Green sulfidated iron oxide nanocomposites for efficient removal of Malachite Green and Rhodamine B from aqueous solution

With the acceleration of global urbanization and industrialization, water pollution has become a major threat to human survival. Dyes are one of the most common pollutants in wastewater. The efficient removal of the dyes from wastewater has been a fascinating topic of debate for the scientific community. Malachite Green (MG) and Rhodamine B (RhB) are common dyes in wastewater. MG, a triphenylmethane based dye, has been extensively applied in the food industry, aquaculture, dyeing and textile industries. The MG dye is highly cytotoxic to human cells and has tumour-enhancing property. The presence of MG in the water may lead to teratogenic, carcinogenic and mutagenic effects on human cells (Kan et al. 2015). RhB, a red dye, has been widely used in the paper, plastics, leather, dyeing, textile and printing industries. RhB is toxic and carcinogenic because it contains four N-ethyl groups on either side of a xanthene ring (Mohammadi et al. 2010). Until now, various methods have been employed to remove dyes from water, for example photocatalysis, adsorption, membrane separation, etc. (Mohammadi et al. 2010; Xiao et al. 2020). Among these methods, the adsorption method is regarded as a superior method for treating wastewater considering its straightforward operation, nontoxicity and lower cost (Xiao et al. 2020).

Nanomaterials as adsorbents have attracted considerable attraction because of their larger specific surface areas, well-defined shapes, and manageable size. Iron nanoparticles (including zero-valent iron and iron oxide nanoparticles) have been extensively applied in environmental remediation with their distinctive physical and chemical characteristics, great specific surface areas, high reaction activity and abundant raw material sources (Handojo et al. 2020; Bounab et al. 2021; Golabiazar et al. 2021). However, due to the slow removal rate of some pollutants, narrow range of active pH value (Bounab et al. 2021), easy agglomeration and other problems, researchers focus more on modification of iron nanoparticles to overcome their shortcomings. At present, the main modification measures include green synthetic iron nanoparticles, sulfide-modified iron nanoparticles, supported iron nanoparticles, and stabilized-modified iron nanoparticles (Zhang et al. 2018, 2020; Liang et al. 2021). Although there are numerous modification methods at present, sulfide modification materials are a breakthrough because of their environmental friendliness and high reactivity. Sulfide modification refers to the physical and chemical modification of iron nanoparticles by using sulfidation reagents to form iron sulfides (Fe_1+αS or Fe_1-αS) on the surface of iron nanoparticles (Kim et al. 2011). Several studies have shown that modification of iron nanoparticles, such as pits formation and roughness changes, can lead to significant increases in removal efficiency (Wu et al. 2022). Currently, there are two sulfidation modification methods, namely one-step and two-step synthesized sulfidated iron nanoparticles. The one-step method is synthesized through reaction of ferric iron with borohydride and a sulfidation reagent. The two-step method is synthesized by adding a sulfidation reagent to the iron nanoparticles suspension after iron nanoparticles formation (Han & Yan 2016). The sulfidation reagents used include sodium dithionite, thiosulfate and sulfides (e.g., sodium sulfide). Different reagents have no significant effect on the reactivity of these sulfidated iron nanoparticles (Han & Yan 2016). Considering that the one-step method is simpler, this paper adopts the one-step method to the synthesis of sulfidated iron nanoparticles. Sulfides and sodium dithionite may release toxic H₂S gas during the synthesis process; sodium thiosulfate is used as the sulfidation reagent in the next experiment.

