Ultrasonic role to activate persulfate/chlorite with foamed zero-valent-iron: Sonochemical applications and induced mechanisms

Highlights

• •The novel foamed-ZVI activating persulfate/chlorite by ultrasonic was applied.
• •The real-time and online continuous detections tracked radicals transformation.
• •The EPR spectra with different systems in different mixtures were detected.
• •The free or non-free radicals processes induced by US were explored.
• •The dynamics on cavitation and the US-field simulation were presented.

Abstract

The novel system, consisting of composite oxidants (persulfate/chlorite, S₂O₈²⁻/ClO₂⁻) and stationary phase activator (zero-valent-iron foam, Fe⁰_f) driven by ultrasonic (US) field, was applied to treat the triphenylmethane derivative effectively even at low temperature (≈ 289 K). By comparisons of sub-systems, the US roles to S₂O₈²⁻, ClO₂⁻, and Fe⁰_f were seriatim analyzed. US made the reaction order of multi-component system tend to within 1 (leading to de-order reaction), and widened pH activating range of the Fe⁰_f by sonicate-polishing during the process of ClO₂⁻ co-activating S₂O₈²⁻. US and Fe⁰_f were affected by fluid eddy on activating S₂O₈²⁻/ClO₂⁻. The Fe⁰_f had slight effect on the temperature of US bubble-water interface but the addition of ClO₂⁻ lowered it. The partitioning capacity of the above US reactive zone increased during the reaction. US and ClO₂⁻ could enrich the kinds of degradation intermediates. The contributions of free radicals (ClO_x-based radicals, sulfate radicals (SO₄•⁻), and hydroxyl radicals (•OH)) and non-free radicals (ClO₂, and O = Fe^IV/V from ionic Fe under “-O-O-” of S₂O₈²⁻ and cyclic adjustment reaction of ClO₂⁻) processes by sonochemical induction were equally important by corresponding detection means. Especially, real-time and online high-resolution mass spectrum by self-developing further confirmed the chain transfers of different free radicals due to US role. The findings expanded the application of sono-persulfate-based systems and improved understanding on activation mechanism.

1. Introduction

A series of trace triphenylmethane derivatives detected in aquatic environment with low concentration (<μM level) as emerging contaminants have drawn the attention [1], therein no lack of triphenylmethane crystal-violet (tpm_CV). Besides as the dye, the tpm_CV (Fig. S1a), could be used not only in clinical treatment, but also as antimicrobial and antiparasitic agents in aquaculture to treat and prevent fungal and protozoal infections [2]. However, the United States, the European Union, China, and others have not approved the use of tpm_CV in aquaculture due to it with the possible carcinogen, mutagen and teratogen effects [1], [2], [3]. Even so, the use of the prohibited and persistent tpm_CV would be common occurance, leading to its frequent detection in aquatic environment. Specially, the bio-concentration of organisms made tpm_CV convert into leuco crystal-violet (_Ltpm_CV, Fig. S1b) (more toxic and more difficult to be metabolized) [1], [3], which might be seen as the persistent pollutants.

$$H_{2}O\stackrel{\left)\right)\left)\right)}{\to }\bullet OH+\bullet H$$ (1)

The alternative US-based advanced oxidation processes (AOPs), as the promising and friendly treatment technologies with physicochemical processes of cavitation (gas-phase reaction zone effective temperature, ≈ 5200 K, and high pressure, in the range of hundreds of bars), pyrolysis (breaking chemical bonds, and yielding hydroxyl radicals (Eq. (1)) [4]), supercritical water oxidation (depth and temperature of the liquid-phase reaction zone, ≈ 0.2 μm and ≈ 1900 K, respectively), and violent turbulence [5], [6]; meanwhile, especially as the sono-catalysis technologies to activate oxidizing agents to produce active species to form combined system [5], [6], [7], have attracted particular concern for environmental remediation. Herein, as a kind of oxidizing agent, persulfate (involving peroxymonosulfate (HSO₅⁻, PMS) or peroxydisulfate (S₂O₈²⁻, PDS)) with direct oxidizing or producing SO₄•⁻ (E⁰ (SO₄•⁻/SO₄²⁻) = 2.60 – 3.10 V_NHE) by activation to improve the oxidizing capacity [8] via physicochemical means (e.g. alkali, Eq. (2); heat, Eq. (3) [9]; photo-; electro-; microwave; and US, Eq. (4)), metal-free catalysis [10], metal-based materials (e.g. zero-valence metal, Eq. (5)) [11], and other methods (e.g. ozone, superoxide, and plasma), could be used to treat environmental contaminants as well.

$${2S}_2{O}_8^{2-}+{2H}_{2}O\stackrel{{\mathit{OH}}^{-}}{\to }S{O}_4^{\bullet -}+O_2^{\bullet -}+3S{O}_4^{2-}+{4H}^{+}$$ (2)

$$S_2{O}_8^{2-}\stackrel{\mathrm{h}\mathrm{e}\mathrm{a}\mathrm{t}}{\to }2S{O}_4^{\bullet -}\mathrm{k}=1.0\times {10}^{-7}{s}^{-1}\left(298K\right)\mathrm{o}\mathrm{r}5.7\times {10}^{-5}{s}^{-1}\left(343K\right)$$ (3)

$$S_2{O}_8^{2-}\stackrel{\mathit{hv}/))))}{\to }2S{O}_4^{\bullet -}$$ (4)

$$\equiv {\mathrm{M}}^{n+}+S_2{O}_8^{2-}\to \equiv {\mathrm{M}}^{\left(n+1\right)+}+S{O}_4^{2-}+S{O}_4^{\bullet -}$$ (5)

Therefore, further works for sono-persulfate based AOPs were worth exploring to extend acoustic catalysis and persulfate activation bilaterally, seeming to become prospective treatment methods for the elimination of refractory organic contaminants above.

At present, the zero-valent-iron (ZVI, Fe⁰)-based materials in disperse mediums by modified techniques (pre-magnetized [12], sulfidated [13], loaded/embedded [14], micro/nano [15]) or in stationary phase ZVI (e.g. foamed-ZVI (Fe⁰_f) consisting of three-dimensional porous form (Fig. S1c) with properties [16], [17] of permeability, controlled pore size, foveolate structure, shape stability, and machinability) as the activated intermediary to connect the US and persulfate, have obtained great attention. Herein, the Fe⁰_f for oxidant activation was rarely reported, however, functional materials based on Fe⁰_f have drawn important attention, such as, electrode materials for hydrogen evolution reaction or oxygen evolution reaction (electrocatalytic water splitting) [17], which also had greater potential for environmental remediation. However, the surface passivation of ZVI has been a fatal problem affecting its activation to persulfate. Nevertheless, it’s just that introducing US could polish the above ZVI in activation process, and ease the burden of passivation to provide the fresh ZVI.

