2.1 Materials

Fe₃O₄ was selected as the magnetic material owing to its environmental friendliness. The following materials were used in the synthesis of the BEA zeolite/Fe₃O₄ composite (magnetic RHAS-BEA). RHAS was extracted from rice husk ash provided by CP All Public Co., Ltd., Thailand, using an alkaline extraction, followed by acid precipitation (see Supplementary Information for details). All other chemicals used in this study were of reagent grade. The Al source, alkali source, structure-directing agent (SDA), and magnetic particles were Al₂(SO₄)₃·16H₂O (Nacalai Tesque, Japan), NaOH (Nacalai Tesque, Japan), tetraethyl ammonium hydroxide (TEAOH; 35 wt%, Tokyo Chemical Industry, Co., Ltd., Japan), and Fe₃O₄ (Kanto Chemical, Co., Inc., Japan) with Heywood diameters of primary particles in the range 100–300 nm. To perform the adsorption experiments, paraquat solutions were prepared by dissolving appropriate amounts of paraquat dichloride (C₁₂H₁₄Cl₂N₂, Tokyo Chemical Industry, Co., Ltd., Japan) in distilled water (DW). The pH of the solution was adjusted using 0.1 mol/L hydrochloric acid, diluted from concentrate hydrochloric acid (HCl; 36 wt%, Nacalai Tesque, Japan), and 0.1 mol/L NaOH, prepared by dissolving NaOH (Nacalai Tesque, Japan) in DW. A sodium chloride solution (0.01 mol/L) was prepared by dissolving sodium chloride (NaCl; Nacalai Tesque, Japan) in DW and was used to determine the pH at the point of zero charge. Silicon (Si, 1.00 mgSi/mL, Nacalai Tesque, Japan), aluminum (Al, 1002 mg/L, Fujifilm Wako Pure Chemical Corporation, Japan), and iron (Fe, 1001  mg/L, Fujifilm Wako Pure Chemical Corporation, Japan) standard solutions were used in the quantification of leached ions during the adsorption test. Diquat dibromide monohydrate (Fujifilm Wako Pure Chemical Corporation, Japan) and Cibacron blue 3GA (CB, Sigma-Aldrich, USA) were used in the paraquat co-adsorption test.

2.2 Synthesis of magnetic RHAS-BEA

Magnetic RHAS-BEA was synthesized based on a previously reported DGC protocol (Phouthavong et al., 2020) and modified for RHAS utilization (Fig. 1). RHAS (3.63 g) was dissolved in 1.00 g of a 20 wt% NaOH solution (prepared by dissolving NaOH pellets in DW) and 1.033 g of a 10 wt% TEAOH solution (diluted from 35 wt% TEAOH in DW) by employing mechanical stirring for 30 min. Ultrasonication was then applied, using an ultrasonic bath (US-104, SND Co., Ltd., Japan), for 2 h to improve the dispersion and dissolution of RHAS before the addition of 0.30 g of the Al source. Ultrasonication was performed again for 1 h. The mixture was then continuously stirred in a closed container for 16 h before the temperature was increased to 80 °C. After the solution was stirred for 8 h at 80 °C, 1.12 g Fe₃O₄ was introduced to the mixture. For comparison, Fe₃O₄ was excluded when preparing the non-magnetic RHAS-BEA sample (abbreviated as RHAS-BEA hereafter). The container was opened until the precursor solution transformed into a homogeneous and dry precursor gel. The dried gel was collected and crushed before being placed in a small container inside an autoclave. An equivalent weight of water was added to the autoclave without direct contact with the gel. The autoclave was then heated to 180 °C for 12 h in a convection oven (SOFW-300SB, AS ONE Corporation, Japan) to obtain the product, which was easily collected (i.e., it was dry) without washing. The pre-crushed products were calcined in an electric furnace (VF-3000, SK Medical Electronics Co. Ltd., Japan) at 450 °C for 12 h (in air) to remove the SDA before use.

Close

Fig. 1

Synthesis of magnetic RHAS-BEA using DGC method.