However, current studies have investigated that sulfidated iron nanoparticles had some limitations in the practical application, such as the complex synthesis schemes, the use of dangerous reducing reagents and the introduction of boron impurities during the synthesis process (Liang et al. 2021). Therefore, it is necessary to develop a more environmentally and economically friendly method. In recent years, some researchers have discovered the use of reducing substances in plant extracts to reduce iron ions instead of the dangerous reducing agents needed for chemical methods, and then irons were oxidized and hydrolysed to form iron oxide nanoparticles (Zhang et al. 2018). There are many reports in the literature on the preparation of iron nanocomposites from plant extracts. Wang et al. used green tea and eucalyptus leaf extracts to produce iron nanoparticles for the elimination of nitrate from aqueous solution (Wang et al. 2014). Iron nanoparticles for the removal of Cr (VI) were synthesized using eucalyptus leaf extracts (Jin et al. 2018), etc. But no literature has been reported on the one-step green synthesis of sulfidated iron oxide nanocomposites using plant extracts as a reducing agent and for the removal of pollutants.

Herein, it is anticipated that some performance improvements can be achieved through a combined approach of sulfide modified and green synthesis iron nanoparticles. So, an attempt was made to fabricate sulfidated iron oxide nanocomposites (S-Fe NCs) using the leaf extract of Viburnum odoratissimum as a reducing agent. The Viburnum odoratissimum leaf contains a high amount of the chemical components that include flavonoids, sesquiterpenes, lignans, coumarin glycosides, etc. (Ma et al. 2014). In this study, we described a simple, economical and green approach for the preparation of S-Fe NCs, which had the inherent advantages of having a wider application potential.

The S-Fe NCs were investigated for their ability to remove the dyes MG and RhB. A range of characterization methods were used to analyze the morphologies and properties of S-Fe NCs. The effects of adsorbent dosage, pH, ionic strength, concentration and temperature on the adsorption capacity for MG and RhB were studied. Meanwhile, the adsorption kinetics, adsorption thermodynamics and adsorption mechanism were also studied.

In this study, the Viburnum odoratissimum leaves were collected from Zhengzhou University, Zhengzhou, China. The chemical reagents (i.e. Iron chloride hexahydrate (FeCl₃·6H₂O), Sodium thiosulfate pentahydrate (Na₂S₂O₃·5H₂O), Malachite Green (MG) and Rhodamine B (RhB)) are of analytical grade.

The Viburnum odoratissimum leaves were rinsed three times in deionized water to remove superficial dirt, dried at room temperature, and cut into smaller pieces afterwards. The small leaves were mixed with deionized water (100 g L⁻¹), and the temperature maintained at 353 K for 120 min. The leaves extract was filtrated and was stored in an airtight vessel at 277 K for any future use.

Sodium thiosulfate (0.0465 g) was put into 150 mL of the extract of Viburnum odoratissimum leaf under vigorous stirring at room temperature. After 20 min, 50 mL of FeCl₃ solution (0.075 mol L⁻¹) was put dropwise into the leaves extract under continued vigorous stirring for a further 1 h. The colour of the solution changed from yellow to black, revealing the production of the solid material. The solid was separated by centrifugation at 10,000 rpm for 3 min and rinsed three times in deionized water. At the end, the product was vacuum freeze dried for 12 h. Figure 1 illustrates the plausible synthesis mechanism; first, disproportionation of sodium thiosulfate produces elemental sulfur and sulfite (Han & Yan 2016). Second, the catechin hydroxyl group forms a complex with Fe³⁺ and reduces Fe³⁺ into Fe⁰, followed by oxidation and hydrolysis to form Fe oxide nanoparticles, along with the elemental sulfur being reduced in the presence of Fe⁰ to sulfide (Han & Yan 2016). Finally, Fe oxide and FeS grow together to form S-Fe NCs. Biomolecules as stabilizers and capping agents formed by components of the plant extracts are wrapped around the S-Fe NCs surface.

