distributions, which demonstrate that MX1 has a smaller particle size compared to MC4.
Moreover, Table 2 shows the effect of catalyst contents on the surface area and pore volume of the magnetic xerogels. The results of the RF gels using sodium carbonate as a catalyst show that MC200 had a higher average pore diameter (5.16 nm) than MC100 (4.03 nm), but MC200 had a lower surface area (529.47 m 2 /g) than MC100 (545.09 m 2 /g). However, MC200 (529.47 m 2 /g) exhibited a surface area lower than MC100 (545.09 m 2 /g). These findings are consistent with those of [52], reported that increasing the molar ratios of R/C in gels prepared with Na2CO3 leads to an increase in average pore width. When lower molar ratios of R/C are used for RF gel preparation, a higher concentration of The mesoporosity of RF gels increases with an increase in the R/C ratio, as reported in previous studies by [39,45,51]. This indicates that the porosity of RF gels can be controlled by adjusting the R/C ratio in the synthesis process. Mesopores are pores with diameters between 2 and 50 nm and are desirable for various applications such as adsorption. Table 2 shows the textural properties of magnetic xerogels, the effect of direct, and indirect sonication on the textural properties of materials, specifically MX1 and MC4, respectively. The surface area of the magnetic xerogels for both MX1 and MC4 increased significantly compared to the xerogel. MC4 exhibited a higher surface area of 529.47 m 2 /g, whereas MX1 had a surface area of 472.41 m 2 /g. Additionally, the total pore volume and average pore diameter of MC4 were lower than those of MX1. This can be explained by the fact that MX1, prepared through direct sonication with a shorter sonication time for gelation, resulted in a lower surface area but higher total pore volume and larger average pore size. The effect of pore structure in magnetic xerogels using the ultrasound-assisted sol-gel method was investigated through N 2 physisorption analysis. It can be explained that direct sonication promotes the formation of smaller and more uniform pores within the xerogel structure, as the energy can be precisely targeted to specific regions. On the other hand, indirect sonication may result in the generation of larger or more irregularly shaped pores in the xerogel due to less controlled and localized energy transfer. This observation is consistent with the findings in Figure 3 of SEM images and Figure 4 depicting particle distributions, which demonstrate that MX1 has a smaller particle size compared to MC4.
Moreover, Table 2 shows the effect of catalyst contents on the surface area and pore volume of the magnetic xerogels. The results of the RF gels using sodium carbonate as a catalyst show that MC200 had a higher average pore diameter (5.16 nm) than MC100 (4.03 nm), but MC200 had a lower surface area (529.47 m 2 /g) than MC100 (545.09 m 2 /g). However, MC200 (529.47 m 2 /g) exhibited a surface area lower than MC100 (545.09 m 2 /g).
These findings are consistent with those of [52], reported that increasing the molar ratios of R/C in gels prepared with Na 2 CO 3 leads to an increase in average pore width. When lower molar ratios of R/C are used for RF gel preparation, a higher concentration of Na 2 CO 3 results in the formation of smaller clusters with smaller average-sized pores. Therefore, MC100, with its lower R/C ratios, has a greater number of smaller pore diameters, and a higher surface area, making it more suitable for use in water treatment adsorption.
In this study, the obtained results of the mesoporous nature of magnetic xerogels with varying R/C demonstrated the effect on the surface area and pore volume of the RF polymer in magnetic xerogels. It can be explained that the pH of the RF solution is associated with the quantity of catalyst utilized during the synthesis. When the pH was decreased, both the surface area and pore volume of the RF polymer in xerogels increased. This indicates that lower pH values result in the formation of a greater number of pores and increased surface area within the RF polymer retained in the xerogels [53]. Moreover, it can be explained that the larger carbonate ions have a trigonal planar molecular geometry, which may cause steric hindrance. Consequently, the condensation of the intermediates leads to the generation of larger pores of the samples [54]. The use of a higher amount of catalyst leads to more rapid gelation, resulting in a less uniform structure with fewer and larger pores. Alternatively, the catalyst itself may interfere with the formation of crosslinks within the RF polymer, leading to a less porous structure. Therefore, higher amounts of catalyst used during the synthesis have a similar effect on the surface area and porosity of the resulting material, as observed in the results obtained by [55].
The determination of the isoelectric point (IEP) and point of zero charges (pH pzc ) of xerogels and magnetic xerogels was carried out by measuring the zeta potential and pH, as shown in Table 2. The IEP and pH pzc of MX1 and MC4 prepared by direct and indirect sonication, respectively, with R/C 200 are in a similar range of values. However, the RF xerogel exhibits lower IEP and pH pzc values compared to the other materials. These findings are similar to the results reported by [56], where organic xerogels demonstrated a pH pzc value of 3. Figure 6 shows the particle distribution of xerogel and magnetic xerogels prepared using the sol-gel method under ultrasonic irradiation. The particle size distribution in the obtained xerogels may vary because of sonication-assisted synthesis and variations in the R/C ratios. RFX exhibits a broader particle size distribution with larger particles compared to MX1 and MC4, which were prepared with the same molar ratios and drying process. RFX, prepared without sonication, showed a larger particle size, which is consistent with the findings of [57].
It can be observed that the average particle diameter of MX1 (28.05 nm) was lower than that of MC4 (32.65 nm), which is similar to the results obtained from SEM analysis. The use of direct sonication in the preparation of MX1 resulted in a narrower particle distribution due to localized energy transfer, leading to more consistent particle sizes in the obtained xerogel. On the other hand, MC4 exhibited a wider range of particle sizes due to less precise control of sonication energy distribution. Therefore, the direct method of sonication generally leads to a lower particle size distribution compared to the indirect method, due to the more localized and intense energy transfer that promotes effective fragmentation and reduction in particle size. Similar findings of the study of [58].
The initial pH of the solution is a factor influencing the polymerization of xerogels, especially when varying the molar ratio of the catalyst. The pH values of the RF solutions for MC1, MC2, MC3, and MC4 were 7.26, 7.05, 6.92, and 6.82, respectively, within the similar range of the study of [52]. It can be observed that higher catalyst concentrations with lower R/C molar ratios result in smaller particles and pore sizes, as reported by [53].
influenced by pH. Higher pH values generally promote faster gelation, while lowe values slow down the process. The gelation kinetics can significantly impact the ov pore structure and porosity of the xerogel. When the condensation reaction occurs i presence of small particles resulting from the high pH conditions, it produces mat with smaller pores, leading to a higher density or more compact RF gel structure On the other hand, lower pH values may result in a less densely crosslinked structu   The pH of the precursor solution plays a crucial role in determining the final structure of the obtained xerogel. It affects the kinetics of polymerization and crosslinking reactions, as well as the condensation and gelation processes. The mechanism of polymerization in RF gels involves two steps: the addition reaction to form hydroxymethyl derivatives of resorcinol and the condensation of these derivatives to form methylene or methylene ether bridged compounds [49]. In a high pH solution, the first addition reaction is favored. This leads to a higher rate of polymerization and crosslinking, resulting in a more extensively crosslinked network structure and a relatively quick process. This process often yields small nodules and narrow mesopores. Gelation kinetics, which refers to the rate of transition from a liquid precursor solution to a gel network, is strongly influenced by pH. Higher pH values generally promote faster gelation, while lower pH values slow down the process. The gelation kinetics can significantly impact the overall pore structure and porosity of the xerogel. When the condensation reaction occurs in the presence of small particles resulting from the high pH conditions, it produces materials with smaller pores, leading to a higher density or more compact RF gel structure [59]. On the other hand, lower pH values may result in a less densely crosslinked structure. Figure 7a depicts the FTIR spectra of RFX and magnetic gels prepared using ultrasonication with direct and indirect techniques, covering a wavelength range of 4000-400 cm −1 . The characteristic FTIR bands of RFX, MX, and MC are similar. However, MC4, MX1, and MX2 exhibit an FTIR band at 478 cm −1 attributed to Fe-O stretching vibration [60,61]. The profiles of RFX, MC4, and MXs show the presence of six absorption bands: (i) O-H stretching at 3300 cm −1 , (ii) C-H stretching at 2900 cm −1 , (iii) C = C stretching in the aromatic ring at 1600 cm −1 , (iv) C-H bending vibration at 1400 cm −1 , (v) C-O stretching at 1200 cm −1 , and (vi) methylene ether C-O-C linkage stretching between two resorcinol molecules at 1000 cm −1 [62]. The FTIR spectra of RF gel and MC1-MC4 can be observed in Figure 7b, and all of them exhibit bands that are correlated with the bands described above.
ing at 1200 cm −1 , and (vi) methylene ether C-O-C linkage stretching between two resorcinol molecules at 1000 cm −1 [62]. The FTIR spectra of RF gel and MC1-MC4 can be observed in Figure 7b, and all of them exhibit bands that are correlated with the bands described above. Regarding the characterization of MCs and MXs, it can be observed that the preparation of monolithic resorcinol-formaldehyde xerogels involved different methods of sonication, utilizing low and high intensity, respectively. However, the results of their XRD and FTIR analyses show significant similarities. This is in contrast to the study conducted by [48], where the preparation of ZnO nanoparticles using direct and indirect sonication had an impact on the crystalline structure (XRD analysis) and resulted in different IR spectra of the samples. Due to the probable growth mechanisms of ZnO nanoparticles, various crystallization mechanisms were proposed. However, in the case of xerogels, ultrasonic irradiation aids in promoting aging and hydrophobization reactions. Additionally, [37] discovered that the preparation of silica xerogels can be accomplished in less than 1/5 of the time required by conventional methods.