Export to PPT

2.3 Characterization

The presence of the BEA zeolite and iron oxides in the samples was analyzed by performing X-ray diffraction (XRD; UltimaIV, Rigaku Corporation, Japan) equipped with a Cu-Kα X-ray source (α = 1.5405 Å). The morphologies of the synthesized samples were observed by applying field-emission scanning electron microscopy (FE-SEM; JSM-6330F, JEOL, Japan). Ultra-high-resolution FE-SEM-energy dispersive X-ray spectroscopy (FE-SEM-EDS; S-4800, HITACHI, Japan) was used to determine the Si/Al ratio of the samples and observe the distribution of the magnetic particles in the magnetic RHAS-BEA sample.

2.4 Paraquat adsorption experiments

2.5 Separability of the adsorbents from aqueous solutions

The separability of the adsorbents from the aqueous solutions was examined by monitoring their sedimentation in the absence and presence of magnets (Hagio et al., 2019; Phouthavong et al., 2020). Suspended RHAS-BEA and magnetic RHAS-BEA particles at a concentration of 100 mg/L in DW were shaken for 2 h before being allowed to stand in the absence or presence of a neodymium (Nd) magnet (surface flux density of 0.45 T). The turbidity of the supernatant was measured in formazin turbidity units; specifically, the turbidity of the supernatant was consecutively measured for 0.5–60 min using a turbidimeter (HI93703, Hanna Instruments, USA). Suspended solids (SS) (mg/L) at each monitored standing time were calculated from the turbidity values using a calibration curve.

2.6 Reusability of the adsorbents and ions leached from the adsorbents

Spent RHAS-BEA and magnetic RHAS-BEA were collected after one paraquat adsorption test, regenerated by recalcination at 450 °C for 6 h, and reused in subsequent paraquat adsorption cycles. Adsorption and regeneration cycles were repeated to evaluate the reusability. Each adsorption cycle was performed in duplicates. The ions leached from the adsorbents were examined by shaking the adsorbents (1 g/L) in water and paraquat solution (1 mg/L) for 3 h at 26 °C. The solutions were collected to quantify potentially leached elements, such as Si, Al, and Fe, using inductively coupled plasma-atomic emission spectrometry (S-II, Seiko Instrument, Japan).

2.7 Paraquat co-adsorption test with blue dye

The selectivity in the presence of blue dye added to commercial paraquat was tested in practical applications. Aqueous solutions of paraquat (40 μmol/L, equivalent to 10 mg/L) were prepared separately and mixed with CB at two different concentrations (4 and 40 μmol/L). Single and binary adsorption tests were performed using a ratio of 1 g adsorbent per 1 L solution, contact time of 24 h at 26 °C, and the pH values of the solutions were not adjusted. The concentrations of CB and paraquat in the solutions were determined from the absorbances recorded using a UV–Vis spectrophotometer (UV-2450, Shimadzu Co., Japan) at 600 and 257 nm, respectively. The blue dye also showed intense UV absorption at λ_257nm which is the analytical detection wavelength of paraquat (Fig. S1). Thus, the co-absorption of CB at λ_257nm was subtracted from the absorption of paraquat in the binary mixture using the constant ratio between the absorbance at λ_257nm and λ_600nm of a single CB.

2.8 Paraquat co-adsorption test with diquat

Aqueous solutions of paraquat (40 μmol/L, equivalent to 10 mg/L) were prepared separately and mixed with diquat at equivalent molarity. The adsorption was performed using 1 g adsorbent per 1 L solution, contact time of 24 h at 26 °C, and the pH values of the solutions were not adjusted (pH 5.8). UV absorption spectra of paraquat and diquat were recorded using the UV–Vis spectrophotometer.