Scanning electron microscopy (SEM, ZEISS Gemini 300) and transmission electron microscopy (TEM, Tecnai G2 TF20) were applied to observe structural morphology of the S-Fe NCs. Energy dispersive X-ray spectroscopy (EDS, OXFORD Xplore) was undertaken to observe elemental distribution. X-ray diffraction (XRD, Bruker D8 Advance) images were acquired to identify the structural composition. Fourier transform infrared spectroscopy (FTIR) (PE-1710 spectrophotometer, USA) was employed to identify sample surface functional groups. X-ray photoelectron spectroscopy (XPS, Thermo ESCALAB 250Xi) was employed by monochromatic Al Kα excitation (k= 0.834 nm) to determine elemental oxidation state. Surface area analyser (NOVA1200e) was used to obtain the pore size distribution and Brunauer–Emmett–Teller (BET) specific surface area. The point of zero charge (pH_PZC) of S-Fe NCs was determined and the value of pH_PZC was 3.85.

The morphology and the size of S-Fe NCs were examined using SEM and TEM as illustrated in Figure 2, indicating the successful synthesis of nanocomposites. Many small particles on the surface of S-Fe NCs are observed in Figure 2(a), which may be attributed to the precipitate formation of FeS (Kim et al. 2011). After adsorption with MG solution (Figure 2(b)) and RhB solution (Figure 2(c)), a slight increase in the size of S-Fe NCs is noted and some aggregations appear on the adsorbent surface, indicating that MG or RhB are being adsorbed on the surface. The large amount of adsorption capacity resulted in a significant increase in the particle size of the S-Fe NCs, which is similar to previous reports (Wang et al. 2014). TEM images for the S-Fe NCs are shown in Figure 2(d) and 2(e). The S-Fe NCs have diameters of ∼100 nm. The average size of S-Fe NCs obtained from TEM is close agreement with average sizes measured by SEM. As shown in Figure 2(e), the lack of lattice stripes in S-Fe NCs suggested that the S-Fe NCs is amorphous.

To better investigate the composition of the S-Fe NCs, the elemental information of S-Fe NCs was measured using EDS as shown in Figure S1. Fe, C, N, O and S Kα EDS mapping indicated that the five elements are uniformly distributed throughout the nanocomposites. The composition of the surface of S-Fe NCs was mainly C (63.66 wt%), S (0.36 wt%), N (2.59 wt%), O (20.34 wt%), Fe (13.05 wt%) (Figure S1 g). The C and O signals were mainly derived from C, O-containing molecules in plant extracts, which can be ascribed to the surface of S-Fe NCs being covered with biomolecules of Viburnum odoratissimum leaf extracts.

The XRD pattern of the S-Fe NCs is shown in Figure S2. The lack of obvious diffraction peaks manifested that the material was mainly amorphous. The wide shoulder peak at around 2θ = 20° was determined for the adsorbed organic materials from plant extract as a capping or stabilizing agent (Jin et al. 2018). Furthermore, a small characteristic peak occurring at 2θ = 44° corresponding to zero-valent iron (Fe⁰) was also noticed (Jin et al. 2018). However, it was reported that the Fe⁰ was hardly involved in the removal process and the removal mechanism was based on an adsorption process (Wang et al. 2014). This result also proved that S-Fe NCs contain primarily iron oxides and iron sulfides.

FTIR spectra were applied to gain the information of surface functional groups (Figure S3). As shown in Figure S3, the characteristic peak located at 3,393 cm⁻¹ corresponded to the O-H stretching vibration. The characteristic peaks at 1,620 and 1,400 cm⁻¹ were ascribed to C = C and carboxyl C = O stretching vibration, respectively (Zhang et al. 2020). The peak at 1,079 cm⁻¹ was assigned to C = O characteristic peak of the carboxylic acid (Zhang et al. 2018). The peak at 1,025 cm⁻¹ was caused by C-O-C and O-H characteristic peak. A weak peak in the 816 cm⁻¹ was noted and can be attributed to C-H out-of-plane bending mode (Zhang et al. 2020). The characteristic peak at 534 cm⁻¹ refers to the Fe-O stretches of iron oxides, demonstrating the formation of S-Fe NCs. These results clearly showed that some oxidized polyphenols capped the surface of S-Fe NCs and improved stability properties of S-Fe NCs.