Performance of Adsorption of Arsenic Using MCs and MXs
In the batch adsorption experiment of As(V) using MCs and MXs, the effect of pH in the range of 2 to 7 was used to evaluate their adsorption capacities, as shown in Figure 8. MC1 and MC2 demonstrated high adsorption capacities, qe were more in the range of 63.26-73.47 µg/g and 59.18-61.22 µg/g, respectively, than other materials. MX1 and MX2 showed higher adsorption capacity in the acidic solution. Due to the pHpzc being the zero net charge on the surface of the adsorbent, the adsorbent surfaces are charged positively or negatively, depending on whether the pH of the solution is lower or higher than the pHpzc values, respectively [55]. The analysis result of pHpzc of MX1 was 4.54, meaning that MX1 adsorbed As(V) at pH values lower than this value. The same can be described for the adsorption of MC1 and MC2, whose pHpzc values were 6.63 and 6.12, respectively. Regarding the characterization of MCs and MXs, it can be observed that the preparation of monolithic resorcinol-formaldehyde xerogels involved different methods of sonication, utilizing low and high intensity, respectively. However, the results of their XRD and FTIR analyses show significant similarities. This is in contrast to the study conducted by [48], where the preparation of ZnO nanoparticles using direct and indirect sonication had an impact on the crystalline structure (XRD analysis) and resulted in different IR spectra of the samples. Due to the probable growth mechanisms of ZnO nanoparticles, various crystallization mechanisms were proposed. However, in the case of xerogels, ultrasonic irradiation aids in promoting aging and hydrophobization reactions. Additionally, [37] discovered that the preparation of silica xerogels can be accomplished in less than 1/5 of the time required by conventional methods.