In addition, in order to promote corrosion of ZVI and regulate the activation cycle (e.g. ionic Fe cycle) under the action of US, the chlorite (ClO₂⁻, Cl(III)) could be introduced into the aforementioned process and combined with persulfate to form composite oxidants. Commonly, ClO₂⁻ with a higher quantum yield at 254 nm, 1.0–1.53 mol⋅Einstein^-1 [18], [19], comparing with H₂O₂ (0.5 mol⋅Einstein^-1) and persulfate (0.7 mol⋅Einstein^-1) [20], could be activated by physical fields including UV-light stimuli ([18], Eq. (6)), heat [21], and US, or chemical agents including acid (Eq. (7)) and PDS [22] to produce chlorine dioxide (ClO₂).

$$3\mathrm{C}\mathrm{l}{O}_2^{-}+H_{2}O\stackrel{\mathit{hv}}{\to }2Cl{O}_2+{\mathit{Cl}}^{-}+{2OH}^{-}+0.5O_2$$ (6)

$$5\mathrm{C}\mathrm{l}{O}_2^{-}+4H^{+}\to {\mathit{Cl}}^{-}+4Cl{O}_2+2H_{2}O$$ (7)

The ensuing surface passivation of ZVI (Cl role) may impede its further activation, however, introducing US, as a “bridge”, could not only activate composite oxidants, persulfate/chlorite (O_SC), to generate SO₄•⁻ and ClO₂ by co-catalysis (Eq. (8)) [23], but also polish the Fe⁰_f to avoid aforementioned Fe⁰_f passivation and to activate O_SC deeply (Eqs. (9)-(13)).

$$S_2{O}_8^{2-}+Cl{O}_2^{-}\stackrel{))))}{\to }S{O}_4^{\bullet -}+S{O}_4^{2-}+Cl{O}_2$$ (8)

$${\mathit{Fe}}_f^0+S_2{O}_8^{2-}\to {\mathit{Fe}}^{\mathit{II}}+2{\mathit{SO}}_4^{2-}$$ (9)

$${\mathit{Fe}}^{\mathit{II}}+S_2{O}_8^{2-}\to {\mathit{Fe}}_{(aq\&s)}^{\mathit{III}}+{\mathit{SO}}_4^{\bullet -}+{\mathit{SO}}_4^{2-}$$ (10)

$${\mathit{Fe}}^{\mathit{II}}+{\mathit{ClO}}_2\to {\mathit{Fe}}_{(aq\&s)}^{\mathit{III}}+{\mathit{ClO}}_2^{-}$$ (11)

$$4{\mathit{Fe}}^{\mathit{II}}+{\mathit{ClO}}_2^{-}+2H_{2}O\to {4Fe}_{(aq\&s)}^{\mathit{III}}+{\mathit{Cl}}^{-}+4{\mathit{OH}}^{-}$$ (12)

$$2{\mathit{Fe}}_{(aq)}^{\mathit{III}}+polished{\mathit{Fe}}_f^0\to {3Fe}^{\mathit{II}}$$ (13)

In this study, the novel treatment systems, namely, the combined system (US/Fe⁰_f-O_SC) based on co-catalysis involving acoustic catalysis and multiple activation of composite oxidants, and sub-systems, were established. The commonly apparent effects of US may have been mentioned, however, because of that US has always been seemed as just an assistant method in environmental remediation, the researches on its applied potential (e.g. couple function as a “bridge” of systems above by series connection) and feasible mechanism of US role (e.g. US role to transformations of generated active species) may be neglected.

This study was therefore designed to (1) evaluate applications of the combined- and sub- systems based on US from comparisons of performances, affecting parameters, and pollutant degradation paths; (2) reveal the detailed sonochemical roles in apparent remediation by comparisons of coupling systems; and (3) explore the roles of US-induction in remediation mechanism by comparisons of free radicals and non-free radicals processes.

2. Experiment

2.1. Operating

The operating processes including chemicals, self-developed US/catalytic/activated reactors (UCARs), etc. were presented in Section S1.

2.2. Analytical methods

The intermediates of tpm_CV and methyl phenyl sulfoxide (PMSO) were detected by gas chromatograph-mass spectrometer (GC/MS, Trace 1300 ISQ-QD, Thermo Fisher). The temperature programming (GC/MS) was: column temperature 323 K, hold time 3 min; 10 K·min⁻¹ up to 573 K, hold time 10 min. Before the above analysis, 50 mL subsample at a certain treating time (5, 15, 25, 35, and 45 min) was extracted with 15 mL dichloromethane, respectively. Afterward, the extracted solution by mixing was dehydrated by anhydrous sodium sulfate and concentrated to 2 mL. The ClO₂ and tpm_CV concentration were detected by UV–vis spectrum (TU-1810S, PGENERAL, China) at 360 nm [19] and 584 nm, respectively. The concentrations of “Fe” were determined by phenanthroline spectrophotometry at 510 nm (HJ/T 345–2007). The accurate acoustic-intensity entering the system was measured by the calorimetry method [24], presenting in Section S2 (Eq. S1). The compositions and micrographs of the Fe⁰_f, before and after treatment, were detected and examined by X-ray diffraction (XRD, HAOYUAN Instrument, DX-2700) and scanning electron microscopy (SEM, Zeiss, Germany), respectively. Before XRD and SEM, the Fe⁰_f samples taken out from reaction solution needed dehydration immediately by adding into absolute ethyl alcohol, then drying out with filter papers, finally vacuum drying. The anions of S₂O₈²⁻/ClO₂⁻ were detected by ion chromatography (IC, ThermoScientific, Aquion).

2.3. Spin trapping and EPR measurements

The US filed was applied in solution consisting of NaClO₂ (1–20 mM) and K₂S₂O₈ (ClO₂⁻: S₂O₈²⁻ = 1:5 in molar ratio) to conduct sonochemical experiments, then Fe⁰_f was added as needed. DMPO with equal proportion of O_SC was utilized to trap the generated active species above. In addition, dimethyl sulfoxide (DMSO, as needed) was added into water to form mixture for comparison.