The S-Fe NCs XPS measurements were performed to determine the surface element composition and their valence states in Figure 3 and Table S1. Figure 3(a) shows the XPS spectrum of S-Fe NCs in the Fe 2p, O 1 s, C 1 s, N 1 s and S 2p region. Detailed XPS spectra of Fe 2p for the S-Fe NCs are shown in Figure 3(b). The three characteristic peaks were at 709.6 eV, 711.32 eV and 714.93 eV, which were attributed to FeS, Fe(III) oxides and the trace amounts of FeCl₃, respectively (Han et al. 2019). In addition, no significant peak at around 707 eV could be found, suggesting that Fe⁰ was negligible in S-Fe NCs. This conclusion was consistent with the result of XRD analysis. The XPS spectrum of O 1 s was decomposed in Figure 3(c), where two compositions were contributed by OH⁻ (531.49 eV) and C-O (532.86 eV). The C1 s spectrum of S-Fe NCs showed three peaks at 284.81 eV, 286.34 eV and 288.33 eV (Figure 3(d)), which could be attributed to C = C, C-OH and carboxyl/epoxy C = O, respectively. The N 1 s XPS spectrum displayed two peaks at 400.39 (NH₂-O) and 402.24 eV (NH₂-Fe), which were considered to be N elements in the biomolecules. The S 2p spectrum of S-Fe NCs showed two peaks at 163.41 eV and 168.6 eV (Figure 3(f)), which were attributed to polysulfide (S_n²⁻) and surface bound SO₄²⁻, respectively (Cao et al. 2017). The presence of SO₄²⁻ might be due to the oxidation of sodium thiosulfate by oxygen in the air. The presence of O 1 s, C 1 s and N 1 s suggested that some biomolecules were covered on the surface of the S-Fe NCs by chemical bonding. These results provided favourable evidence that the S-Fe NCs are mainly consisting of iron oxides, biomolecules and iron sulfides, further confirming that the S-Fe NCs had a good adsorption capacity.

The specific surface area of S-Fe NCs was determined by BET analysis. As shown in Figure S4, a type-IV isotherm with a hysteresis loop was noted, suggesting the presence of mesoporous matrix in the S-Fe NCs. The specific surface area of the S-Fe NCs was determined to be 14.05 m² g⁻¹. The pore volume and average pore diameter of the S-Fe NCs were 0.101 cm³ g⁻¹ and 3.41 nm, respectively.

As shown in Figure S5, the adsorption capacity of S-Fe NCs for MG and RhB decreased with an increase in adsorbent dose. This result was ascribed to unsaturated active sites. Conversely, the removal efficiency of S-Fe NCs was observed to increase from 21.3 to 97.4% for MG and 13.2 to 94.5% for RhB under identical conditions, respectively. This result was attributed to an increase in number of available active sites, thus achieving a higher removal efficiency.

The solution pH plays a crucial factor in the functional groups form of dye molecules and the surface charge of S-Fe NCs, thus influencing the adsorbent adsorption capacity. Based on the pK_a value of the dye molecules, the distribution curves of dye molecules at different pH values are made in Figure 4. As shown in Figure 4, the adsorption capacity of S-Fe NCs for MG increased from 3.77 to 5.27 mmol g⁻¹ with the initial pH increasing from 3.15 to 9.6. This is because at low pH values, high concentrations of H⁺ ions may compete with cationic MG⁺ ions for limited sorption active sites on the S-Fe NCs. At pH <3.85 (pH_pzc of S-Fe NCs), the surface of the S-Fe NCs is positively charged. There is electrostatic repulsion between the S-Fe NCs and the cationic dyestuff MG, reducing the adsorption capacity. The adsorption performance of S-Fe-NC for RhB had the opposite effect compared to MG. The adsorption capacity of S-Fe NCs for RhB decreased gradually from 3.31 to 2.27 mmol g⁻¹ with the initial pH increasing from 2.38 to 9.7. RhB exists in both cationic and zwitterionic forms in water. At pH >3.7 (pK_a of RhB), RhB can transform the cationic form to the zwitterionic form due to deprotonation of the COOH groups in RhB, which brings about an electrostatic repulsion between RhB and the negatively charged S-Fe NCs (Mohammadi et al. 2010). Moreover, the zwitterionic form of RhB in water can gather a greater molecule aggregate and become unable to enter into the pore of S-Fe NCs, resulting in reduced adsorption capacity. Therefore, the adsorption capacity gradually decreased with increase in solution pH. Even so, S-Fe NCs exhibited a coincident high adsorption capacity in a broad range of pH, illustrating its application potential in various environments.