3.1 Characterization of the RHAS-BEA and magnetic RHAS-BEA

The XRD pattern of RHAS (Fig. 2a) has a broad peak, indicative of its highly amorphous structure; additionally, no crystalline phase was detected. The RHAS particles have ultrafine sizes (i.e., they are below 100 nm) (inset of the scanning electron microscopy (SEM) image in Fig. 2a). For synthesizing advanced materials, amorphous SiO₂ is the preferred source of SiO₂ owing to its high reactivity (Santana Costa and Paranhos, 2018; Steven et al., 2021). Prior to the inclusion of the Fe₃O₄ particles, preliminary studies were conducted to confirm the crystallization of BEA zeolites using RHAS via the DGC method. Initially, the precursor gel was prepared by mixing RHAS with an Al source in NaOH and TEAOH solutions, using simple mechanical stirring at room temperature (26 °C). The DGC process primarily produced an amorphous product with a minor crystalline phase at 2θ = 32.0° and 34.0°, which can be classified as a CHA zeolite (Criado et al., 2007; Dang et al., 2020), as indicated by the XRD pattern and the inset SEM image in Fig. 2b. This product may have been brought forth due to the insufficient dissolution of the SiO₂ source for the generation of SiO_4/2 tetrahedra to form a dense aluminosilicate gel. Therefore, the amount of NaOH, temperature, and aging time were increased to facilitate optimal dissolution. These increases significantly improved the synthesis, as indicated by the characteristic XRD peaks (2θ = 7.6° and 22.4°) which were congruent with the standard card (JCPDS 48–0074) of BEA zeolite (El Maksod et al., 2017), and some crystals with clear boundary shapes in the inset SEM image in Fig. 2c. Overall, the crystallinity and morphology of the BEA zeolite significantly improved, as shown in Fig. 2d, with the assistance of ultrasonication for the dissolution of RHAS and mixing with the alumina source. The observed XRD peaks agreed well with those of BEA zeolites that were synthesized in other studies using commercial colloidal SiO₂ via the DGC method (Inagaki et al., 2007; Matsukata et al., 2002; Phouthavong et al., 2020). Ultrasonication was used to prepare a dry precursor gel containing Fe₃O₄ particles. The BEA zeolite phase was still detectable in the presence of Fe₃O₄, as observed through XRD (Fig. 2e).

Close

Fig. 2

(a) XRD pattern of RHAS with its inset SEM image, and XRD patterns with their inset SEM images of as-synthesized products using gels prepared at room temperature (b), at 80 °C with increased NaOH (c), at 80 °C with increased NaOH and ultrasonication for non-magnetic (d) and magnetic RHAS-BEA prepared under the same conditions as (d) (e).

Export to PPT

The as-synthesized non-magnetic and magnetic RHAS-BEA samples were calcined at 450 °C for 12 h to remove the SDA and open their pores for paraquat adsorption. The XRD patterns of the calcined products are shown in Fig. 3a and 3b, displaying the characteristic peaks of BEA in both samples. These results confirm that the BEA structure did not collapse after the removal of the SDA through calcination. The XRD peaks in the 2theta range of 30°–40° corresponded to the incorporated Fe₃O₄ (JCPDS 19–0629) in the magnetic sample (Fig. 3b), while that of a small amount of hematite (α-Fe₂O₃, JCPDS 01–1053) was also detected, indicative of the slight oxidation of Fe₃O₄ caused by in-air calcination (Fauzi and Ratnawulan, 2021).

Close

Fig. 3

XRD patterns of (a) RHAS-BEA and (b) magnetic RHAS-BEA samples calcined at 450 °C (12 h).

Export to PPT

The FE-SEM images in Fig. 4 show the morphologies of the calcined samples. We found that the crystals of both samples predominantly had rectangle-like faces; this is typical of BEA-type zeolites prepared by using the DGC method (Inagaki et al., 2007; Matsukata et al., 2002). SEM observations revealed the presence of aggregates with particle sizes below 1 μm (Fig. 4a). However, the aggregate size of the magnetic RHAS-BEA sample was smaller than that of the non-magnetic RHAS-BEA sample. The presence of Fe₃O₄ particles could have decreased the aggregate size of the product; this could have occurred because of the larger number of nuclei created between the Fe₃O₄ particles in the precursor gel (Hagio et al., 2019; Miao et al., 2022). Additionally, the aggregate size distribution of the precursor gel may affect the aggregate size of the crystallized products. Higher-magnification images (Fig. 4b) showed that the samples have rough surfaces, indicative of nanoparticle aggregation (Inagaki et al., 2007). Notably, the incorporation of the Fe₃O₄ particles into the magnetic samples hardly changed the morphology of the BEA zeolite. FE-SEM-EDS elemental mapping revealed that the Si/Al ratios of RHAS-BEA and magnetic RHAS-BEA were 48.8 and 45.0, respectively, and therefore could be classified as high-SiO₂ zeolites (Li et al., 2023). Fig. 5 depicts the distribution of Si and Fe in magnetic RHAS-BEA at almost the same positions; this shows that all BEA particles contain Fe₃O₄ particles. Therefore, these results confirm the success of using amorphous RHAS as the SiO₂ source in the DGC method to crystallize BEA-type zeolites and that the inclusion of Fe₃O₄ particles did not hinder the crystallization of these zeolites.