Performance of Adsorption of Arsenic Using MCs and MXs
In the batch adsorption experiment of As(V) using MCs and MXs, the effect of pH in the range of 2 to 7 was used to evaluate their adsorption capacities, as shown in Figure 8. MC1 and MC2 demonstrated high adsorption capacities, q e were more in the range of 63.26-73.47 µg/g and 59.18-61.22 µg/g, respectively, than other materials. MX1 and MX2 showed higher adsorption capacity in the acidic solution. Due to the pH pzc being the zero net charge on the surface of the adsorbent, the adsorbent surfaces are charged positively or negatively, depending on whether the pH of the solution is lower or higher than the pH pzc values, respectively [55]. The analysis result of pH pzc of MX1 was 4.54, meaning that MX1 adsorbed As(V) at pH values lower than this value. The same can be described for the adsorption of MC1 and MC2, whose pH pzc values were 6.63 and 6.12, respectively.  Table 3 presents the molar ratios used in the synthesis of five magnetic xerogels, specifically MX3-MX7. These xerogels were prepared with molar ratios of M/R of 0.03, 0.05, 0.1, 0.15, and 0.2, respectively. The xerogels were synthesized using direct ultrasonicassisted synthesis with an ultrasonic VCX130 operating at 130 watts and a 1/4" diameter probe. The M/R ratios increased with the increasing Fe contents, as determined by chemical composition analysis using ICP-Optical Emission Spectroscopy. However, MX6 and MX7 exhibited similar Fe content values. In this case, it can be explained that the high quantity of magnetite may not have fully incorporated into the gel matrix and some of it may have washed out during the solvent exchange, as evidenced by the observation of a brown solution after changing the acetone solution.  MX8-MX11 were synthesized using indirect sonication via the Q700 sonicator, which has a power output of 700 watts and a 1/2" diameter probe. The molar ratios used in the synthesis, along with the corresponding Fe content, are shown in Table 4. At the same molar ratios of M/R at 0.15 for direct (MX7) and indirect (MX11) sonication, MX11 demonstrated a higher Fe content than MX7. The theoretical calculations of Fe content for MX8, MX9, MX10, and MX11 are 4.27%, 6.85%, 12.52%, and 17.29%, respectively. These values are similar to the results obtained from ICP-OES analysis. This can be explained that increasing the power output to 700 watts makes the system more homogenous.  Figure 9, with an increasing amount of magnetite through direct sonification. Figure 9a-e shows the SEM images of the surface morphology of magnetic xerogels composed of large numbers of microclusters with a three-dimensional network. However, some parts of them are agglomerated, and some bright particles can be observed. The elemental distribution of these particles can be confirmed with the corresponding EDX spectra, which demonstrate the existence of iron (Fe), oxygen (O), carbon (C), aluminum (Al), and sodium (Na). Gels 2023, 9,618 15 of 30  Elemental mapping was analyzed to observe the distribution of synthesized magnetite on RF matrix gels, as shown in Figure 9f-j. The individual EDX mapping of Fe element distributions: blue = low, green = medium, and red = high. Magnetite particles were observed in a blue color and were evenly distributed on the RF surface, with a higher quantity corresponding to the increasing M/R ratios. Similar results were obtained from Table 3. The results of the mapping analysis show that the incorporation of Fe into the structure of the samples is homogeneously distributed, similar to the results obtained from activated carbon xerogels doped with iron (II) phthalocyanine by ultrasonication [63]. However, some of them had some accumulation of Fe due to the increase of high concentration of magnetite in the RF solution. It can be observed in Figure 9h,i, where EDX mappings for Fe display a red color in several regions, indicating a high concentration of Fe within the RF gels. The agglomeration of the microclusters and the presence of bright particles on the surface of the MX5-MX7 xerogels suggest that the synthesis process could be improved. Further studies are needed to optimize the synthesis conditions in order to produce magnetic xerogels with improved properties.
The XRD patterns of MX3-MX7, prepared by direct sonication, and MX8-MX11, prepared by indirect sonication, are shown in Figure 10a,b, respectively. Both sets of samples were synthesized with different molar ratios and utilized different ultrasonic processors. However, both sets varied the M/R ratios from 0.03 to 0.2 for MX3-MX7 and from 0.03 to 0.15 for MX8-MX11. Consequently, the XRD analysis of the magnetic xerogels demonstrated the presence of magnetite, in accordance with the JCPDS card assignments, as described in Figure 1. The diffraction peaks at d311 (2 θ = 35.68 • ) appeared high and sharp for all materials, indicating their magnetic properties [42], the intensity of the iron phase peaks increased with higher M/R ratios in the synthesis. These results are particularly relevant for the analysis of the chemical composition, as presented in Tables 3 and 4. The FTIR spectra of monolithic resorcinol-formaldehyde xerogels prepared by direct sonication, with varying M/R molar ratios of 0.05, 0.1, 0.15, and 0.2 (referred to as MX4, MX5, MX6, and MX7, respectively) are similar, as shown in Figure 11a. Similarly, Figure 11b presents FTIR spectra of MX8-MX11, prepared by indirect sonication, which exhibit similarities. The resulting FTIR spectrum displays peaks corresponding to different vibrational modes of the molecules in the sample, as discussed in detail in Figure 7. Both groups of materials exhibit an FTIR band at 468 cm −1 , attributed to Fe-O stretching vibration [60,61]. Therefore, the use of different sonication methods and power outputs of the ultrasonic processor has no effect on the functional groups and chemical compounds present in the samples of monolithic resorcinol-formaldehyde xerogel, based on the absorption of infrared radiation with wavelength ranges of 4000-400 cm −1 . The FTIR spectra of monolithic resorcinol-formaldehyde xerogels prepared by direct sonication, with varying M/R molar ratios of 0.05, 0.1, 0.15, and 0.2 (referred to as MX4, MX5, MX6, and MX7, respectively) are similar, as shown in Figure 11a. Similarly, Figure 11b presents FTIR spectra of MX8-MX11, prepared by indirect sonication, which exhibit similarities. The resulting FTIR spectrum displays peaks corresponding to different vibrational modes of the molecules in the sample, as discussed in detail in Figure 7. Both groups of materials exhibit an FTIR band at 468 cm −1 , attributed to Fe-O stretching vibration [60,61]. Therefore, the use of different sonication methods and power outputs of the ultrasonic processor has no effect on the functional groups and chemical compounds present in the rect sonication, with varying M/R molar ratios of 0.05, 0.1, 0.15, and 0.2 (referred to as MX4, MX5, MX6, and MX7, respectively) are similar, as shown in Figure 11a. Similarly, Figure 11b presents FTIR spectra of MX8-MX11, prepared by indirect sonication, which exhibit similarities. The resulting FTIR spectrum displays peaks corresponding to different vibrational modes of the molecules in the sample, as discussed in detail in Figure 7. Both groups of materials exhibit an FTIR band at 468 cm −1 , attributed to Fe-O stretching vibration [60,61]. Therefore, the use of different sonication methods and power outputs of the ultrasonic processor has no effect on the functional groups and chemical compounds present in the samples of monolithic resorcinol-formaldehyde xerogel, based on the absorption of infrared radiation with wavelength ranges of 4000-400 cm −1 .

