Microwave-Induced Deep Oxidation of Brilliant Green Using Carbon Nanotube-Supported Bismuth Ferrite

Abstract

Microwave-induced oxidation has emerged as an effective approach for water purification. In this study, bismuth ferrite-supporting carbon nanotubes with strong microwave absorption and magnetism were successfully fabricated for the degradation of brilliant green. The reactivity of bismuth ferrite in microwave fields and the role of carbon nanotubes was revealed by systematic characterization methods. Our results demonstrated that the addition of bismuth ferrite in microwave-induced system can enhance the ability of microwave-induced absorption and further induce the degradation and mineralization of brilliant green within 10 min, significantly surpassing conventional heating methods. The brilliant green decomposition by bismuth ferrite in microwave-induced process is a heterogeneous process. Its excellent performance achieved by active species-trap experiments can be attributed to microwave-induced holes. Overall, this study presented a promising material for microwave-induced elimination of brilliant green and other dyes in aqueous media, which can provide the basis for the environmental application of microwave radiation to water purification and wastewater treatment.

1. Introduction

Major classes of modern synthetic dyes exhibit strong structural and color stability and are produced and utilized substantially in the textile industry, with approximately 700,000 tons of dye produced per year [1,2]. Brilliant green (BG), as a representative dye, has the characteristics of wide application, large usage, and high toxicity, and is considered to be a priority pollutant [3], posing great threat to the aquatic environment and human health. It must be specially treated before disposing off even at low concentrations owing to its relatively higher toxicity. Various mature and new technologies have been utilized for BG treatment; nevertheless, each possesses distinct application scenarios based on BG concentration, ambient conditions, and reaction temperature [4]. Effective and efficient methods for remediating dye-contaminated waters still need to be explored.

Advanced oxidation processes (AOPs), such as ozonation [5], Fenton-like [6], photocatalysis [7], and microwave-induced (MW-induced) reactions [8], have been demonstrated effective in elimination of BG from wastewater. Compared with other AOPs, MW-induced reactions [9] and microwave-assisted photochemical processes [10] are a form of ultra-high-frequency electromagnetic waves, characterized by energy efficiency, high efficacy, safety, and non-harmfulness [11]. They are able to significantly accelerate the rotation of polar molecules, accumulate a significant amount of energy in a brief period of time, and subsequently diminish the chemical bond, accelerate the reaction rate, and reduce the activation energy [12]. MWs are able to efficiently degrade BG due to the rapid and selective heating induced by the microwave radiation [13].

Previous studies have demonstrated that azo dye Direct Black BN (DB BN) was removed by the adsorption on MgFe₂O₄-SiC and further microwave-induced catalytic oxidation [14]. After the finished adsorption equilibrium, the degradation of DB BN was slower than that without adsorption. Furthermore, BG can be degraded by UV irradiation assisted with microwaves in the presence of ZnO, but a long reaction time hinders the application of this technology [15]. So far, the high polarization of water molecules could plausibly reduce the effectiveness of these MW-induced processes for the degradation of dyes in wastewater, and the performance of MW-induce catalysts needs further improvement [8,16,17]. Nanosized perovskite-type ferrite has been developed as a reusable heterogeneous MW-induced catalyst for the selective reduction of nitroaromatic compounds [18]. Bismuth ferrite (BFO), a typical perovskite-type ferrite, has been used as a heterogeneous Fenton catalyst [19], visible-light catalyst [20] and magnetic material [21], which is thought to be chemically active in the MW field. Our previous studies showed that secondary calcined BaFeO exhibited high removal efficiency and total organic carbon (TOC) removal of BG in an MW-induced catalyst process [8]. However, nano-BaFeO samples cannot convert magnetic energy into thermal energy under an MW field. Carbon nanotubes (CNTs), as a metallic conductor or semiconductor, combined with silicon or metallic oxide, are thought to form heterojunctions [22,23] which improve reactivity. Meanwhile, CNTs have a high dielectric loss tangent among microwave absorbing materials, implying that they have greater adaptability for transforming electromagnetic energy into heat energy [24]. Previous studies showed that MFe₂O₄ (M=Co, Cu, Mn, etc.) composites, such as CoFe₂O₄, used with modified CNTs had better microwave adsorption capability and higher MW-induced catalytic activity than pure MFe₂O₄ [25,26,27]. However, the MW-induced reaction with bismuth ferrite-modified CNTs acting as the MW absorption material used for the oxidative degradation of dyes is seldom reported, and the mechanism is also unclear.