The electron paramagnetic resonance (EPR) spectra were obtained by a Bruker A200 spectrometer. General instrument settings were as follows unless otherwise noted: center field 3490 G, modulation frequency 100 kHz, modulation amplitude 1 G, receiver gain 1.42 × 10⁴, time constant 20.5 ms, conversion time 40 ms, sweep time 41 s, frequency 9.82 GHz, and power 19.47 mW. The hyperfine splitting constants were measured by using Bruker Win-EPR SimFonia (Version 1.2).

2.4. Real-time and online continuous detection

In Scheme 1, the real-time and online continuous (ROC) detections of reaction mixtures were realized with a self-developed pneumatic nebulization (PN)-droplet spray ionization (DSI) (electrospray ionization (ESI) with CH₃OH as an assist as required) coupled to LTQ and LTQ-Orbitrap high resolution mass spectrum (HRMS) (Thermo Scientific) in positive or negative mode with a resolution of 60 000. Online MS/MS analysis of the captured intermediates could be performed during the reaction process. Data were acquired in full scan in the m/z range 50–300 and in MS/MS with a collision-induced dissociation energy of 25. The instrument was calibrated routinely with the provided calibration solution by Thermo Scientific to ensure mass accuracy within ± 3 ppm.

Scheme 1.

Schematic diagram of ROC detection device.

The self-developed PN-DSI as Scheme 1^① shown: the sealed glass device (7 mL) was equipped with mixture (5 mL) of S₂O₈²⁻ at 5 × 10^-6 M, ClO₂⁻ at 1 × 10^-6 M, and DMPO at 1 × 10^-4 M under US generator; then the liquid along the capillary tube entered into the site of triple valve by N₂ inflating; and the charged samples from the above liquid, suffering the action of PN (its description in Section S3) by another N₂ flow, were transmitted into HRMS.

2.5. Simulation on US reactor

The Acoustic-Piezoelectric Interaction of COMSOL Multiphysics, including pressure acoustics Eq. (14), was applied to simulate heterogeneous reaction process and visualize the action of US-field in the solution. The detailed simulation process were described in Section S4 (Eqs. S2-S25). The UCAR2 was simulated as shown in Fig. 1a-c and Fig. S2a-g.

$$\mathrm{\nabla }\bullet \left(-\frac{1}{{\rho }_c}\left(\mathrm{\nabla }{p}_t-{\mathit{q}}_d\right)\right)-\frac{k_{\mathit{eq}}^2{p}_t}{{\rho }_c}=Q_m$$ (14)

Fig. 1.

Simulation of acoustic pressure distribution and piezoelectric mechanics deformation at 28 kHz with Fe⁰_f: (a) sound pressure level, (b) vertical slice, (c) horizontal slice; and comparisons of O_SC (6 mM)-based systems and PDS (6 mM)-based systems: (d) removal efficiencies, (e) fitting kinetics (_{molar ratio}S₂O₈²⁻/ClO₂⁻ 5:1, f 28 kHz, I 30 W⋅L^-1, pH without adjustment (initial pH 6.5, near-neutral), Fe⁰_f 0.358 mM, tpm_CV 26.47 μM, 0 r⋅min⁻¹, and initial T 289 K).

3. Results and discussion

3.1. Comparisons of US-based systems

At low temperature (≈ initial 289 K), low pollutant concentration (10^-6 M, μM level), and near-neutral pH (without adjustment), the novel US/Fe⁰_f-O_SC presented good performance on removing tpm_CV. O_SC, consisting of S₂O₈²⁻ and ClO₂⁻ with molar ratio 5:1, was sourced from our previous study. From Fig. 1d, the treating efficiencies of tpm_CV by US/O_SC (6 mM)-based systems were greater than those by US/PDS (6 mM)-based systems, due to the degradation of tpm_CV in US/PDS-based systems presenting weak implementation on activation in later stage (e.g. after 15 min).

The exponential analysis of apparent dynamics could be utilized to further understand the removal process of tpm_CV (details in Section S5, Eqs. S26-S28). By fitting concentration functions of the systems mentioned above, the optimal fitting orders (non-integer, R²_Adj > 0.999, Fig. 1e) of O_SC, Fe⁰_f-O_SC, US-O_SC, and US/Fe⁰_f-O_SC systems were 1.35, 1.10, 0.90, and 0.75, respectively.

In order to further measure the synergistic effect among the above-mentioned systems (Fig. 1e) conveniently, the S and f_i (details in Section S6, Eqs. S29-S33) were calculated to evaluate interactions of each O_SC-based system with pseudo-first-order kinetics (R²_Adj > 0.990) as below (k_obs values (10^-3, min⁻¹) of O_SC, Fe⁰_f-O_SC, US-O_SC, and US/Fe⁰_f-O_SC systems: 46.82, 51.56, 57.22, and 85.64, respectively).

$$\mathrm{S}=\frac{{k_{\mathit{obs}}}_n}{\sum {k_{\mathit{obs}}}_i}$$ (15)

$$f_i=\frac{{k_{\mathit{obs}}}_n}{{k_{\mathit{obs}}}_{n-1}}$$ (16)

where S is the synergistic effect of the multicomponent system, f_i is the action of single i system, and k_obs represents pseudo-first-order kinetic constant unless otherwise indicated.

The S was 1.83 (>1), and the f_US and f_Fe0f in US/Fe⁰_f-O_SC (1.66, 1.50) > those in US-O_SC (1.22, 1.10), which indicated that: i) US/Fe⁰_f-O_SC had well synergistic effect; ii) The action of US was slightly better than that of Fe⁰_f; iii) Compared with dual systems, both action of US and Fe⁰_f enlarged in ternary systems (further confirming their synergistic effect).

In addition, the adding of US (or ClO₂⁻) may enrich the varieties and amounts of active substances in the reaction system, which greatly weakened the influence of pollutant concentration, leading to de-order reaction, but Fe⁰_f alone seemed to play not obvious role on the above.