Performance of Adsorption of Arsenic Using MX4-MX7 and MX8-MX11
Figure 12a presents the removal efficiency of As(V) using MX4-MX7 prepared by direct sonication via a sonicator with 130 watts of power. The removal efficiency of MX4-MX7 was higher than RFX, which was prepared without using magnetite. In par-  Figure 12a presents the removal efficiency of As(V) using MX4-MX7 prepared by direct sonication via a sonicator with 130 watts of power. The removal efficiency of MX4-MX7 was higher than RFX, which was prepared without using magnetite. In particular, MX4 with a lower loading of Fe 3 O 4 (M/R = 0.03) gave the highest arsenic removal of 58.78%. Meanwhile, arsenic removals were lower with MX5, then increased and remained constant for MX6 and MX7. This can be explained by the capacity of the sonicator. With a low power output sonication and small diameter tip, it was possible to homogenize the solution well with a low quantity of magnetite. However, with increasing magnetite loading into the RF solution with M/R of 0.05, 0.07, and 0.15, the As removal results were similar. This can be confirmed with SEM/EDX analysis (Figure 9), which showed that magnetite was more homogeneously distributed in MX4 than in the other materials.  (Figure 9), which showed that magnetite was more homogeneously distributed in MX4 than in the other materials. Figure 12b shows the arsenic removal using MX8-MX11 prepared by indirect ultrasonic-assisted synthesis with 700 watts. The removal of MX8 (M/R = 0.03) and MX9 (M/R = 0.05) increased dramatically from 36.49% to 58.78%. With the increasing of M/R to 0.1 and 0.15, their arsenic removal of MX10 and MX11 remained constant, which demonstrated the same behavior as MX4-MX7.

Performance of Adsorption of Arsenic Using MX4-MX7 and MX8-MX11
Additionally, the effect of the molar ratio of R/W and M/R on the total solids content of the materials is shown in Tables 3 and 4. The solid content increases with increasing magnetite loading. At the same molar ratios of M/R, the total solids content also increases with increasing R/W. Moreover, low solids contents result in fragile structures, and very high solids contents result in increased densification of the material that lowers porosity. Therefore, the optimum solids content of the xerogel is 20 w/v% [45].   Additionally, the effect of the molar ratio of R/W and M/R on the total solids content of the materials is shown in Tables 3 and 4. The solid content increases with increasing magnetite loading. At the same molar ratios of M/R, the total solids content also increases with increasing R/W. Moreover, low solids contents result in fragile structures, and very high solids contents result in increased densification of the material that lowers porosity. Therefore, the optimum solids content of the xerogel is 20 w/v% [45].