The aim of this study is to develop an innovative MW catalyst for the thorough degradation of BG. Carbon nanotubes (CNTs) are regularly employed as supports for catalysis because of their unique structural features, chemical properties, and outstanding regeneration abilities. Furthermore, bismuth ferrite exhibits both high catalytic activity and remarkable microwave absorption qualities. Therefore, CNT-embedded magnetic bismuth ferrite (BFNTs) was fabricated as magnetic catalysts, and their absorption and thermal conversion capacities in the microwave field were investigated in detail and compared with traditional BFO. The principal active species involved in the degradation of BG during the microwave-induced process were investigated through quenching experiments. The intermediates and products resulting from the degradation of BG were identified using liquid chromatography–mass spectrometry (LC-MS) and gas chromatography–mass spectrometry (GC-MS). Furthermore, various characterization techniques facilitated the proposal of potential degradation pathways and mechanisms.

2. Results and Discussion

2.1. Characterization of the BFNT Catalyst

The X-ray diffractometer (XRD) patterns of BFNT composites with increasing carbon nanotube loadings are shown in Figure 1. All five samples’ loadings with CNTs exhibited well-defined diffraction peaks corresponding to the (012), (104), (110), (006), (202), (024), (122), (116), (018), (214), (208), (220), and (131) crystallographic planes of BiFeO₃ (JCPDS File No. 20-169), respectively, confirming successful preservation of the host matrix crystallinity, which was in accordance with an earlier study [28]. Notably, the addition of CNTs (≤20 wt%) maintained comparable peak intensities (FWHM <0.35°), indicating negligible structural distortion of the bismuth ferrite. However, at the loading of 40% CNTs, the diffraction peak intensity was reduced significantly, suggesting that the bismuth ferrite particles associated with the CNTs became smaller, which implied possible interfacial strain induced by the CNTs-BFO heterojunction formation. Additionally, the absence of secondary phases confirms effective CNT dispersion without compromising phase purity, consistent with previous reports on analogous systems.

Transmission electron microscopy (TEM) analysis revealed significant morphological evolution with CNT loadings. Pure BFO exhibited regular particles averaging 200 nm (Figure 2A), while CNT-loaded BFNT composites showed comparative size reduction proportional to the CMTs’ mass ratio (Figure 2B–D). At 10% CNT loading, BFO aggregates (80–120 nm) anchored on the surface of the CNTs (Figure 2B). When the ratio of CNTs increased to 20%, the CNTs appeared to be securely associated with the BFO particles (Figure 2C), suggesting optimal interfacial contact in this condition. This configuration facilitated magnetic phase stabilization, consistent with previous reports [29]. However, excessive loading of 40% CNTs induced partial BFO detachment from CNT surfaces despite maintaining the particle sizes (Figure 2D). Increasing the mass ratio of CNTs from 20% to 40% did not appear to result in a substantial change in the average particle size of bismuth ferrite, which ranged from 30 nm to 50 nm, suggesting interfacial saturation. Furthermore, high-resolution TEM analysis identified preferential exposure of {104} facets with d-spacing of 0.27 nm (Figure 2E), but the {012} face could also be partially exposed (Figure 2F), revealing anisotropic growth tendencies in the optimized composite.

The Fourier transform infrared spectroscopy (FT-IR) analysis also exhibited critical structural and interfacial characteristics of the BFNT composite. As shown in Figure 3A, the absorption vibration peak of metal oxides is below 800 cm⁻¹, while the most prominent bands of ferrite are in the region of 800–2000 cm⁻¹ [30]. The intensities of the two adsorption bands at 442.6 and 556.5 cm ⁻¹ were manifested in the infrared spectrum of 10-BFNTs, which were attributed to (O-Fe-O) bending and (Fe-O) stretching modes in the perovskite-type sheets of BiFeO₃, respectively [9,31]. Additionally, a small band at 478.7 cm⁻¹ ascribed to the symmetric stretching of (Bi-O) was observed [32]. Compared to titania-coated CNTs prepared by a hydrothermal method [14], hydroxyl stretching vibrations (3000–3600 cm‒¹) were also attenuated, indicated enhanced surface hydrophobicity of BFNTs, which should be beneficial to the adsorption of aqueous phase organic contaminants. Diagnostic carbon skeleton vibrations at 1387 cm⁻¹ and 1450 cm⁻¹ were assigned to the (C=C) stretching mode, confirming CNT integration [33,34]. The new and intensified peaks at 835 and 1387 cm⁻¹ in BG-loaded samples directly demonstrated the dye adsorption through π-π interactions. After microwave induced decomposition of BG, the intensity of these peaks diminished, confirming effective BG decomposition and interfacial regeneration of adsorption sites.