To simulate the complicated situation of real wastewater, the effect of salt concentration on the adsorption properties was researched. The effects of various concentrations of NaCl and CaCl₂ on the adsorption process were studied, and the results are shown in Figure S6. With an increase in CaCl₂ concentration from 0.01 to 0.1 mol L⁻¹, the adsorption capacities of S-Fe NCs for MG and RhB decreased from 4.06 to 3.18 mmol g⁻¹ and 2.7 to 2.56 mmol g⁻¹, respectively. However, the adsorption capacities of MG and RhB on S-Fe NCs had no obvious change with the NaCl concentration increase, which is similar to previous reports (Liu et al. 2020). The difference in the adsorption process was due to their different ionic charge and size. Ca²⁺ had one more positive charge than Na⁺, which made it a stronger ionic effect. This confirms the existence of electrostatic interactions between the two dyes and S-Fe NCs.

Figure 5 exhibits the effect of reaction time on the adsorption of MG and RhB onto S-Fe NCs in aqueous solution. The adsorbed rates of MG and RhB on the S-Fe NCs were very fast in the first 30 min. The adsorption capacity of MG was 78, 80 and 80% of the equilibrium adsorption capacity, with the initial concentrations 0.882, 1.37 and 2.055 mmol L⁻¹, respectively. Likewise, the RhB absorption reached 70, 62 and 71% of equilibrium adsorption capacity in the first 30 min, respectively. From Figure 5, it was clear that the adsorption process of both dyes onto S-Fe NCs was through two main stages: the fast initial stage and the slower stage leading to equilibrium. This represents a typical process of adsorption, as available active sites decrease as time goes on and the adsorption reaches the maximum values (Vigneshwaran et al. 2021). Another phenomenon could also be observed from the diagram: that the adsorption capacity increased with increasing initial concentration, indicating that a larger concentration gradient could be used to overcome the resistance to mass transfer during adsorption process.

The fitted curves and corresponding parameters are shown in Figure 5 and Table 1, respectively. According to the low SSE and high R² from Table 1, it was obviously observed that the Elovich model best described the adsorption process of MG and RhB on S-Fe NCs. This result suggested that the adsorption process of MG and RhB on S-Fe NCs was not homogeneous. In addition, the pseudo-second-order model and double-constant rate equation model indicated higher R² values and lower SSE in Table 1, which illuminated the applicability of the pseudo-second-order model and double-constant rate equation model. The application of the pseudo-second-order model illustrated chemisorption as a rate-controlling step in adsorption. The value of R² for the pseudo-first-order model (0.854 ≤ R² ≤ 0.972) was lower than the pseudo-second-order model (0.934 ≤ R² ≤ 0.988), and the calculated adsorption capacity also differed significantly from the experimental value. Moreover, higher values of SSE (0.167 ≤ SSE ≤ 1.350) further indicated that the pseudo-first-order model was not quite satisfying.

The intraparticle diffusion model was also employed to investigate the adsorption behaviour of MG and RhB on S-Fe NCs. As shown in Figure 5(c) and 5(d), the adsorption process of S-Fe NCs to dyes can be divided into two stages. The first stage was the diffusion of MG or RhB molecules from the aqueous phase to the outer surface of S-Fe NCs. In the second stage, the adsorption rate gradually decreased until the dynamic equilibrium of adsorption was reached. The value of k_t1 was greater than that of k_t2, and the boundary layer-related constant C₂ was larger than C₁ in Table 1. The results suggested that mass transfer of MG and RhB molecules predominated in initial adsorption process. In addition, the fitted curves did not cross the original point, suggesting that intraparticle diffusion was not the sole control step.