Characterization of Fe 3 O 4 -Monolithic Resorcinol-Formaldehyde Carbon Xerogels
Some parts of the RF surface of Fe 3 O 4 -Monolithic resorcinol-formaldehyde xerogels (MXRF) were agglomerated due to the formation of magnetite, as shown in Figure 13a. The presence of Fe in the RF gels was determined to be 14.83 w% by AAS, compared to 24.67% of Fe as quantified by EDX in the solid sample. It can be observed that the morphology and EDAX analysis did not change significantly after the adsorption process (Figure 13b).  Figure 14a shows N2 adsorption-desorption isotherms at 77 K, and Figure 4b illustrates the pore size distributions of XRF and MXRF. The analysis results of BET surface area, total pore volume, and average pore size of xerogel adsorbent were 399.19 m 2 /g, 0.517 cm 3 /g, and 5.228 nm, respectively. When magnetite composites were added to xerogels, the porous properties of MXRF for BET surface area, total pore volume, and average pore diameter were 292 m 2 /g, 0.279 cm 3 /g, and 3.81 nm, respectively. Figure 14a shows that the adsorption isotherms of RFX and MXRF adsorbents at a constant temperature of 77 K with N2 as the adsorptive exhibit a linear relationship be-   Figure 14a shows N 2 adsorption-desorption isotherms at 77 K, and Figure 4b illustrates the pore size distributions of XRF and MXRF. The analysis results of BET surface area, total pore volume, and average pore size of xerogel adsorbent were 399.19 m 2 /g, 0.517 cm 3 /g, and 5.228 nm, respectively. When magnetite composites were added to xerogels, the porous properties of MXRF for BET surface area, total pore volume, and average pore diameter were 292 m 2 /g, 0.279 cm 3 /g, and 3.81 nm, respectively. As shown in Figure 15, FTIR analysis of MXRF before and after adsorption of As(III) was obtained using attenuated total reflection (ATR) technique. Absorption peaks at 558 cm −1 are characteristic peaks of Fe-O-Fe, which are indicative of magnetite, confirming the presence of Fe3O4 on the MXRF adsorbent [67]. The bending vibration of the hydroxyl groups (Fe-OH) confirmed the formation of iron oxide in xerogels [68] and O-H groups on the gel surface. These groups are possible to facilitate the adsorption of arsenic by iron oxides composites in the matrices of RF magnetic xerogels [28].
XRD patterns of MXRF and MXRF600 before and after adsorption (Figure 16) clear-  Figure 14a shows that the adsorption isotherms of RFX and MXRF adsorbents at a constant temperature of 77 K with N 2 as the adsorptive exhibit a linear relationship between relative pressure and amount adsorbed. RFX and MXRF exhibited type IV adsorption isotherms with H2 and H4 hysteresis loops, respectively. This implies that RFX contained typical mesoporous materials and MXRF contained micro-and mesoporous adsorbents, similar to the results of the pore size distributions. Figure 14b shows that the main pore diameter sizes of RXF and MXRF are in the range of 2-50 nm, which is defined as mesoporous material. The pore size distribution of MXRF reveals that the average pore diameter was 3.81 nm, which is similar to the results of the narrow centering of PSD of Fe, Co, and Ni doped carbon xerogels [64]. This indicates that the doping with transition metals, such as magnetite, into the xerogels has a similar effect to the composite of magnetite, which affects the reduction of surface areas and total pore volume of the material and makes alterations to their textural properties [65,66]. Similar results were found from SEM analysis, which showed increased agglomeration of particles in RF gels.
As shown in Figure 15, FTIR analysis of MXRF before and after adsorption of As(III) was obtained using attenuated total reflection (ATR) technique. Absorption peaks at 558 cm −1 are characteristic peaks of Fe-O-Fe, which are indicative of magnetite, confirming the presence of Fe 3 O 4 on the MXRF adsorbent [67]. The bending vibration of the hydroxyl groups (Fe-OH) confirmed the formation of iron oxide in xerogels [68] and O-H groups on the gel surface. These groups are possible to facilitate the adsorption of arsenic by iron oxides composites in the matrices of RF magnetic xerogels [28].
As(III) was obtained using attenuated total reflection (ATR) technique. Absorption peaks at 558 cm −1 are characteristic peaks of Fe-O-Fe, which are indicative of magnetite, confirming the presence of Fe3O4 on the MXRF adsorbent [67]. The bending vibration of the hydroxyl groups (Fe-OH) confirmed the formation of iron oxide in xerogels [68] and O-H groups on the gel surface. These groups are possible to facilitate the adsorption of arsenic by iron oxides composites in the matrices of RF magnetic xerogels [28]. XRD patterns of MXRF and MXRF600 before and after adsorption ( Figure 16) clearly demonstrated that they had high intensity peaks that contained a crystalline phase and corresponded to Fe3O4 with the Joint Committee on Powder Diffraction Standards (JCPDS) card No. 19-0629. Therefore, the chemical and structural properties of MXRF and MXRF600 did not change significantly following the carbonization and adsorption process. XRD patterns of MXRF and MXRF600 before and after adsorption (Figure 16) clearly demonstrated that they had high intensity peaks that contained a crystalline phase and corresponded to Fe 3 O 4 with the Joint Committee on Powder Diffraction Standards (JCPDS) card No. 19-0629. Therefore, the chemical and structural properties of MXRF and MXRF600 did not change significantly following the carbonization and adsorption process.

Adsorption of Low and High Concentration of As(III) and As(V) with MXRF and MXRF600
In the adsorption process, contact time is one parameter that is a time-dependent process. Adsorption kinetic studies are important in water treatment. These studies can describe the mechanism of the adsorption process and provide kinetic adsorption constants and valuable information. The experimental data were analyzed with four kinetic models: pseudo first-order, pseudo second-order, Elovich, and Power function.
The effect of contact time on the adsorption process was varied from 10 to 1440 min with different ranges of initial concentration for the low range of As(III) concentrations (25, 50, and 75 µg/L) and high range of concentration for As(III) and As(V) were 514 µg/L and 1034 µg/L, respectively. The adsorption kinetic of As(III) on MXRF is shown in Figure 16. XRD diffractogram of Fe 3 O 4 -monolithic resorcinol-formaldehyde xerogels (MXRF) and carbon xerogel (MXRF600) before and after adsorption of As(III).

Adsorption of Low and High Concentration of As(III) and As(V) with MXRF and MXRF600
In the adsorption process, contact time is one parameter that is a time-dependent process. Adsorption kinetic studies are important in water treatment. These studies can describe the mechanism of the adsorption process and provide kinetic adsorption constants and valuable information. The experimental data were analyzed with four kinetic models: pseudo first-order, pseudo second-order, Elovich, and Power function. The effect of contact time on the adsorption process was varied from 10 to 1440 min with different ranges of initial concentration for the low range of As(III) concentrations (25,50, and 75 µg/L) and high range of concentration for As(III) and As(V) were 514 µg/L and 1034 µg/L, respectively. The adsorption kinetic of As(III) on MXRF is shown in Figure 17. The removal efficiency for As(III) concentration of 75 µg/L increased faster in 10 min and remained constant until 240 min at 97.33%.  In this study, the experimental data were analyzed with nonlinear equations using the Langmuir and Freundlich isotherm models to describe the adsorption of As(III) and As(V) on MXRF600. The Langmuir isotherm model assumes that a monomolecular layer of adsorbate molecules is formed on the adsorbent surface, with each molecule having the same adsorption energy. The Freundlich isotherm model describes the heterogeneity of the surface and the distribution of adsorption energies.
The conditions for the isotherm adsorption were as follows: adsorbent dose of 2 g/L, initial solution pH of 3.0, and contact time of 24 h. The initial concentrations of the As(III) and As(V) solutions were in the range of 0.05-1.27 mg/L and 0.12-3.0 mg/L, respectively. The Langmuir and Freundlich model parameters and regression coefficients are shown in Table 6. The experimental data for the adsorption of As(III) and As(V) on magnetic carbon xerogel monoliths were fitted to the Langmuir models, and the maximum monolayer adsorption capacity (qmax) of As(III) and As(V) were 694.3 and 1720.3 µg/g, respectively, with R 2 values (RSME) of As(III) and As(V) were 0.897 (3.865), and Kinetic parameters and correlation coefficients for As(III) and As(V) adsorption by using MXRF600 were obtained by nonlinear regression as presented in Table 5, including residual root mean square error (RMSE). The condition of As(III) and As(V) adsorption kinetics were pH of 3, dosage of 2 g/L, and initial concentration of As(III) and As(V) solution of 0.514 mg/L and 1.034 mg/L, respectively. The adsorption kinetic models that presented the best fit in the As(III) and As(V) adsorption process were the Power equation and Elovich chemisorption model. It can be observed that MXRF600 demonstrated greater adsorption of As(III) and As(V) than MXRF, implying a higher adsorption capacity. The final step of preparing MXRF600 was to produce a carbon xerogel with a carbonization process for removing the rest of the oxygen and hydrogen groups and improving a thermally stable nanostructure [49]. With the use of high temperature under an inert atmosphere, MXRF and MXRF600 demonstrated modifications in their chemical composition and texture properties, which can be identified with the analysis of XRD, FTIR, N 2 physisorption, and SEM/EDAX, as discussed above.
In this study, the experimental data were analyzed with nonlinear equations using the Langmuir and Freundlich isotherm models to describe the adsorption of As(III) and As(V) on MXRF600. The Langmuir isotherm model assumes that a monomolecular layer of adsorbate molecules is formed on the adsorbent surface, with each molecule having the same adsorption energy. The Freundlich isotherm model describes the heterogeneity of the surface and the distribution of adsorption energies.
The conditions for the isotherm adsorption were as follows: adsorbent dose of 2 g/L, initial solution pH of 3.0, and contact time of 24 h. The initial concentrations of the As(III) and As(V) solutions were in the range of 0.05-1.27 mg/L and 0.12-3.0 mg/L, respectively. The Langmuir and Freundlich model parameters and regression coefficients are shown in Table 6. The experimental data for the adsorption of As(III) and As(V) on magnetic carbon xerogel monoliths were fitted to the Langmuir models, and the maximum monolayer adsorption capacity (q max ) of As(III) and As(V) were 694.3 and 1720.3 µg/g, respectively, with R 2 values (RSME) of As(III) and As(V) were 0.897 (3.865), and 0.901(9.220), respectively. Table 6. Isotherm parameters and correlation coefficients for As(III) and As(V) adsorption on MXRF600.