Close

Fig. 4

FE-SEM photomicrograph of RHAS-BEA and magnetic RHAS-BEA samples observed at × 5000 magnification (a) and × 100000 magnification (b).

Export to PPT

Close

Fig. 5

FE-SEM-EDS elemental mapping of the calcined magnetic RHAS-BEA sample.

Export to PPT

3.2 Paraquat adsorption studies

In kinetic studies conducted at 26 °C, a significant amount of paraquat was quickly adsorbed on RHAS-BEA and magnetic RHAS-BEA within 1 min in the first stage of contact time (Fig. 6a) and sharply increased within 10 min. This is indicative of the high adsorption rates of the synthesized adsorbents. Subsequently, adsorption slowed and reached saturation at 60 min. Fast adsorption of paraquat has also been reported in previous studies using zeolites such as FAU (Pukcothanung et al., 2018), modified FAU (Nur et al., 2005), and MFI (Walcarius and Mouchotte, 2004). The non-magnetic and magnetic RHAS-BEA revealed similar adsorption kinetic trends, but the amount of adsorbed paraquat was lower for the magnetic sample. This is likely due to the lower fraction of the adsorbent (i.e., BEA zeolite). These details are discussed in the next paragraph along with an isotherm study. Using widely applied linear adsorption reaction kinetic models, the adsorption of paraquat on both BEA samples best fit the pseudo-second order model according to the coefficient of determination (R²) shown in Fig. 6b and 6c. This indicates that the surface reaction (i.e., chemisorption) occurred in the electrostatic interaction between paraquat and the adsorbents (Moradi, 2014). Similar pseudo-second-order kinetic constants (k₂) for paraquat adsorption on the RHAS-BEA and magnetic RHAS-BEA samples (i.e., 0.0087 and 0.0078 g/mg min, respectively) revealed that the inclusion of Fe₃O₄ particles in the BEA zeolite had a negligible effect on the paraquat adsorption rate. The intraparticle diffusion model also showed a good fit (Table S1) considering multiple-step adsorption (Özacar, 2003); however, considering that the trend line did not pass through the origin in the first step (Fig. 6d), it was concluded that this was not the primary adsorption mechanism in this study. In conclusion, the overall rates of paraquat adsorption on both RHAS-BEA and magnetic RHAS-BEA may be controlled by a surface reaction step occurring through a chemisorption mechanism, as described by the pseudo-second-order model, rather than by the intraparticle diffusion step.

Close

Fig. 6

(a) Change in amount of paraquat adsorbed on RHAS-BEA and magnetic RHAS-BEA over time and linear adsorption kinetic model plots including the (b) pseudo-first-order, (c) pseudo-second-order, and (d) intraparticle diffusion model. The initial concentration of paraquat was 75 mg/L, pH = 5.8 (non-adjusted).

Export to PPT

The results of the adsorption isotherm study (Fig. 7a) show an increase in the amount of adsorbed paraquat with increasing C_e which became saturated at approximately 70 and 50 mg/g for RHAS-BEA and magnetic RHAS-BEA, respectively. Two common linear adsorption isotherm models, namely, the Langmuir and Freundlich models, were used to explain the adsorption behavior. The Langmuir model was found to better describe the adsorption behavior of paraquat for both non-magnetic and magnetic RHAS-BEAs (Fig. 7b) than the Freundlich model (Fig. 7c). This model showed that paraquat is adsorbed as a monolayer on homogeneous adsorption sites on the surface of RHAS-BEA or magnetic RHAS-BEA (Rasaie et al., 2021). The Langmuir adsorption model was found to be the best fit for describing the adsorption of paraquat in zeolites (Insuwan and Rangsriwatananon, 2017; Keawkumay et al., 2019; Osakoo et al., 2017; Rongchapo et al., 2013; Rongchapo et al., 2015), mesoporous SiO₂ (Osakoo et al., 2017; Rasaie et al., 2021; Rongchapo et al., 2013, 2015), zeolite/mesoporous SiO₂ composites (Osakoo et al., 2017), graphene oxide/mesoporous SiO₂ composites (Dehghani et al., 2021), clays (Ait Sidhoum et al., 2013), and modified biopolymers (Fernandes et al., 2017; Huang et al., 2019a; Junthip et al., 2019). A plot based on the Langmuir model was used to determine the maximum adsorption amount of paraquat (q_m) of RHAS-BEA or magnetic RHAS-BEA to be 75.64 ± 1.09 and 54.71 ± 4.19 mg/g, respectively. Theoretically, the calculated BEA zeolite mass percentage in the calcined magnetic RHAS-BEA sample is approximately 72.6% (see Supplementary Information), which is believed to be the major contributor to paraquat adsorption. Because the maximum adsorption amount of magnetic RHAS-BEA was 72.3%, which is similar to 72.6%, the reason for the lower adsorption amount when using the magnetic sample was likely due to the fraction of iron oxide in the sample.