3.2. Effects of factors

3.2.1. Effect of pH with/without US

The pH values greatly affected the systems of Fe⁰_f-O_SC and US/Fe⁰_f-O_SC on removing tpm_CV (Fig. 2a). Generally, the acidic condition was beneficial for degradation. After introducing US, organic pollutants could evaporate into the cavitation bubbles easily, and then be pyrolyzed directly in them under acid condition. Since pH decreased from 11 to 2, the k_obs (10^-3, min⁻¹) increased from 41.37 to 345.80 (from 38.54 to 132.73) of US/Fe⁰_f-O_SC (Fe⁰_f-O_SC). However, due to US role, the pH application rage of system widened (specifically, the k_obs of Fe⁰_f-O_SC at pH 2 ≈ that of US/Fe⁰_f-O_SC at pH 4; the k_obs of Fe⁰_f-O_SC at pH 6.8 ≈ that of US/Fe⁰_f-O_SC at pH 10.1) as shown in embedded Fig. 2a. Meanwhile, when pH presented alkaline, the reaction rate both in Fe⁰_f-O_SC and US/Fe⁰_f-O_SC seemed to be accelerated, but it would decelerate. The greater the alkaline, the more obvious the above-mentioned phenomenon, which could be ascribed to ClO₂⁻ (alkali activation to form ClO• being different from acidic reaction to generate ClO₂) [20], [22]. The S₂O₈²⁻ was dominant, thus the alkali role to ClO₂⁻ could not sustain over the whole reaction process. Due to US polishing (comparing the Fig. 3) and ClO₂⁻ co-catalysis, the systems actually kept certain ability of treatment, and the passivation of Fe⁰_f obtained remission.

Fig. 2.

Effects of (a) acid-base (comparisons of US/Fe⁰_f-O_SC and Fe⁰_f-O_SC), (b) fluid eddy (ibid. (a)), (c) temperature, and (d) ultrasonic intensity & frequency (basic conditions referring to Fig. 1).

Fig. 3.

Comparisons of Fe⁰_f with SEM (1 μm, 2 μm, and 20 μm) and XRD (10-90°): (a) original sample without treating, (b) with US-O_SC 30 min, and (c) & (d) with O_SC but without US, 15 min & 30 min (_{molar ratio}S₂O₈²⁻/ClO₂⁻ 5:1, f 28 kHz, I 30 W⋅L^-1, pH without adjustment, 0 r⋅min⁻¹, and initial T 289 K).

3.2.2. Effect of fluid eddy with/without US

The fluid eddy was realized by mechanical mixing. The liquid height [4] was determined based on the wavelength (λ) of the applied frequency (f), namely, λ = c_s/f where c_s was obtained as 1469.60 m⋅s⁻¹ with the modified temperature equation (Eq. S34). The site of mechanical mixing was placed at 1/2λ (distance from bottom) [5]. In order to describe the inherent attribute of fluid by the action of mechanical mixing, herein, the angular momentum (L), moment of inertia (I_v), and rotational kinetic energy (E_rot) were introduced to describe the role of fluid eddy with following Eqs. S35-S37 (Section S7).

By comparison of Fe⁰_f-O_SC and US/Fe⁰_f-O_SC systems in Fig. 2b, the fluid eddy both affected the removal of tpm_CV varying in three stages (increased → decreased → restrained), which was partly consistent with the study of [5]. The difference was that the introduction of O_SC (especially, ClO₂⁻) and Fe⁰_f led to the appearance of restrained stage. Enlarging the fluid eddy made S₂O₈²⁻ and ClO₂⁻ react preferentially to consume themselves by Eq. (17) (sonic catalysis weakening) and accelerate Fe⁰_f surface oxidation (O_SC and O₂) and even passivation.

$$S_2{O}_8^{2-}+2Cl{O}_2^{-}\to 2S{O}_4^{2-}+2Cl{O}_2$$ (17)

Due to the action of US, the peak-shift (a right-ward shift) occurred in terms of variation of k_obs. The enhancement in the sonochemical oxidation was attributed mainly to the direct disturbance of the ultrasound transmission and the resulting change in the cavitation-active zone.

3.2.3. Effect of initial temperature with US

The temperature affected the performance of US/Fe⁰_f-O_SC on removing tpm_CV (k_obs, (10^-3, min⁻¹) increasing from 85.64 to 309.47 when T increasing from 289 K to 318 K in Fig. 2c). The reaction thermodynamics of tpm_CV removal was described by Eyring equation (Eqs. (18), (19)) and Arrhenius equation (Eq. (20)) to obtain the experimentally derived free energy of activation (ΔG^≠), fitted enthalpy (ΔH^≠), entropy (ΔS^≠) of activation, and activation energy (E_a), respectively.

$$k_{\mathit{obs}}=\frac{k_{b}T}{h}{e}^{-(\mathrm{\Delta }{G}^{\ne }/RT)}$$ (18)

$$\mathrm{\Delta }{G}^{\ne }=\mathrm{\Delta }{H}^{\ne }-T\times \mathrm{\Delta }{S}^{\ne }$$ (19)

$$k_{\mathit{obs}}=A{e}^{-(E_a/RT)}$$ (20)

where R is the gas constant (8.314 J·mol⁻¹·K⁻¹); k_b is the Boltzmann constant (1.38 × 10^-23 J·K⁻¹); h is the Planck’s constant (6.626 × 10^-34 J·s); and the above equations were extended in Section S8 (Eqs. S38 and S39).

As shown in Fig. 2c, the liner relation in ln(k_obs/T) vs. 1/T (R²_Adj = 0.984) and ln(k_obs) vs. 1/T (R²_Adj = 0.985). Therefore, the ΔH^≠, ΔS^≠, E_a, and A were calculated as 31.117 kJ·mol⁻¹, −156.956 J·mol⁻¹·K⁻¹, 33.227 kJ·mol⁻¹, and 91.858 × 10³, respectively. Therefore, for US/Fe⁰_f-O_SC system, the relationships (ΔG^≠ vs. T and k_obs vs. T) were expressed as ΔG^≠= 31.117 + 0.157 T (kJ·mol⁻¹) and k_obs = 9.186 × 10⁴exp(–33.227 × 10³/RT) (min⁻¹).

3.2.4. Effect of US intensity and frequency

The effects of ultrasonic intensity and frequency (Fig. 2d) were discussed in Section S9.

3.3. Comparative analysis of degradation intermediates with/without US

In terms of US/Fe⁰_f-O_SC, mono-(S1-S13), di-(D1-D9), and tri-(T1-T4) phenylic compounds (Table S1) were detected. Comparing with double-systems Fe⁰_f-O_SC and US-O_SC, the US/Fe⁰_f-O_SC seemed to make the intermediates more diversified (Fig. S3) [25], especially, to produce more Cl-adducts (because of ClO₂⁻) [22] and mono-benzene compounds (Fig. 4a) [26]. The detailed analysis was shown in Section S10.

Fig. 4.

(a) Comparisons of intermediates detected by GC/MS under different systems and possible degradation pathway of tpm_CV on US/Fe⁰_f-O_SC (S: single benzene ring, D: double benzene rings, T: three benzene rings, Z: straight chain compounds, SMC: small molecule compounds (organic acid, CO₂, H₂O, etc.), the detailed intermediates with t_R and m/z seeing in Table S1); (b) GC/MS detection of PMSO and PMSO₂ (initial PMSO 6 mM) (the detailed intermediates with t_R and m/z seeing in Table 1); and (c) the transformations of PMSO in US/Fe⁰_f-O_SC and Fe⁰_f-O_SC.