Conclusions
The ultrasonic-assisted synthesis of Fe 3 O 4 -monolithic resorcinol-formaldehyde xerogels using direct and indirect sonication methods as an easier recovery of adsorbent was shown to reduce the gelation time and improve the textural properties of the final product. The optimal mixing time for magnetite dispersion in an RF aqueous solution was determined to be 5 min using direct sonication and 60 min using indirect sonication, as confirmed by SEM/EDX analysis. This study investigated the effect of different molar ratios of R/C, M/R, R/W, and thermal treatment on RF xerogel. The results show MXRF600 was synthesized by indirect sonication with R/F = 0.5, R/C = 100, R/W = 0.05, and M/R = 0.15 and enhanced adsorption capacity for As(III) and As(V) from groundwater due to the influence of sonication assistance and the carbonization process. However, the optimization of the process parameters for the adsorption of magnetic carbon xerogels should be studied to find out the optimum condition and improve their performance in removing contaminants from the environment. The desorption process, regeneration efficiency, and the lifecycle assessment of magnetic carbon xerogels are suggested for future research.

Synthesis of Adsorbent Materials
The gels were synthesized by polymerizing resorcinol (R, C 6 H 6 O 2 ) and formaldehyde (F, CH 2 O) in water (W), using sodium carbonate (C, Na 2 CO 3 ) as a catalyst, following the procedure described by [69]. The synthesis utilized molar proportions of R/C = 200, R/F = 0.5, and R/W = 0.06 [70,71]. While keeping other factors constant, the effect of loading of Fe (via direct and indirect sonication), Fe content (M/R = 0.03-0.2), water (R/W = 0.04-0.06), and catalyst (R/C = 50-200) ratios were varied in the realization of the monoliths. Then, they were evaluated for their impact on the physicochemical properties of the resulting materials, as well as their ability to remove As(III) and As(V). Iron oxides (M) used in this study were magnetite obtained from Lanxess, García, Nuevo León, Mexico.
The procedure for synthesizing the gels involved placing half of the deionized water and the mass of R in a 100 mL beaker, which was then vigorously shaken to homogenize the solution. The F solution was added, followed by the addition of C, and the mixture was stirred magnetically until homogeneous. pH of RF solution was controlled between 5.5-6.0 to obtain high surface areas of resulting materials [52]. The resulting solution was then placed in Pyrex ® glass tubes, which were sealed with a stopper to prevent evaporation. The temperature of the RF solution during reaction of an ultrasonic processor was controlled to be in the range of 80 to 85 • C. To evaluate the optimal mixing time and the dispersion of magnetite in the RF aqueous solution, three types of ultrasonic devices were applied. First, a digital ultrasonic device (UP400St; Hielscher, Teltow, Germany) with an output of 400 watts, a frequency of 24 kHz, and a 1-inch diameter probe was used in the synthesis of RFX, MC1-MC4, MX1-MX2, and MXRF. The device was equipped with automatic frequency tuning and adjusting an amplitude ranging from 80% to 100%. An ultrasonic processor (VCX 130; Sonics & Materials, Inc., Newton, CT, USA) with a power output of 130 watts, a frequency of 20 kHz, and a 1 4 -inch diameter tip was applied in the synthesis of MX3-MX7. A sonicator (Q700; Qsonica L.L.C, Newtown, CT, USA) with a power rating of 700 watts, a frequency of 20 kHz, and a 1/2-inch diameter probe was used in the synthesis of MX8-MX11. All ultrasonic processors were used for homogenization, dispersal, and deagglomeration of magnetite particles in the RF aqueous solution, using both direct and indirect sonication methods before the gelation process.