The graph of Bt value versus time t forms a line with a slope of B. The fact that the straight line did not cross the original point indicated that the velocity control step during adsorption procedure was not an intra-diffusion, but a membrane diffusion. The values of B calculated for MG were 0.01671, 0.01906 and 0.01676 min⁻¹. For RhB, the values of B were 0.02646, 0.02673 and 0.01399 min⁻¹.

In this study, the average radius of S-Fe NCs is 50 nm. The values of D_eff for MG were 4.24 × 10⁻¹⁴, 4.83 × 10⁻¹⁴ and 4.25 × 10⁻¹⁴ cm² s⁻¹, respectively. Likewise, the values of D_eff for RhB were 6.71 × 10⁻¹⁴, 6.78 × 10⁻¹⁴ and 3.55 × 10⁻¹⁴ cm² s⁻¹, respectively. The value of D_eff is less than 10⁻¹¹ cm² s⁻¹, suggesting that intraparticle diffusion is not the sole controlling step in the adsorption procedure of S-Fe NCs on MG and RhB. It is controlled by a combination of membrane diffusion and intraparticle diffusion.

Figure 6 shows the effect of the concentrations of MG or RhB on the adsorption process at three temperatures. From the obtained results, it was seen that the adsorption capacity of MG onto S-Fe NCs increased from 0.66 to 4.44 mmol g⁻¹ with the initial concentrations increasing from 0.137 to 2.055 mmol L⁻¹ at 303 K. The increase in concentration meant a large number of the dye molecules competing for a limited number of active sites, thereby bringing about an increase in the concentration gradient drive. Moreover, the adsorption capacity of MG onto S-Fe NCs was observed to increase from 4.22 to 4.69 mmol g⁻¹ with the temperatures increasing from 293 to 313 K, suggesting that the adsorption process was endothermic. The adsorption of RhB on S-Fe NCs had the same trend. The adsorption capacity of MG was significantly greater than that of RhB. This was due to the larger molecular structure of RhB compared to MG, which posed a steric hindrance between RhB and the adsorbent.

The isotherm parameters are summarized in Table 2. Based on the high R² and low SSE, for the adsorption of MG on S-Fe NCs, the isotherm models' fit was in the order Koble-Corrigan model > Redlich-Peterson model > Temkin model > Dubinin-Radushkevich model > Langmuir model > Freundlich model. The parameter n of Koble-Corrigan model was between 0 and 1, suggesting that the adsorption process was complex. The parameter g of Redlich-Peterson model was closer to 1, indicating that the adsorption process of MG molecules had the behaviour of the Langmuir model. The Temkin model also fitted the experimental data well, indicating that S-Fe NCs had a uniform distribution of binding energy for MG adsorption. As shown in Table 2, The values of E were 6.45, 6.89 and 6.86 kJ mol⁻¹ at 293, 303 and 313 K, respectively, suggesting that the MG adsorption on S-Fe NCs was mainly physisorption. The Langmuir model could also better describe S-Fe NCs for MG adsorption. The calculated maximum adsorption capacity of S-Fe NCs for MG was 4.09, 4.31 and 4.73 mmol g⁻¹ at 293, 303 and 313 K, respectively.

According to the low SSE and high R² from Table 2, Koble-Corrigan, Redlich-Peterson, Langmuir, Temkin and Dubinin-Radushkevich models could describe the adsorption of RhB onto S-Fe NCs. The maximum uptake of RhB on S-Fe NCs determined by Langmuir model was 2.53, 2.88 and 3.25 mmol g⁻¹ at 293, 303 and 313 K, respectively.