The X-ray photoelectron spectroscopy (XPS) patterns elucidated the changes in the surface elemental composition associated with MV-induced degradation of BG (Figure 3B). Different response intensities of Bi4f might result from different amounts of adventitious carbon. The ratio of Bi/Fe changed slightly from an initial ratio of 1.14 to 1.28 after MW radiation, and the value after regeneration returned to 1.16. This indicated that there is no obvious change in elemental composition. A previous study has reported the higher than anticipated (1:1) ratio could result from possible Bi segregation to the surface of BFO [35]. Notably, compared to the control system where BFNTs are absent (Table S1), the amount of Fe²⁺/Fe³⁺ released into solution after MW radiation is negligible, consistent with the premise that the MW-induced reaction involving BFNTs is a heterogeneous process. Consistent with the presence of Fe³⁺ in bismuth ferrite, there is a satellite peak on the curves of the synthesized BFNTs before and after dye adsorption [36] (Figure 3C), yet this peak disappears after MW radiation. The reason can be explained by the partial reduction to Fe²⁺ after MW radiation. The Bi 4f binding energy shifted minimally from 159.1 eV before to 158.6 eV after MW radiation. Both values are ascribed to Bi³⁺ and are significantly below the Bi³⁺ to Bi⁰ transition with a typical value of 3.0 ± 0.1 eV, [37]. The abnormally low binding energy shift (0.5 eV) may be due to a deficiency in oxygen during the MW-induced reaction, either in the perovskite lattice structure or in the Bi₂O₂ layer [38]. After regeneration, Fe²⁺ is completely retransformed into Fe(III), and the peak of Bi 4f centers at 159.1 eV again, which demonstrates material stability. It is further demonstrated by the negligible leaching of Fe²⁺/Fe³⁺ into solution owing to the existence of a stable Bi-Fe bond. Figure 3D shows a small C 1s peak at 288.7 eV, assigned to the -COO group formed on the surface of CNTs during sol–gel combustion process [39]. However, the peaks associated with O 1s did not change throughout the process, indicating oxygen’s non-redox participation. These findings collectively validate the heterogeneous catalytic nature of MW-driven degradation, where surface-mediated electron transfer predominates over homogeneous ion leaching mechanisms.

BiFeO₃ nanoparticles exhibit a high saturated magnetization of approximately 13 emu g⁻¹, which was not substantially reduced as the CNT mass ratio increased from 10% (wt/wt) and 40% (wt/wt) (Figure S1). As typical soft magnetic materials, both BFO and BFNTs demonstrate low coercive fields, characteristic of soft magnetic materials, making them highly suitable for magnetic recovery applications in nanomaterial systems. The UV–Vis spectral analysis of BG dye associated with MW radiation revealed a distinct degradation pattern (Figure S2). The aqueous BG solution exhibited a maximum absorption wavelength (λmax) at 624 nm. During the MW-induced degradation process, λ_max remained stable for the first 8 min, after which a decrease in BG concentration coincided with a hypsochromic shift of λ_max to 614.5 nm. The small hypsochromic shift indicates that the cleavage of ethyl groups in the BG structure, followed by rapid degradation of the conjugated chromophore system.