Monolithic Resorcinol-Formaldehyde Xerogels Effect of Loading of Magnetite with Direct and Indirect Sonication, and Modification of Catalyst
The study investigated the optimal mixing time and dispersion of magnetite in RF aqueous solution, using both direct and indirect ultrasonication methods prior to the gelation process. Magnetic xerogel monoliths (MCs) were prepared by indirect sonication with molar ratios of R/F = 0.5, R/W = 0.06, and M/R = 0.01, and varying proportions of resorcinol and catalyst. MC1, MC2, MC3, and MC4 were identified based on R/C ratios of 50, 100, 150, and 200, respectively. The homogenization process was carried out using ultrasonic-assisted synthesis, with digital ultrasonic equipment (UP400St; Hielscher, Teltow, Germany), starting at room temperature. After 5 min of sonication, the temperature reached 85 • C. Magnetite (M) was added into the homogeneous RF aqueous solution and subjected to indirect sonication for 60 min to disperse the magnetite particles before the gelation process. Additionally, the variable factors studied in this work include loading of magnetite with direct and indirect ultrasonication. Therefore, MX1 was prepared using the same method as MC4 but with direct sonication to compare their properties and adsorption capacity of arsenic in aqueous solution. Afterward, the materials were placed in the oven at 80 • C for 5 days. In the case of MX2, the gelation and curing process was changed to be left at room temperature for 5 days. In this study, monoliths of magnetic xerogels (MXs) were prepared by the sol-gel polymerization of resorcinol with formaldehyde, using an alkaline catalyst and direct sonication of magnetite to incorporate them into the xerogels. Different proportions of iron oxides were modified to achieve the maximum adsorption capacity. Initially, batches of magnetic xerogel monoliths (MX3-MX7) were prepared by varying the M/R ratio from 0.03 to 0.2. The molar ratios of R/F = 0.5, R/C = 200, and R/W = 0.04 were maintained for a small portion of the batches. The preparation process involved the use of an ultrasonic processor VCX 130 with a power output of 130 watts and a frequency of 20 kHz.

Monolithic Resorcinol-Formaldehyde Carbon Xerogels by Indirect Sonication
The monolithic resorcinol-formaldehyde xerogels (MXRF) were synthesized in a larger batch using UP400St equipment with the relations of molar ratio of R/F = 0.5, R/C = 100, R/W = 0.05, and M/R = 0.15. Then, the gels were cured in a conventional oven for three days at 80 • C. The gels were taken off the glass tubes and allowed to cool to room temperature. After that, the gels were cut using a diamond disk into pellet forms of 5 mm in diameter. The materials were then exchanged with acetone, sealed in a jar with the lid tightly closed, and wrapped with paraffin film. The jar was placed in a shaking water bath (BS-11; Lab Companion, Daejeon, Republic of Korea) at 150 rpm for two days, with fresh acetone being added daily. Subsequently, the gels were dried for three days in a conventional oven at 80 • C.
MXRF were then pyrolyzed using a tube furnace (STF55346C-1; Lindberg/Blue M, Asheville, NC, USA) with the following conditions: temperature of 600 • C, heating ramp of 3 • C/min, time of 6 h, and nitrogen flow of 100 mL/min. The resulting product was monolithic resorcinol-formaldehyde carbon xerogels, which were labelled as MXRF600.

Characterization through Analytical Techniques
To assess the physicochemical characteristics of the synthesized materials, the following techniques were employed: X-ray diffraction (XRD) analysis was used to identify the main constituents and mineralogical phases of the synthesized materials. The analysis was performed using an X-ray diffractometer on MCs, MXs, and MX3-MX11 samples (XPert PW3040; Philips, Almelo, The Netherlands), and on MXRF and MXRF600 (D8 ADVANCE; Bruker, Karisruhe, Germany). Sample preparation involved sieving the sample through a 200-mesh sieve, resulting in an average particle size of 74 µm. A high-temperature chamber attached to the X-ray diffractometer was used to measure diffraction patterns up to 900 • C. Cu(Kα) radiation was applied in a 2θ range from 10 • to 80 • .
Fourier transform infrared spectroscopy (FTIR) was employed to investigate the surface functional groups of the adsorbents before and after arsenic adsorption, in order to understand the mechanism of ion adsorption. FTIR analysis was conducted using a Shimadzu IRAffinity-1S instrument (Shimadzu Corp., Kyoto, Japan) on dry powder samples. Infrared spectra were measured by connecting to the attenuated total reflection (ATR) contained in the disk of crystal (type IIIa monocrystalline diamond). Before the analysis, the samples were sieved through a standard test sieve No. 142 to obtain a uniform particle size of 106 µm. Subsequently, the powder samples were dried in an oven at 60 • C for 15 h under dry air to avoid interference from water vapor adsorption in the infrared region, which could affect the analysis result. After installing the ATR with infrared spectroscopy, the solid samples were directly added to the crystal plate and pressed for surface analysis. All spectra were recorded between the wavenumbers of 400-4000 cm −1 , with 45 scans per sample.
The surface morphology, pore structure, and element analysis of the magnetic xerogels were analyzed using a scanning electron microscope (SEM). MCs, MXs, MXRF, and MXRF600 were analyzed using a field emission scanning electron microscope (FE-SEM) (7800F Prime; JEOL, Tokyo, Japan) after gold coating. MX3-MX11 were analyzed using a scanning electron microscope (SEM) (JSM-IT300; JEOL, Tokyo, Japan). The samples were coated with graphite before the analysis. The acceleration voltages used were between 5 and 20 kV. The textural properties of magnetic xerogels, and magnetic carbon xerogels were characterized by physical adsorption of N 2 at 77 K, using physisorption apparatus (ASAP 2020; Micromeritics, Norcross, GA, USA and NOVA touch 2LX; Quantachrome Instruments, Boynton Beach, FL, USA). The samples were dried at 110 • C for 15 h prior to N 2 physisorption analysis.
Particle Size Distribution (PSD) was determined using Dynamic Light Scattering Analyzers (PMX 500; Microtrac, Meerbuch, Germany), and the data report was generated by FLEX software version 11.1.0. 1 The determination of the point of zero charge (pH pzc ) and the isoelectric point (IEP) was conducted following the methods described by [72]. The pH solutions were prepared by adjusting deionized water to pH values of 2, 4, 6, 8, 10, and 12 using 0.1 M HCl or 0.1 M NaOH solutions. The pH was measured using a multi-parameter device (Orion Star A211; Thermo Scientific, Beverly, MA, USA). A zeta potential analyzer (PMX 500; Microtrac, Meerbuch, Germany) was employed to measure the zeta potential, with pH variations ranging from 2 to 11 for the determination of the isoelectric point (IEP).
The amount of Fe in the magnetic xerogel monoliths was determined by Inductively Coupled Plasma (ICP) Optical Emission Spectrometer (OES) (Optima 8300; Perkin Elmer, Shelton, CT, USA).