The maximum adsorption capacity of S-Fe NCs for MG and RhB was compared with different materials and the results are shown in Table S2. Compared to other materials, S-Fe NCs showed a significant adsorption capacity for MG and RhB, which was primarily because of the great specific surface area of the S-Fe NCs and the presence of many reactive functional groups such as carboxylic and hydroxyl groups on the surface of S-Fe NCs. Thus, the S-Fe NCs will be a green and excellent adsorbent for the purification of MG and RhB from contaminated water.

The stability of S-Fe NCs was evaluated and the results are shown in Figure S7. It was found that the adsorption capacity for MG and RhB remained consistent at 4 and 2.4 mmol g⁻¹, suggesting that the S-Fe NCs were stable over a longer time period. The high stability of S-Fe NCs may be associated with the presence of biomolecules and sulfide-modified iron nanoparticles (Zhang et al. 2018).

The FTIR spectra of S-Fe NCs before and after adsorption of MG and RhB were displayed to analyse the adsorption mechanism of MG and RhB onto S-Fe NCs (Figure 7). After adsorption of MG, two new characteristic peaks were noted: 1,173 cm⁻¹ (C-N stretching vibration of aromatic compounds) and 726 cm⁻¹ (benzene ring). This result indicated that the MG molecules were adsorbed to the surface of the S-Fe NCs. In addition, the FTIR spectra of S-Fe NCs after adsorption of MG showed that the peak at 1,079 cm⁻¹ decreased and peak at 1,400 cm⁻¹ shifted to a lower wavelength, due to electrostatic interaction between the -COO⁻ group of S-Fe NCs and the positively charged nitrogen centre of MG dye (Zhang et al. 2018). At the same time, hydrogen bonds were formed between the -COOH of S-Fe NCs and the free electron pair N of the amidogen of the MG. The peaks at 1,620 cm⁻¹ and 1,400 cm⁻¹ shifted to a lower wavelength after MG adsorption, which may be ascribed to the π-π stacking interaction between the MG molecule and the aromatic rings of S-Fe NCs. In addition, the nitrogen atom in the MG molecule has coordination interactions with Fe on the S-Fe NCs (Huo & Yan 2012). Figure S6 showed that the MG adsorption capacity of S-Fe NCs was less affected with the increase of NaCl concentration, indicating that the inner-sphere complex was formed during the adsorption of MG by S-Fe NCs (Gao et al. 2020). In summary, the adsorption of MG on S-Fe NCs can be described (Figure 8) by the electrostatic interaction, the hydrogen bonding, the π-π stacking interactions and inner-sphere surface complexation.

As for RhB adsorption onto S-Fe NCs, the adsorption mechanism could also be described (Figure 8) by the electrostatic interaction, the hydrogen bonding, the π-π stacking interactions and surface complexation. In addition, the wavelength shift from 1,079 cm⁻¹ (-COO⁻ group) to 1,077 cm⁻¹ was noted, suggesting the Fe³⁺ of S-Fe NCs edge sites can intercalate with the COO⁻ group of RhB to form a cation bridge (Liu et al. 2011).

S-Fe NCs was successfully synthesized via a rapid green method and was applied to the adsorption process of MG and RhB in water. Combining SEM, EDS, XRD, FTIR and XPS analysis, it is clear that the S-Fe NCs is mainly composed of iron oxides, iron sulfides and biomolecules. The S-Fe NCs exhibited high-efficiency adsorption performance. The adsorption mechanisms of MG and RhB on S-Fe NCs can be described by the electrostatic interaction, the hydrogen bonding, the π-π stacking interactions, surface complexation and cation bridging. Moreover, the S-Fe NCs exhibited good stability, which might be attributed to biosynthesis and sulfidated modification. These results confirmed that the S-Fe NCs has great potential as an adsorbent for dyes remediation. This research also provides a new perspective for the applications of the remediation of other contaminants.

This work was supported by the Henan Science and Technology Department in China (No. 162300410016)

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