3.4. Assessing US bubble/water interface

The k_obs could be measured at the macro-level based on the total UCAR solution. However, in order to obtain concrete rate constants for catalyactivation according to effectively reactive volume (k_int at bubble–liquid interface zone of cavitation nucleus where the reaction sites for catalysis and activation of US (or Fe⁰_f) to O_SC (or PDS) mainly located), the mass balance expression, as Wei et al. [27] mentioned, could be used to make a conversion from k_obs to k_int:

$$k_{\mathit{int}}=\frac{V_{\mathit{total}}}{V_{\mathit{int}}}{k}_{\mathit{obs}}$$ (21)

where k_int was the rate constant at the phase interface, V_total was the volume of wastewater in treatment, and V_int was the volume of interfacial shell surrounding cavitation bubbles. Here, assuming the interface thickness was 10% of the radius of the cavitating bubble: an empirical value of V_total/V_int was obtained from the calculation of [27].

The effective mean temperature of the interfacial region was estimated by k_int and Arrhenius equation. The Arrhenius parameters (lnA = 36.6 and E_a = 134 kJ·mol⁻¹) were referred to the study of thermal dissociation of PS [27], [28]. Therefore, at optimal degradation order, the corresponding temperatures of US/Fe⁰_f-O_SC, US-O_SC, US/Fe⁰_f-PDS, and US-PDS were calculated to be 531 K, 535 K, 737 K, and 744 K, respectively, which illustrated that i) Fe⁰_f had slight effect on the interface temperature; ii) ClO₂⁻ lowered the interface temperature. Although the interfacial zone would have a steep temperature gradient between the hot cavitation bubble and the ambient temperature of the bulk solution [27], the calculated temperatures above could be considered as the mean effective temperatures of the interfacial region for O_SC (or PDS) dissociation (activated by Fe⁰_f under the US roles), but < 1900 K found in the liquid region surrounding the collapsing cavitation bubbles in alkane solvents at 5 Torr vapor pressure under argon [29], [30]. As Misik et al. [31] described, the contaminants with different structures may also influence the interface temperature at sonochemical regions.

In addition, the calculated E_a of US/Fe⁰_f-O_SC was 33.227 kJ·mol⁻¹ > 10–13 kJ·mol⁻¹, indicating that degradation process was dominated by an intrinsic chemical reaction rather than a diffusion process. According to study of Cui et al. [32], the modified Freundlich Isotherm (Eq. (22)) could be applied to describe the relationship between the amounts of molecules partitioning (adsorption) at the interface of cavitation bubble and the initial concentration at equilibrium in US-based systems.

$$\log \left(r_t\right)=\frac{1}{k_r}\log \left(c_{{\mathit{tpm}}_{\mathit{CV},0}}\right)+\mathrm{l}\mathrm{o}\mathrm{g}(K_f)$$ (22)

where r_t (µM·min⁻¹) was the decomposed rate of tpm_CV (Eq. S40), k_r was called as the reactivity constant, and K_f was the adsorption capacity. According to the calculating of r_t, the Section S11 was presented as the general form.

Therefore, according to US/Fe⁰_f-O_SC, the log(r₀) vs. log(c_tpmCV,0) had a good liner relation (R²_Adj = 0.996, Fig. 5a), and the k_r,0, indicating the adsorption intensity, was 0.823, while the K_f,0, as [32] described, represented the capacity of tpm_CV molecules adsorbing or partitioning to the reactive zone (gas–liquid interface) and was 0.0442. Similarly, the k_r,end was 0.728, while the K_f,end was 0.0659 (R²_Adj = 0.987). By comparison, the adsorption or partitioning capacity of tpm_CV to the reactive zone was increased as the degradation progressed.

Fig. 5.

(a) Freundlich degradation isotherms of tpm_CV in Fe⁰_f-O_SC solution media under US irradiation (f 28 kHz, I 30 W⋅L^-1), (b) inhibition of different systems via screening agents (6 M MA and 1 M TBA), (c) variation of various Fe concentrations in US/Fe⁰_f-O_SC, and (d) ClO₂ detection of US-O_SC by UV–vis spectrum (T 303 K).

3.5. Discussion of degradation mechanism induced by US

3.5.1. Contribution of main active species with/without US

The reaction rate of tert butyl alcohol (TBA) vs. •OH (3.8–7.6 × 10⁸ M⁻¹·s⁻¹) was much higher than that of TBA vs. SO₄•⁻ (4.0–9.1 × 10⁵ M⁻¹·s⁻¹) [33], while the reaction rate of methyl alcohol (MA) vs. •OH (9.7 × 10⁸ M⁻¹·s⁻¹) was close to that of MA vs. SO₄•⁻ (2.5 × 10⁷ M⁻¹·s⁻¹) [34]. The results of inhibition reactions of different systems with different scavengers were shown in Fig. 5b and Fig. S4. According to Eq. (23), the relative degrees of dependence in •OH or SO₄•⁻ could be obtained.

$${\delta }_i=\frac{\mathrm{\Delta }{k_{\mathit{obs}}}_i}{{k_{\mathit{obs}}}_{(none)}}$$ (23)

where δ_i represented the shielding effects of different systems with different scavengers, Δk_{obs, i} represented the difference values of different scavengers, and i represented MA or TBA.

By comparison of δ (MA 6 M, TBA 1 M, relatively enough):

i) δ_MA-(Fe⁰_f-O_SC) (0.692) ≈ δ_MA-(O_SC) (0.691): both generated •OH and SO₄•⁻ of O_SC and Fe⁰_f-O_SC could be controlled by 6 M MA.

ii) δ_TBA-(Fe⁰_f-O_SC) (0.283) ≪ δ_TBA-(O_SC) (0.528) but k_obs-(Fe⁰_f-O_SC) > k_obs-(O_SC): some other active species (maybe Fe^IV/V) existed.

iii) δ_MA-(US-O_SC) (0.805) > δ_MA-(O_SC) and δ_TBA-(US-O_SC) (0.594) > δ_TBA-(O_SC): the introduction of US could promote the system to generate more both of •OH and SO₄•⁻.

iiii) δ_MA-(US/Fe⁰_f-O_SC) (0.515) < δ_MA-(all the others) but δ_TBA-(US/Fe⁰_f-O_SC) (0.643) > δ_TBA-(all the others): the •OH and SO₄•⁻ were also the main active species of systems, but the introduction of US tended to promote the active species to transform to more •OH.