Close

Fig. 7

(a) Adsorption isotherm of paraquat on RHAS-BEA and magnetic RHAS-BEA (error bars indicate the standard deviation), and a linear adsorption isotherm model plotted using (b) Langmuir and (c) Freundlich models. The initial concentration of paraquat was in the range 10–750 mg/L, pH = 5.8 (non-adjusted), contact time = 24 h. The experiments were carried out in duplicates.

Export to PPT

3.3 Effect of initial pH on paraquat adsorption and possible adsorption mechanism

In this study, the effect of the initial pH of the paraquat solution on adsorption capacity was investigated. The initial unadjusted pH of the tested paraquat solution was approximately 5.8, and we evaluated its adsorption amount in the pH range 3–9. pH values outside this range were not considered because: natural water is generally neutral; highly acidic conditions induce dealumination; and highly alkaline conditions trigger the dissolution of SiO₂. Overall, we found that the adsorption of paraquat onto the samples was pH-dependent in aqueous solutions. The results are shown in Fig. 8a. The q_e values were lower when the pH decreased from 5.8 to 3, but tended to increase when the pH was above 5.8, and plateauing once it went above 7. This trend was observed for both non-magnetic and magnetic RHAS-BEA samples. As shown in the paraquat structure (inset in Fig. 8a), no protons remained at the two positively charged quaternary amine groups (i.e., for further protonation or deprotonation), and its cationic form thus existed over a wide pH range. By determining the pH_pzc, we found that the pH variation greatly affected the surface charges on the adsorbents (Fig. 8b). Specifically, the pH_pzc values for RHAS-BEA and magnetic RHAS-BEA were 6.3 and 5.9, respectively. At pH values below these points, the adsorbent surface charge is positive mainly because of the protonation of the hydroxyl groups (Mokrzycki et al., 2022), which electrostatically repel the positively charged amine of paraquat. Therefore, the values of q_e were lower in this pH region. Conversely, hydroxyl groups are deprotonated at pH values above the pH_pzc, generating negative charges on the adsorbent surface, which promote electrostatic interactions with the paraquat cation in addition to cation exchange at the alumina sites on the BEA zeolite framework. These contributions increased the q_e values. According to the pH results, paraquat adsorption can be improved through an increase of the initial pH. In this study, the q_m values were evaluated through adsorption isotherm studies, which were performed using the non-adjusted initial pH of the paraquat solutions (pH ≈ 5.8). However, the use of RHA-BEA and magnetic RHA-BEA at this pH provided sufficient paraquat adsorption and complied with the typical pH range for most natural waters (Pukcothanung et al., 2018).

Close

Fig. 8

(a) Effect of pH on the amount of paraquat adsorbed on RHAS-BEA and magnetic RHAS-BEA (error bars indicate the standard deviation of duplicate tests using the initial concentration of paraquat of 100 mg/L, contact time = 24 h), and (b) pH_pzc of the RHAS-BEA and magnetic RHAS-BEA.

Export to PPT

At pH ≈ pH_pzc, the surface charge becomes zero; however, but greater q_e values than those in the lower pH region. Hence, interactions other than the aforementioned electrostatic forces may contribute to the adsorption mechanism. The paraquat molecule is smaller than size of the NaBEA zeolite cavity (Rongchapo et al., 2013), allowing it to be easily adsorbed via pore diffusion. Thus, cation exchange between cationic paraquat and Na⁺ at the aluminum sites is thought to be the main mechanism by which adsorption occurs (Osakoo et al., 2017; Rongchapo et al., 2018) in this region. Several studies have reported the adsorption mechanism of paraquat using similar adsorbents. Hydrogen bonding also occurred during the adsorption of paraquat on mesoporous SiO₂-modified bentonite (Rasaie et al., 2021), and was found between the H atom of paraquat and the O atom centered around Al in the FAU zeolite (Keawkumay et al., 2019). Using X-ray photoelectron spectroscopy, Osakoo et al. showed that the binding energy between the C and N atoms of paraquat and the centered O atom in an FAU zeolite/mesoporous SiO₂ composite (Osakoo et al., 2017). The aromatic rings and methyl groups of paraquat also interact with the FAU zeolite framework (Rongchapo et al., 2018). These interactions supported the possible mechanism of action of paraquat on RHAS-BEA and magnetic RHAS-BEA synthesized by DGC.