2.2. Degradation of BG

No degradation of BG was observed from exposure to MW radiation alone. However, pristine bismuth ferrite, a poor adsorbent of BG, modified with CNTs achieved significant enhancement, particularly at 40% CNT loading (wt/wt), adsorbing 95% of BG within 2 h (Figure 4). In the presence of BFNTs with CNT mass ratios from 5% to 20%, the initial aqueous BG concentration decreased from ~ 15 mg L⁻¹ to ~ 13 mg L⁻¹ over a period of 2 hr. When the mass ratio of CNT was 40%, the concentration of aqueous phase BG diminished by 95% after equilibration. Predictably, the adsorption efficacy of BFNTs for BG increased with increasing BET surface area, showing a large increase in BG uptake from water, with a 20% to 30% (wt/wt) loading of CNTs on BFO. The BET surface area of 20-BFNTs is 2.7 times greater than that of 5-BFNTs (

Table 1. BET surface areas of bismuth ferrite (BFO) with increasing loadings (0–40% wt/wt) of CNTs.

CNT Mass (%)	Surface Area (m2 g−1)
0	5.9 ± 0.1
5	6.3 ± 0.1
10	12.1 ± 0.1
20	16.3 ± 0.2
40	82.9 ± 0.3

). A slight reduction in the initial rate of degradation of BG was observed for the BFNTs compared to BFO and is attributed to the adsorption/desorption effect of CNTs. This implies that when dealing with BFNTs, BG adsorbed on CNTs may not be degraded efficiently as that on bismuth ferrite, though the difference appears small. It is apparent that the degradation rate of BG is faster and more extensive when exposed to MW radiation compared to conventional heating, indicating that MW radiation has a specific effect on the degradation of BG beyond just heat. In addition to BG, other dye solutions at various concentrations, e.g., 20 mg L⁻¹ of methyl orange, 100 mg L⁻¹ of methyl violet and malachite green, and 20 mg L⁻¹ of alizarin yellow R, were also found to be rapidly degraded (> 94%) by BFNTs under the same reaction conditions in 10 min (Figure S3). The results show that not only triarylmethane family dye (BG, methyl violet and malachite green) but also azo family dyes (methyl orange, alizarin yellow R) were amenable to extensive degradation in the presence of BFNTs, suggesting that BFNTs are broadly effective for organic dye wastewater remediation. In addition, as shown in Figure S4, the removal efficiency of BG slightly declines while remaining at a high level (>95%) after six tests. The decline is mainly attributed to the coverage of the reactive sites on the catalyst by degradation intermediates of BG. To test this hypothesis, we washed the catalyst with ethanol, used them again, and then found that the removal rate of BG was recovered by the seventh time. Furthermore, acceptable Fe leaching (0.007 mg/L) after seven cycles can be observed based on inductively coupled plasma–mass spectrometry (ICP-MS). This low ion dissolution may be ascribed to the fact that BFO is confined by porous carbon, which can greatly reduce ion dissolution and improve the long-term stability of the catalyst. We also investigated the application of the reaction system in different actual water bodies and found that it exhibited a good removal rate in tap water, lake water, and river water (Figure S5), highlighting its potential for application in real environments.

The TOC value also decreased rapidly during the MW-induced process, with a 78% reduction observed within 6 min of irradiation (Figure 5A). This TOC decline was synchronized with brilliant green (BG) decomposition, indicating near-instantaneous mineralization of intermediates—a stark contrast to the delayed TOC removal reported in NiOx-mediated triarylmethane dye degradation systems [40,41]. The pH increased from 5.04 to 5.77 during treatment, consistent with oxidative mineralization of acidic decomposition byproducts rather than mere adsorptive removal. The adsorption recovery MW (ARMW) experiment results demonstrated effective BG elimination (>92%) from BFNT surfaces (Figure S6). The BG adsorbed on BFNTs could be removed effectively in the ARMW process. This indicates that BG could be degraded either by in situ degradation or by redispersion into aqueous solution in MW field. In either form, bismuth ferrite plays an irreplaceable part in the process.

The MW-induced degradation of BG using BFNTs was systematically compared across three mass ratios of CNTs (5%, 10%, and 20%) against conventional heating methods. The MW-induced degradation rate of BG increased, with a positive correlation between degradation rate and CNT loading, achieving 85%, 92%, and 97% BG removal for 5-, 10-, and 20-BFNTs, respectively, within 10 min (Figure 5B). In contrast, conventional heating exhibited an inverse trend at lower CNT ratios (<20%), with 10-BFNTs showing marginally faster degradation (78% removal) than 20-BFNTs (72%), though statistical analysis (p > 0.05) confirmed no significant difference. The above results confirm the advantage of the MW-induced effect for dye decomposition by BFNTs due to a special heating effect of microwave radiation.