3.5.2. Variation of common Fe and detection of ClO₂ with/without US

The variations of the dissolved Fe (Fe^II and Fe^III), suspending Fe & dissolved Fe, and Fe weight loss (suspending Fe, dissolved Fe, and sedimentary Fe) in US/Fe⁰_f-O_SC were shown at Fig. 5c. The dissolved Fe seemed to exist a dynamic (fluctuant) equilibrium, indicating a ferrikinetics utilization process. By comparisons of Fe⁰_f morphology changes among Fe⁰_f-O_SC and US/Fe⁰_f-O_SC (SEM, Fig. 3), and Fe⁰_f-PDS and US/Fe⁰_f-PDS (previous study), US also had the polishing role in US/Fe⁰_f-O_SC, but more serious Fe-corrosion in O_SC indicated that ferrikinetics may be not sourced from Fe-self circulation (different from PDS), which be mainly attributed to the ClO₂⁻ role. ClO₂⁻/generated-ClO₂ took part in the reaction of Fe, enriching the transformation ways, due to US polishing role, to not only ease Fe⁰_f surface passivation but also accelerate Fe⁰_f surface corrosion (Fig. 3b). But interestingly, without US polishing, the Fe corrosion and formation process of Fe_xO_y on Fe⁰_f surface could be observed under enhanced role of O_SC (Fig. 3a smooth → Fig. 3c rough → Fig. 3d “flower”-like). Meanwhile, the relative intensities of characteristic peaks (three) of Fe⁰_f were also changed under O_SC.

In addition, the ClO₂ was detected by UV–Vis spectrum (ClO₂ different from other chlorines in water, existing absorption peak around 360 nm) [20], [21]. From Fig. 5d, the produced ClO₂ presented throughout the reaction process (gradually increasing) on US-O_SC by the reaction of Eq. (17). In addition, the anions of S₂O₈²⁻/ClO₂⁻ after reaction consisted of ClO₂⁻, Cl⁻, SO₄²⁻, and a few ClO₃⁻ (Fig. S5) by IC detection.

3.5.3. ROC-HRMS and EPR detection with/without US

3.5.3.1. ROC-HRMS detection

The spin trapping of DMPO•-SO₄⁻ was the predominant O-centered adduct for SO₄•⁻, and the adduct could be protonated and decomposed over time to DMPO•–OH [35].

According to US-O_SC system (at negative mode), the peak at m/z 209 was detected by LTQ and LTQ-Orbitrap, respectively. The MS/MS results of m/z 209 in US-O_SC system was totally different to those in background, and the peak (m/z 209) in background was weak and sporadic but in US-O_SC system was stable by parallel experiments, thus indicating that the above peak in background belonged to signal noise but in US-O_SC system was a new peak. By MS/MS results, the peak at m/z 209 was much close to DMPO•-SO₄⁻ (Interestingly, the peak at m/z 209 being observed in both positive and negative modes with similar but different MS/MS results). Especially, the fragment “m/z 97” (failing to capture MS/MS/MS result) in m/z 209 in negative mode implied an unstable whole structure just like “-OSO₃(H)” [36], [37], [38]. In addition, the unstable peaks at m/z 209(+1, +2) with Δ m/z (peaks at m/z 209(+1, +2) and 112(+1)) around 97(-1, +1), would appear alternately in positive mode (inferring as Scheme 2b).

Scheme 2.

Transformations of DMPO-R, -R⁻[35], [40], [41], [42].

PN-HRMS analysis also confirmed the presence of a peak at m/z 130 (only in positive modes) by MS/MS results, which could be explained with DMPO•–OH after abstraction of •H from another molecule in the sample mixture (observed mass was shifted at + 1 Da relative to the mass of DMPO•–OH [39]) (inferring as Scheme 2a). Previous studies associated m/z 130 with an adduct between the DMPO and •OH [40], [41], [42] (Δ m/z 18 between m/z 130 and 112 representing de –OH and -H), which was also observed in EPR in our experiments.

As continuous reaction and detection, the peak at m/z 209 gradually weakened even disappeared, but the peak at m/z 130 enhanced (compared with Fig. 6b & 6d, the peak could be clearly captured in LTQ-Orbitrap over time). Meanwhile, the ESI with the adding MA (compared with Fig. 6c & 6d) attempted to enhance the effect of peaks appearance but resulted in that peaks at m/z 130 and 209 were weaken, and other peaks remained unchanged or indeed enhanced. Herein, two points (mutual supporting) could be further obtained: i) MA indeed could quench both of SO₄•⁻ and •OH; and ii) peaks at m/z 130 and 209 most likely corresponded to DMPO•–OH and DMPO•-SO₄⁻, respectively.

Fig. 6.

ROC detection by MS or MS/MS (_{molar ratio}S₂O₈²⁻/ClO₂⁻ 5:1, pH without adjustment): (a) O_SC, positive mode, LTQ; (b) US-O_SC, positive mode, LTQ& LTQ-Orbitrap (long-time); (c) US-O_SC, negative mode, LTQ& LTQ-Orbitrap; and (d) US-O_SC, positive mode, LTQ& LTQ-Orbitrap (note: i. operating parameters in Section 2.4; ii. MS/MS results in embedded figure; iii. here presenting comprehensive MS results of continuous period not representing one time point, and continuous period ranging from 3 to 15 min).

According to O_SC system (Fig. 6a), however, the peak (m/z 209) was clearly detected but the peak (m/z 130) was hardly detected, suggesting that the main free radicals of O_SC were SO₄•⁻. Actually, because of the Scheme 3 [35], the difficulty of free radical identification was increased [43]. However, by PN-HRMS with ROC detecting, the detection (consumption) was along with the reaction (from “without US originally” to “after introducing US”) to gain the variations of peaks at m/z 130 and 209 in real time, illustrating that i) •OH mainly originated from reaction of US-O_SC, not Scheme 3; ii) after introducing US, the type of dominant free radicals tended to be •OH in the system and the transformation may be mainly initiated from SO₄•⁻ (being consistent with Section 3.5.1-iiii); and iii) the PN-HRMS with ROC detecting (comparing with traditional detecting means, liquid chromatograph/mass spectrometer (LC/MS) or EPR with intermittent sample testing one by one) at a certain degree could avoid the potential misjudgement due to Scheme 3.

Scheme 3.

Intramolecular cleavage decomposition of O-centered sulfate radical adducts [35].