Batch Adsorption Experiment
Groundwater used in experimental study was obtained from a well approximately 70 m deep located at Jiutepec, Morelos Mexico. Physical and chemical characteristics of groundwater sample used in this study were analyzed. pH (7.  [4]. Since there was no arsenic in the selected water, arsenic was added to the stock solution prepared for adsorption tests on synthetic samples. This water was used to prepare the corresponding arsenic solution to the required concentrations by adding sodium arsenite (NaAsO 2 , Sigma-Aldrich) and sodium arsenate dibasic heptahydrate (HAsNa 2 O 4 ·7H 2 O, Sigma-Aldrich) for studying As(III) and As(V) adsorption processes, respectively.
The batch adsorption experiment of As(V) using MCs and MXs as adsorbents was conducted to evaluate their adsorption capacities. The effect of solution pH (2-7) on As(V) adsorption was investigated with an initial concentration of 100 µg/L, a dose of 1 g/L, 150 rpm, a contact time of 6 h, and a temperature of 26.2 ± 1 • C.
Batch adsorption of As(V) using MX4-MX7 and MX8-MX11 was studied with direct sonication at low power output, and indirect sonication at high power output, respectively. The following conditions were used: an initial concentration of As(V) of 200 µg/L, pH of 3, a dose of 2 g/L, 150 rpm, a contact time of 6 h, and a temperature of 26.3 ± 1 • C.
XRF600 was carbonized into pellets and used in this form to test kinetics and isotherms. The kinetic study adsorption using MXRF with As(III) concentrations of 0.025, 0.05, and 0.075 mg/L was conducted at a pH of 3, a dosage of 2 g/L, 150 rpm, a temperature of 26.5 ± 1 • C, and contact time ranging from 10 to 1800 min. The adsorption kinetics of As(III) and As(V) using MXRF600 were carried out under a pH of 3, a dosage of 2 g/L, and initial concentration of As(III) and As(V) solution of 0.514 mg/L and 1.034 mg/L, respectively, with a contact time ranging from 10 to 1440 min.
The conditions for the isotherm adsorption using MXRF600 were as follows: an adsorbent dose of 2 g/L, an initial solution pH of 3.0, 150 rpm, a temperature of 26.4 ± 1 • C, and a contact time of 24 h. The initial concentrations of the As(III) and As(V) solutions were in the range of 0.05 to 1.27 mg/L and 0.12 to 3.0 mg/L, respectively.
The importance of kinetic and equilibrium models of adsorption is described in the mechanisms and dynamics of the adsorption system of adsorbents. Adsorption kinetic models that control the adsorption process of arsenic are related to the adsorbate uptake on the adsorbent with chemisorption. Therefore, the Pseudo First-Order (PFO), Pseudo Second-Order (PSO), and Elovich and Power equations were applied to perform the experimental data in this study. The assumptions of the PFO model are: sorption at localized sites, the energy of adsorption is independent of surface coverage, a saturated monolayer of adsorbates, and the concentration of the adsorbate is constant [73]. The assumptions of the PSO model are similar to those of the PFO model. The PSO kinetic equation typically describes metal ion uptake on activated carbons well, as well as the adsorption of dyes, herbicides, oils, and organic compounds from aqueous solutions [73,74]. The Elovich equation is used to describe the kinetics of a heterogeneous diffusion process [74]. It is a semi-empirical equation that is based on the assumption that the rate of diffusion is controlled by the rate of adsorption onto active sites on the heterogeneity of the surface of the adsorbent.
The Langmuir and Freundlich isotherm models are the most commonly used equilibrium models for determining the relative concentrations of the solute adsorbed onto the solid in the solution [75]. The Langmuir isotherm assumes that a solute is adsorbed onto a homogeneous surface with a finite number of similar active sites, forming a monolayer. The Freundlich isotherm is an empirical model that describes multilayer adsorption.
The equations for kinetic and equilibrium models of adsorption used in this study are listed in Table 7. Table 7. Equation of kinetic and isotherm models of adsorption.

Isotherm Models Non-Linear Equations References
Langmuir q e = K L q m C e (1+KLCe) [74] Freundlich q e = K F C e 1 n [74] q t and q e are the amount of adsorbate adsorbed at time t (mg/g) and the equilibrium adsorption capacity (mg/g), respectively. k 1 is the PFO rate constant (min −1 ), and k 2 is the PSO rate constant (min −1 ), respectively. t is the contact time (min). α is the initial adsorption rate (mg/g min), β is related to surface coverage (g/mg), and a and b are constants. C e is the equilibrium concentration of adsorbate in solution (mg/L). q m is the maximum adsorption capacity (mg/g). K L is the Langmuir constant that is related to the adsorption energy (L/mg). K F and n are Freundlich constants that measure the adsorption capacity ((mg/g)(L/mg)1/n) and intensity, respectively.
The arsenic adsorption process was carried out in a batch reactor system. The effect of contact time and initial concentration of arsenic adsorption was investigated on MXRF and MXRF600. Different kinetic and isotherm adsorption models were analyzed using nonlinear regression analysis with the statistical software R v3.5.

Determination of As(III) and As(V)
The determination of arsenic species was performed using hydride generation atomic absorption spectroscopy (HG-AAS) (Varian; SpectrAA220, Mulgrave, VIC, Australia). To analyze As (III) at trace concentrations, AAS must be combined with the hydride generation (HG) technique with citric-citrate buffer [77].