3.4 Separability of the adsorbents from aqueous solutions

Fig. 9a illustrates the separation of the adsorbent particles from aqueous solutions with and without Nd magnets. In the absence of a magnet, both RHAS-BEA and magnetic RHAS-BEA samples slowly settled to the bottom of the containers, and most of the suspension remained in the liquid phase. When the Nd magnet was attached to the side of the containers, the magnetic sample quickly became attracted the magnetic field, resulting in a much clearer liquid phase after 1 min, whereas the non-magnetic sample did not react. Additionally, we checked the change in the concentration of SS in the liquid phase over 60 min (Fig. 9b). We found that the concentration of SS slowly decreased in the absence of a magnetic field for both samples. In the absence of a magnetic field, the magnetic sample sedimented more slowly because of its smaller particle size. However, when a magnetic field was applied, a slow decrease in SS was still observed for the RHAS-BEA sample because it was not being attracted by the magnet. In contrast, the SS sharply decreased in the magnetic RHAS-BEA sample to a concentration lower than that of the other samples, meeting the SS standard value for rivers set by the Japan Environmental Quality Standards for Protecting the Living Environment. These results indicate that the magnetic separability of the magnetic RHAS-BEA sample was not affected by the co-existing α-Fe₂O₃, that is, the weakly magnetic phase. Therefore, the inclusion of Fe₃O₄ in the RHA-BEA zeolite greatly facilitated the collection of adsorbents after the removal of paraquat.

Close

Fig. 9

(a) Separation of RHAS-BEA and magnetic RHAS-BEA suspension in aqueous solutions by settling the adsorbents without magnet and attaching to Nd magnet for 1 min, and (b) separability of the adsorbents by evaluating SS in the adsorbent suspension (100 mg/L) placing with Nd magnet (solid circle and square markers for RHAS-BEA and magnetic RHAS-BEA, respectively) and without Nd magnet (transparent circle and square markers for RHAS-BEA and magnetic RHAS-BEA, respectively).

Export to PPT

3.5 Reusability and ions leached from the adsorbents during adsorption

In addition to adsorption performance and magnetic separability, economic feasibility was considered (i.e., in terms of reusability). This is an important aspect when considering the applicability of a water treatment approach. In most studies, the adsorption of paraquat on adsorbents is mainly ion-exchange, which is inhibited in acidic media or at high salinity. Several studies have reported that paraquat can be desorbed from adsorbents using salts (Ait Sidhoum et al., 2013) or acidic solutions (Dehghani et al., 2021; Jodeh et al., 2019; Kouchakinejad et al., 2022). In this study, we focused on waste handling after adsorbent regeneration. Hence, thermal treatment of the spent adsorbents was considered to remove the adsorbed paraquat. Pure paraquat thermally decomposes at 300 °C (Haley, 1979), thus the adsorbents were recalcined in air at 450 °C after paraquat adsorption. The q_e values for RHAS-BEA and magnetic RHAS-BEA obtained from the first recalcination cycle decreased by approximately 40% and 30%, respectively, compared to those of the pristine samples. However, these samples remained unchanged for three additional re-adsorption cycles (Fig. 10). The lower adsorption performance after the first cycle can be attributed to the incomplete removal of adsorbed paraquat under the applied recalcination conditions. The thermal decomposition residue may remain the least accessible adsorption site, hindering the access of paraquat molecules in the second cycle and cycles thereafter. Regardless of their initial use, the adsorbents could be reused for at least four cycles without showing degradation in their performance. Given the advantages of using recycled agricultural waste, high conversion rate of the DGC method, and magnetic separability, the adsorption capacities of the reused RHAS-BEA and magnetic RHAS-BEA (37.08 ± 7.95 and 35.21 ± 4.04 mg/g, respectively) are still acceptable because they are maintained at levels comparable to those of several types of fresh adsorbents, as shown in Table 1. The adsorption performance of the non-magnetic and magnetic RHAS-BEA samples among the four reuses emphasizes that the inclusion of Fe₃O₄ particles in the BEA zeolite does not negatively impact adsorption.

Close

Fig. 10

Paraquat adsorption capacity of RHAS-BEA and magnetic RHAS-BEA after five consecutive adsorption/regeneration cycles. The tests were performed in duplicate for each cycle using the initial concentration of paraquat (500 mg/L), contact time = 24 h, and regeneration temperature = 450 °C (6 h).

Export to PPT

For the ion-leaching test, Si was found to slightly leach from the adsorbents while Al was not detected in the solutions during the 3-h shaking time at pH ≈ 6. The detected level of Si for RHAS and magnetic RHAS adsorbent in the solutions were 1 mg/L and 0.4 mg/L, respectively, which were approximately 0.1 wt% and 0.04 wt% of the pristine RHAS and magnetic RHAS adsorbents, respectively. The negligible leaching of Si indicates that the developed adsorbents are stable under the tested conditions. For the magnetic RHAS adsorbent, Fe leaching was not detected during the adsorption test, which also supports the idea that the magnetic particles were incorporated inside the zeolite particles.

3.6 Paraquat co-adsorption with blue dye

The results of paraquat adsorption using the developed adsorbents in the presence of CB are shown in Fig. 11. Our adsorbents showed a paraquat removal efficiency of more than 99% in the presence of CB, a representative blue dye commonly added to commercial herbicides to avoid accidental poisoning (Nur et al., 2005; Rongchapo et al., 2013). The presence of blue dye at both concentrations did not suppress the adsorption of paraquat, whereas the adsorbents only adsorbed a small amount of CB since its molecular size is much larger than paraquat. This implies that the developed adsorbent enables the selective removal of paraquat from commercial herbicides.

Close

Fig. 11

Paraquat (PQ) removal efficiency of RHAS-BEA and magnetic RHAS-BEA tested in single (solid-fill bars) and binary (pattern-fill bars) systems.

Export to PPT

3.7 Paraquat co-adsorption with diquat

Herbicides with similar structures to paraquat might also be adsorbed using the developed adsorbents through similar mechanism. Diquat, not only the similarity in structure (Fig. 12), is one of the commonly used herbicides together with paraquat (Ronnen et al., 1995) and has negative effect to human health (Uematsu et al., 2021). Thus, both paraquat and diquat should be removed from water. The results in Fig. 12a and b show the decreasing in UV absorption intensity of the solutions after single adsorption tests to below the detection limit of the spectrophotometer for both paraquat and diquat using RHAS-BEA and magnetic RHAS-BEA. This indicates that diquat is also adsorbed onto the developed adsorbents. In binary adsorption test result (Fig. 12c), no strong UV absorption bands were detected for both paraquat and diquat after adsorption. Therefore, the RHAS-BEA and magnetic RHAS-BEA are also applicable for removal of those two herbicides.

Close

Fig. 12

UV absorption spectra of (a) paraquat and (b) diquat before and after single adsorption tests, and (c) paraquat and diquat before and after binary adsorption test.

Export to PPT

3.8 Adsorption performance comparison with other types of adsorbents

Direct comparison of the adsorption performance among various types of adsorbents is difficult owing to variations in the experimental conditions and nature of a respective adsorbent. The maximum adsorption amounts obtained from the Langmuir model for the adsorbents in this study and from previous reports are summarized in Table 1. Our adsorbents showed higher q_m values than those of several types of adsorbents, such as mesoporous SiO₂, clays, graphene oxide, and H-LTL zeolites, but lower adsorption performance than the low-SiO₂ zeolites reported in several papers, including the BEA zeolite prepared by conventional hydrothermal synthesis using RHAS (Rongchapo et al., 2013). However, the advantages of our adsorbents are as follows: they can be easily collected through magnetic separation; they are created through the use of agricultural waste–derived ingredients; and they can also be reused. Furthermore, our adsorbents exhibited a paraquat-removal efficiency of more than 99% in the presence of blue dye. Hence, RHAS-BEA and magnetic RHAS-BEA could be used as alternative candidates for paraquat removal from contaminated waters.