But most importantly, the ROC-HRMS seemed to prove the radical chain reaction (chain transfer SO₄•⁻ to •OH) directly in the reaction process.

3.5.3.2. EPR signals

The EPR signals of different systems (ClO₂⁻, US-ClO₂⁻, O_SC with different molar ratio, US/PDS, US-O_SC, Fe⁰_f-PDS, Fe⁰_f-O_SC, US/Fe⁰_f-PDS, and US/Fe⁰_f-O_SC) were detected to conduct comparative analyses. Commonly, introducing US would not change the peak shapes of EPR signals (e.g. ClO₂⁻ vs. US-ClO₂⁻, O_SC vs. US-O_SC, Fe⁰_f-O_SC vs. US/Fe⁰_f-O_SC, etc.). Differently, single PDS with DMPO would still have the signals, which was consistent with study of Zamora and Villamena [35], possibly due to -OSO₃(H) and –OH.

The DMPO-X₁ (♥) signal with a single nitrogen (a^N = 16.50 G (16.30–16.60), a triplet 1:1:1) [44] may contribute from the oxidation of DMPO itself, but according to study of Kondo et al. [45], peaks labeled “♥” may be due to DMPO•-H adducts showing a primary triplet (a^N = 16.50 G) further split by two secondary hydrogens (⋄, a^H_β = 22.40 G). From Fig. 7a, no •OH signal was observed in ClO₂⁻ and US-ClO₂⁻, however, besides DMPO-X₁ signal, the remaining signal DMPO-X₃ (×) seemed to be closed to ClO• signal [46] and ClO•+Cl• signal [47]; in addition, according to description of Ozawa et al. [48], combining with the DMPO-X₁ and DMPO-X₃, it may contribute from oxidation of spin-traps by ClO₂•; while according to the study of Nishikawa et al. [49], ClO_x (x = 2, 3) radicals led to the above phenomenon. Due to the Cl(III) in ClO₂⁻, an intermediate valence [50], it could be explained as ClO_x (x = 0–3) radicals contributed to the above-mentioned results.

Fig. 7.

Comparisons of EPR results under different systems: (a) the step-by-step EPR results of coupled systems, ClO₂⁻ & US-ClO₂⁻ 100 mM, DMPO 100 mM, 5 min; S₂O₈²⁻/ClO₂⁻ 50 mM/50 mM, DMPO 100 mM, 3 min; O_SC & US-O_SC (S₂O₈²⁻/ClO₂⁻ 100 mM/20 mM), DMPO 100 mM, 1 min; Fe⁰_f-O_SC & US/Fe⁰_f-O_SC (S₂O₈²⁻/ClO₂⁻ 5 mM/1 mM, Fe⁰_f 0.358 mM), DMPO 6 mM, 30 s. (b) the EPR results of O_SC-based systems after adding DMSO, O_SC (30 s) & US-O_SC (5 min) with DMSO (S₂O₈²⁻/ClO₂⁻ 100 mM/20 mM, DMPO 100 mM, DMSO 100 mM); Fe⁰_f-O_SC (10 s) & US/Fe⁰_f-O_SC (i: 1 min, ii: 2 min) with DMSO (S₂O₈²⁻/ClO₂⁻ 5 mM/1 mM, Fe⁰_f 0.358 mM, DMPO 6 mM, DMSO 6 mM). (c) the EPR results of PDS-based systems after adding DMSO, S₂O₈²⁻ 100 mM, DMPO 100 mM, 10 min; Fe⁰_f-S₂O₈²⁻ (5 min) and US/Fe⁰_f- S₂O₈²⁻ (2 min) (S₂O₈²⁻ 6 mM, Fe⁰_f 0.358 mM, DMPO 6 mM, DMSO 6 mM). (♦ DMPO•-SO₄⁻, • DMPO•–OH, *(*) DMPO•- SO₃⁻ & OH, ♥, × & ++ DMPO-Xs, ⋄ DMPO•-H).

In terms of O_SC and US-O_SC, DMPO-X₁ signal was almost disappeared, the DMPO•-SO₄⁻ (♦) signal (a^N = 13.78 G, a^H_β = 9.97 G, a^H_γ1 = 1.47 G, a^H_γ2 = 0.98 G) (O-centered adducts) and DMPO•–OH (•) signal (a^N = a^H = 14.91 G, a quartet 1:2:2:1) were clearly obtained. Specially, in terms of relative peak intensity in same spectrum, the relative DMPO•-SO₄⁻ signals in O_SC (S₂O₈²⁻/ClO₂⁻ 5:1) or after introducing US were more highlighted than that of O_SC (1:1) or single PDS, which illustrated co-activation between ClO₂⁻ and PDS.

In terms of Fe⁰_f-O_SC and US/Fe⁰_f-O_SC, the DMPO•-SO₄⁻ and DMPO•–OH signals could be still detected but the DMPO-X₁ signal was enhanced obviously, which indicated that the introduction of Fe⁰_f accelerated reaction, meanwhile may produce other strongly oxidizing substance besides radicals to result in oxidation of DMPO.

In addition, DMSO as the probe was added the above systems to detect free radicals indirectly [51] due to the reaction of Scheme 4. As the addition of DMSO (Fig. 7b), unfortunately, the •CH₃ signal could not be detected; instead, the quartet signal DMPO-X₂ (* & *, near to 1:1:1:1, a^N = a^H = 14.50 G) appeared (Fig. 7b & 7c). The phenomenon was also observed by Zalibera et al. [52] (the mixtures (DMSO/H₂O) with the lower dielectric permittivity significantly influenced the hyperfine coupling constants and the observed signals largely deviated from typical DMPO•–OH adduct in aqueous media); but it seemed to be consisted with the study of Jiang et al. [53] (contributing from SO₃•⁻ [54] which had be explained by Ranguelova et al. [55]). And according to study of Zamora and Villamena [35], the above quartet signal may source from SO₃•⁻ (S-centered adducts) and •OH with different proportions. In addition, the DMPO•-SO₄⁻ signal was disappeared, which may result from the competition reaction of DMSO vs. SO₄•⁻ [56].

Scheme 4.

Reactions of DMSO under •OH, SO₄•⁻, and US in literature reports [56], [58], [59], [60].