2.3. The Mechanism of the MW-Induced Reaction

Microwave irradiation demonstrated superior catalytic performance over conventional heating in accelerating dye degradation kinetics on carbon nanotube-modified bismuth ferrite (BFNT), achieving 92% removal of model wastewater contaminants within 10 min. To elucidate the underlying mechanisms, we systematically evaluated the microwave absorption capacities of BFO and BFNTs through electromagnetic parameter analysis (Table S2). The dielectric constant (ε′), which indicates how effectively a material can store microwave energy, reflects the capacity for polarization. Conversely, the dielectric loss factor (ε″) denotes a material’s capability to convert the stored energy into heat. The loss tangent offers an insight into how effectively a material is penetrated by an electric field and its efficiency in terms of energy dissipation as heat. Both BFO and BFNT exhibit high values of ε’. At a standard frequency of 2.45 GHz, corresponding to domestic microwave ovens, the dielectric loss tangents are approximately 0.039 rad for BFO and 0.043 rad for BFNTs. This suggests that compared to BFO, BFNTs show a relatively higher dielectric loss factor (ε″) value, which demonstrates that BFNTs have a greater adaptability for transforming electromagnetic energy into heat energy, and that BFNTs present better heating efficiency than BFO in the microwave field. This is consistent with the behavior of titania-coated CNTs [14], and provides new evidence that CNTs could be added to enhance microwave absorption and improve heating efficiency for a variety of materials, as was shown herein for BFO (Figure 6).

In MW-induced reactions, there are several kinds of unique processes and special effects, including superheating, hot spots, and selective heating. It is believed that MW heating could provide a temperature of 13-26 K above the normal solvent boiling point, which is referred to as overheating or superheating [42]. Previously, we demonstrated that superheating also occurs in microwave-assisted photocatalytic reactions over titania-coated carbon nanotubes [43]; CNTs contributed to the increased heating rate. In the present work, the rate of heating of bismuth ferrite-loaded CNTs was not strongly affected by different CNT mass ratios (Figure 6). This divergence stems from bismuth ferrite’s superior microwave absorption, which dominates the thermal response. Interestingly, CNT loading-dependent superheating emerged in BFNT systems. With an increase in the CNT loading, the suspension temperature increased from 378 K (BFO only) to 389 K at a 20% CNT loading. Microwaves could produce “hot spots” on CNTs that possess high microwave absorbing ability. As the loading of CNTs increased to 20% (wt/wt), a greater mass of CNTs capable of forming “hot spots” would plausibly lead to the observed increase in suspension temperature. This could explain why the 20-BFNTs perform better than the 10-BFNTs in the MW-induced process evaluated, while the opposite was observed for the conventional heating process. This contrasts with conventional heating, where 10%-BFNTs outperformed 20%-BFNTs (78% vs. 72% removal, respectively), highlighting MW-specific interfacial polarization effects. The observed thermal behavior aligns with recent findings on single-atom selective heating in zeolite systems, suggesting universal microwave–catalyst interaction principles [44].

Hydroxyl radical attack is thought to be an important mechanism in MW-induced reactions over CNTs [16]. However, hydroxyl radicals could not be detected by salicylic acid, which was used to probe the presence of hydroxyl radicals in our study. This reveals that ⋅OH plays a negligible role in this system. The MW-induced reaction described here was also conducted in the presence of a hydroxyl radical scavenger of tert-butanol, superoxide radical scavenger of benzoquinone, and a hole scavenger of oxalic acid and sodium oxalate (Figure 7A, Figure S7 and S8). Notably, when the hole scavenger was added, the BG degradation rates reduced dramatically for both BFO and BFNTs. This reveals that in the MW-induced process, microwave radiation can produce holes over BFO and BFNTs. These holes are produced by a thermoelectric effect of the perovskite oxide structure. The inhibitory effect of the hole scavenger when added to BFNTs (the degradation efficiency decreased from 86% to 36% within 10 min of radiation) was significantly higher than that observed with BFO (degradation efficiency decreased from 89% to 56%), suggesting that the CNTs skeleton or the junction between CNTs and bismuth ferrite enhanced the hole-doping effect of the MW radiation. This may be due to the conductivity of CNTs manifesting enhanced thermoelectric properties [45]. It should be noted that compared to BFO, the support of CNTs could transfer more electromagnetic energy into heat energy when exposed to microwave radiation, causing a larger temperature increase for BFNTs, which would also enhance the thermoelectric effect. Alternatively, Clark et al. proposed that BiFeO₃ and metals with work functions of 4.6–4.9 eV could form n-type Schottky barriers, and the Schottky barrier heights are 0.8–0.9 eV over metal [46]. Considering that the work functions of random multiple-walled CNTs sidewalls/tips have been measured to be 4.6-4.9 eV, Schottky barriers between the carbon nanotubes and BFO are also possibly formed. An enhanced hole–electron separation in microwave field is proposed, whereby the electron flows from bismuth ferrite into CNTs, leaving holes on the bismuth ferrite. This could account for the decreased inhibitory effect of BFNTs on BG degradation.

Superoxide radicals exhibited negligible involvement in the dye degradation mechanism, mirroring the minimal contribution of hydroxyl radicals as evidenced by scavenger assays (Figure 7B). Intriguingly, this inhibitory effect of the superoxide radical scavenger on the reaction over BFNTs was higher than that over BFO (Figure S9), indicating that the effect of carbon nanotubes on superoxide radicals is similar to their effect on the formation of holes [47]. The proposed mechanism of active radicals’ production in MW radiation is shown in Scheme S1. Since holes were the most important active species, we expected that oxygen would perform the critical function of eliminating electrons. Evidence for this effect was obtained by comparing reactivity in the presence of a nitrogen, oxygen, or compressed air atmosphere (Figure S9). Oxygen played a critical role in sustaining hole activity by scavenging electrons, as demonstrated by comparative degradation efficiencies. This is further supported by recent studies on Schottky junction-enhanced transfer in CNT–perovskite composites [25].

2.4. Intermediates and Degradation Pathways

The intermediates analysis in this process is shown in Tables S3 and S4, and the proposed degradation pathway is shown in Scheme S2. The results show limited numbers and concentrations+ of intermediates, consistent with the rapid mineralization evidenced by the 85% TOC removal within 30 min. Deethylation initiates the overall degradation reaction, which proceeds at a much faster rate than that achieved by BFO [8]. Notably, the products from the decomposition of the p-π conjugated structure are different from that of typical triphenylmethane in microwave-assisted photocatalytic reactions, implying a different benzene ring removal process from the classical description in hydroxyl radical-induced reactions [48]. The ring-open products are detected by GC-MS, indicating mineralization instead of simple adsorption. In addition, we further verified the biological toxicity of the treated and untreated brilliant green solutions using Escherichia coli. Obviously, the density of Escherichia coli. in the treated system is much higher than that in the untreated sample (Figure S10), emphasizing the lower biological toxicity and safety after treatment.

3. Materials and Methods

3.1. Materials

Brilliant green (BG, laser-grade) was purchased from AppliChem Company (Hangzhou, China). Dye samples representing different structural classes were obtained from Sinopharm Chemical Reagent Co., Ltd (Shanghai, China). Multiwalled carbon nanotubes (MWCNTs, >95 wt%) with a mean external diameter of 10–20 nm, were purchased from Alpha Nano Technology Co. Ltd. (Chengdu, China). All other chemicals, except those used for chromatogram analysis, were of analytical reagent grade. Deionized water with a resistivity of 1.8 × 10⁷ Ω cm was obtained using a Milli-Q water ion-exchange system.

3.2. Preparation of BFNTs

TFNTs were prepared as a reddish-brown precipitate using a modified citrate acid sol–gel combustion method [49]. Accordingly, 0.01 mol of iron nitrate and bismuth nitrate were dissolved in 25 mL 2-methoxyethanol; then, 25 µL of 0.1 mol L⁻¹ HNO₃, certain amounts of CNTs, 0.01 mol citric acid, and 12 mL of ethylene glycol which acted as a dispersant, were added subsequently. The mixture was stirred at room temperature for 2 h, followed by stirring at 333 K for 1 h to form a solid. Then, the solid was heated at 363 K for 10 h evaporating 2-methoxyethanol to further form nitrate–citrate complex gel. The generated residue was ground and heated at 473 K for another 30 min to remove organic matter and residual NO^3-. Ultimately, the powder was annealed in air at 773 K for 2 h. The prepared BFNTs with different loadings of CNTs are hereafter referenced as x-BFNTs, denoting a BFNT with x% (wt/wt) CNTs (based on the mass of bismuth ferrite) added during the preparation process. A short-time microwave-assisted Fenton method was used here to regenerate BFNTs after the reaction [49]. Additionally, 0.5 mL of 1 mol L⁻¹ H₂O₂ was added into the suspension of used BFNTs with 20 mg L⁻¹ BG. This regeneration process lasted for only 1 min.

3.3. Characterization Methods

The composition of the crystalline phases in BFNTs was examined with an X-ray diffractometer (XRD) that utilizes Cu-Kα radiation (Model, Shimadzu LabX XRD-6000). The transmission electron microscopy graphs of the samples were obtained using a JEM-200CX (JEOL Company) instrument. Using a vibrating sample magnetometer (VSM, Lakeshore Cryotronic 7307-9309), the samples’ magnetism was evaluated at room temperature between -10 and 10 kOe. The FT-IR (Nicolet co., ltd, USA, NEXUS870) spectra were acquired in transmission mode from 400 to 4000 cm⁻¹. A vector network analyzer (VNA, Agilent, E8363C) operating in the 0.5–18 GHz frequency range was used to evaluate the electromagnetic parameters of the dielectric constant (ε′) and dielectric loss factor (ε″). A mixture of paraffin and BFNTs and BFO was compressed into a steel mold to create the VNA specimens. Using a PHI5000 Versa Probe electron spectrometer from ULVAC-PHI and 300 W Al-Kα radiation, X-ray photoelectron spectroscopy (XPS) data were obtained. 6.7 × 10⁻⁸ Pa was the estimated base pressure. The binding energies were associated with the C1s peak at 284.6 eV, attributed to adventitious hydrocarbons or C-C bonds in carbon nanotubes [50].

3.4. MW-Induced Degradation of BG

In the MW oven (330 mm × 330 mm × 200 mm, 900 W maximal MW output, made in Beijing Xianghu Science and Technology Development Co., Ltd., Beijing, China), there is a flat-bottom flask (250 mL), with which a 600 mm long water reflux condenser is connected through a communication pipe (Figure S11). There is an aperture the top of the oven for the communication pipe to pass through. Note that an aluminum tube fixed in the aperture is used to eliminate MW leaking. The limit on the safe stray leakage of MW power density was kept below 0.5 mW/cm² at 2450 MHz, measured at 200 mm from the aperture [40]. All experiments were performed using a Wolff bottle with a connection to a thermocouple and a reflux condenser. After continuous stirring for 2 h in the dark at room temperature to achieve adsorption/desorption equilibrium, 50 mL of a 20 mg L⁻¹ BG solution was set in the microwave device, containing 1 g L⁻¹ of BFNTs. At predetermined times (0, 0.5, 1, 2, 4, 6, 8, and 10 min), a 1 mL of sample was withdrawn and quickly separated magnetically to remove the catalyst prior to analysis. As a reference, a conventional heating-based degradation experiment was carried out simultaneously. Before adding BFO, a BG solution was pre-heated and kept the thermostatic water bath at 100 °C with boiling water. All other conditions were identical to those in the MW-induced reaction experiment.

3.5. Analysis Methods

The analysis of BG concentration was conducted using a UV spectrophotometer (V2550, Japan) set at a wavelength of 625 nm. The pH of the solution was measured utilizing a pH meter (PHS-2C, China). The released aqueous Fe ions were determined by atomic absorption spectrometry (AAS, Hitachi Z-8100) after recovering the catalyst using the magnetic method. The experiment in a suspension without BG addition was set as a control group. To characterize the bismuth ferrite particle after the degradation experiment, the solid sample was washed repeatedly and centrifuged at 11,000 rpm for 15 min, then dried in the atmosphere. The identification of intermediates and degradation products was conducted via LC-MS-MS (Thermo-Finnigan LCQ Advantage) and GC-MS (Agilent). The LC-MS-MS was equipped with an electrospray ionization (ESI) source and a Beta Basic-18 column (150 mm × 2.1 mm). Methanol–water was used as mobile phase, at a flow rate of 0.2 mL min⁻¹, and a linear gradient was set as follows (A—water; B—methanol; v: v): t = 0, A = 95, B = 5; t = 10, A = 50, B = 50; t = 30, A = 10, B = 90; t = 40, A = 10, B = 90; t = 42, A = 95, B = 5; t = 50, A = 95, B = 5. After the pretreatment process (Text S1), the extract (10 μL) was analyzed by GC-MS (via a Finnigan trace gas chromatography interfaced with a Polaris Q ion trap mass spectrometer) equipped with an electron impact ion (EI) source and a DB-5 fused-silica capillary column (30 m × 0.25 mm).

3.6. Active Species Identification

Two distinct methods were employed to identify the active species responsible for BG degradation. In the first approach, salicylic acid (SA, hydroxybenzoic acid) was used as both a molecular probe and as a hydroxyl radical scavenger to detect hydroxyl radicals. Specially, 0.5 g of BFNTs were combined with 50 mL of 0.5 mmol L⁻¹ SA in a glass container and subjected to microwave irradiation (10 min). The mixture was then separated by centrifugation, followed by acidification of the supernatant with 0.7 mL 37% HCl. The solid phases were subsequently extracted thrice using dichloromethane (50 mL per extraction). After that, both the solutions received the addition of 0.7 mL of 37% hydrochloric acid, and solid particles were extracted three times by 50 mL dichloromethane each time. All the extracted liquids were evaporated in a rotary evaporator at 323 K, and the residual material was dissolved in 5 mL of methanol for HPLC analysis.

In the second method, oxalic acid and sodium oxalate were used as holes (h⁺) to scavenge to investigate the effect of MW-induced holes [51,52]. Prior to BG adsorption, oxalic acid was added into the BG solution, and the solution’s pH stabilization was ensured through the addition of sodium oxalate. Benzoquinone and tert-butanol were used as the superoxide radical scavenger (⋅O₂^-) and hydroxyl radical scavenger (⋅OH), respectively. The final concentrations of oxalic acid and sodium oxalate in the BG solution were 0.02 g L⁻¹ and 0.5 g L⁻¹, respectively, and the final concentrations of benzoquinone or tert-butanol were 5 mol L⁻¹. The other experimental conditions were the same as those in the MW-induced experiment section.

Both SA and dihydroxybenzoic acid (DHBA, representing 2,3-DHBA and 2,5-DHBA) in the aqueous solution were determined directly by HPLC (Agilent 1200, USA) coupled with a fluorescent detector at an excitation wavelength of 237 nm and an emission wavelength of 405 nm, using a Zorbax SB-C18 column (5 μm, 4.6 mm × 250 mm), at a flow rate of 0.7 mL min⁻¹, with 80% of deionized water containing 0.03 mol L⁻¹ citric acid and 0.03 mol L⁻¹ acetic acid buffer (pH = 3.6) and 20% of methanol as the mobile phase.

4. Conclusions

In this work, we used a modified citrate acid sol–gel combustion method to synthesize BFNTs in microwave-induced reactions for BG degradation. Synthetic BFNTs have a superior adsorption capacity for BG compared to BFO. As the proportion of CNTs was increased from 5% to 20%, the amount of BG degraded by microwave-induced degradation increased from ~80% to ~95%. However, when the proportion of CNTs in the BFNTs was 40%, about 90% of the BG was removed by adsorption. The TOC removal by 20-BFNTs in microwave-induced degradation of BG processes was ~90%, and the pH showed a slight increase in the reactions. About ~65% BG was removed by 20-BFNTs at 20 min under conventional heating induced degradation processes, demonstrating that microwave induction promoted the degradation of BG. A comparison of the ε″ of BFOs and BFNTs further demonstrates that BFNTs have greater adaptability for transforming electromagnetic energy into heat energy and exhibit higher heating efficiency in the microwave field. The results of capture experiments revealed that the holes were the main active species in the reaction of BFNTs with microwave-induced heating. Deethylation initiated the overall degradation reaction, and ring-open products detected by GC-MS indicated mineralization instead of simple adsorption in this system. The novel material, namely BFNTs, demonstrated high microwave adsorption and efficient degradation of dyes, which provides a new solution for the rapid degradation of dyes.