For comparison, the EPR of Fe⁰_f-PDS and US/Fe⁰_f-PDS was operated (Fig. 7c). The detected signal of Fe⁰_f-PDS was matched to the study of Zhu et al. [57] (adopting Fe-base activation), belonging to “DMPOX”, but it seemed to be more closed to composite of DMPO-X₁ and DMPO-X₂ in this study. However, after introducing the US, the signals were complex because of the appearance of DMPO-X₄ (++, ?). Meanwhile, the introducing of Fe⁰_f would force and make the EPR signals right-shift and complex (Fe⁰_f-O_SC vs. US/Fe⁰_f-O_SC, Fe⁰_f-PDS vs. US/Fe⁰_f-PDS).

3.5.4. Exploration of US role to O = Fe^IV/V

Catalytic iron center formed a highly oxidizing oxo-iron species Fe^IV/V (namely, O = Fe^IV and O = Fe^V) [65] which had the reaction with the sulfoxide in Scheme 5 to generate the corresponding sulfone by oxygen atom transfer step [61], [62]. Formally, O = Fe^V was best described as an oxo-Fe(IV)-porphyrin radical cation, namely, O = Fe^IV(Por)•⁺ (O = Fe^IV•⁺) [65], [66]. The PMSO was used as chemical probe in this study, and the methyl phenyl sulfone (PMSO₂) with m/z of 156, 141, 125, etc. in Table 1 was detected by GC/MS (Fig. 4b), which could demonstrate that the existence of Fe^IV/V in Fe⁰_f-O_SC (t_R 9.34 min) and US/Fe⁰_f-O_SC (t_R 9.32 min) systems with larger proportion. Meanwhile, it was also corresponding to the aforementioned δ_TBA of Fe⁰_f-O_SC (namely, TBA could not effectively inhibit the degradation). In terms of Fe⁰_f-O_SC (US/Fe⁰_f-O_SC) system, the formations of Fe^IV/V as shown in Scheme 6, on the one hand, sourced from conversion reaction between “-O-O-” (from PDS) and Fe^II/III (from Fe⁰) [63], [64], [66], [67]; on the other hand, contributed from adjustment reaction between ClO₂⁻ and Fe^II/III [68], which were operated through a Fe^II/Fe^IV cycle [65], [66], [68]. The kinds and relative peak intensities of the intermediates (especially, PMSO₂) in US/Fe⁰_f-O_SC were greater than those in Fe⁰_f-O_SC, implying that US could facilitate the transformation reaction.

Scheme 5.

Reaction between sulfoxides and Fe(IV) in literature reports [61], [62], [63], [64].

Table 1.

The information of generated products after adding PMSO identified by GC/MS.

Intermediates	Retention time (min)		Detected fragments (m/z)	m/z
	US/Fe0f-OSC	Fe0f-OSC
PMSO	8.58	8.66	140 125 97 94 77 65 51	140
PMSO2	9.32	9.34	156 141 125 94 77 65 51	156
P1	10.53	10.55	174 (1 7 6) 158 138 125 97 77 51	174
P2	11.59	11.61	186 (187, 188, 185, 184) 152 125 109 84 (86) 77 65 51	186
P3	14.18	14.21	218 (219, 220) 185 154 140 109 84 65	218
P4	15.16	/	218 152 125 97 77 65 51	218

Scheme 6.

Reactions of Fe^II/Fe^IV cycles [63], [66], [68], [70].

In addition, the detailed information of other formed intermediates P1-P4 was shown in Table 1. PMSO (sulfoxide) would undergo the reaction of “Pummerer rearrangement” to transfer to Ph-thiol (unstable in oxidation system), then further turn into Ph-S-Ph (P2) and Ph-S-S-Ph (P3), meanwhile, P2 would be oxidized into P4 (the detailed process could be seen in Fig. 4c) [69]. In addition, PMSO under the role of ClO_x would turn into P1 via substitution.

3.5.5. Possible degradation process with US

By detected results and literatures, the possible mechanism of degradation process in terms of US/Fe⁰_f-O_SC was inferred and proposed as follows: at neutral pH, Fe^IV would be as the important intermediate in Fenton like reaction [71], but Fe^IV would reduce the contribution in ZVI/PDS as result of the decrease of available Fe^II [72]. However, US/Fe⁰_f-O_SC at near-neutral pH still kept well activity, therefore, the introduction of ClO₂⁻ not only cooperated with PDS to generate co-activation process (SO₄•⁻ and ClO₂) but also promoted the production and circulation of Fe^IV in presence of Fe⁰_f (with the cycle of Fe^II, Fe^III, O = Fe^IV•⁺, and O = Fe^IV). Meanwhile, US could play another role on the circulation by polishing the surface of Fe⁰_f to provide the fresh ZVI source. Meanwhile, the generated ClO_{x (0−3)} would accelerate corrosion of Fe⁰_f phase interface to release ionic Fe. Importantly, US would also promote production and transformation of O = Fe^IV/V, and free radical chain transfer (from SO₄•⁻ to •OH).

The main reaction processes were shown in Fig. 8 and Section S12.

Fig. 8.

The possible reaction mechanism of US/Fe⁰_f-O_SC.

3.6. Evaluation of energy consumption based on US

The electrical efficiency per log order (EE/O) was applied to evaluate the application prospect of US/Fe⁰_f-O_SC (seeing Section S13 for details). Table S2 listed the calculated EE/O and total cost values of different AOPs for treating wastewater. By economic comparison, the US/Fe⁰_f-O_SC with EE/O value 13.44 kWh⋅m⁻³ and total cost value 3.64 $⋅m⁻³.

4. Conclusions

Due to the limitation of ability, the determination (EPR and ROC/MS) of active species in the complex US/Fe⁰_f-O_SC, and -like systems still face the challenges and should be further explored and analyzed (e.g. isotope detection (¹⁸O, ¹⁷O, and ¹⁶O) or (H, D), etc.). The establishment and calculation of US cavitation model involving bubble–liquid dynamics should be further enhanced.

In this study, the EPR was applied to detect the radicals species of different systems for comparison. The ROC-MS (substitution of EPR) was a preliminary attempt to track the radicals transformation. The US apparent roles to removal reaction were expounded (e.g. de-order reaction, widening the applied range of pH and fluid eddy). The inherent roles of US were stated as follows:

i) US could induce ClO₂⁻ to co-activate PDS, low the bubble-water interface temperature, and promote the production and circulation of Fe^II/Fe^IV; ii) US could enrich the kinds of contaminant intermediates, polish the surface the Fe⁰_f to avoid surface passivation to provide fresh Fe source, promote the production and transformation of O = Fe^IV/V, and push free radical chain transfer from SO₄•⁻ to •OH.

In addition, the US/Fe⁰_f-O_SC may reveal other roles: relatively economical treating cost and potential (ClO₂) disinfection.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: