The superior decomposition of 2,4-Dinitrophenol under ultrasound-assisted Fe3O4@TiO2 magnetic nanocomposite: Process modeling and optimization, Effect of various oxidants and Degradation pathway studies

ABSTRACT In the present study, the ultrasound-assisted Fe3O4@TiO2 magnetic nanocomposite was employed to catalytic oxidation of 2,4-Dinitrophenol. The catalyst features were characterised by SEM, TEM, XRD, BET, FTIR, VSM, DLS and TGA techniques. The effect of operational parameters i.e. pH, catalyst amount, 2,4-Dinitrophenol concentration and ultrasound power on DNP removal was examined and optimised in terms of sono-catalytic degradation system using RSM-based CCD approach. Over 91.45% of DNP were removed by Fe3O4@TiO2 /US system (FTU) under optimum conditions within 30 min and the mineralisation degree was found to be 73% and 64.2% based on COD and TOC, respectively. Quenching experiments confirmed that O2•- is dominant radical species in degradation process. FTU system was more successful in activation of S2O82- than IO4− and H2O2. Based on GC-MS analyses, the possible decomposition pathway was proposed. AOS and COS analyses indicated that FTU system can improve the bioavailability of DNP. The effect of CO32− on DNP degradation was more complicated than other anions. Fe3O4@TiO2 could keep its performance to at least the 5th cycle while Fe and Ti leaching was negligible. The performance of FTU system for DNP treatment under real conditions was tested and promising results were achieved.


Introduction
Phenolic compounds have a severely adverse effect on human health and the environment.Among phenolic compounds, 2,4-Dinitrophenol (DNP) was listed as a priority pollutant of the US Environmental Protection Agency due to its high toxicity, carcinogenicity, environmental stability, and low biodegradability properties [1,2].DNP is widely used in several industries such as production of dyes, pesticides, explosive materials, plastics and preservatives [3].Due to the adverse impacts of DNP on the ecosystem and its high harmfulness, this contaminant has become very considered by researchers in the last years [4,5].In other side, exposure to DNP even at low concentrations, can cause many destructive and irreversible effects on human health, i.e. endocrine glands and reproductive organs disorders, eye and skin irritation, headache, liver damage, and irregular heartbeat [6].Hence, DNP removal before discharging to the receiving resources is critical to human health and environmental protection.In the past decades, different physicochemical treatment methods including photocatalytic degradation [7], Fenton oxidation [8], catalytic ozonation [9], and anaerobic elimination [10] have been employed to remove DNP from industrial effluents.However, the mentioned approaches have been some disadvantages, such as high operational and funding costs, high energy consumption, pre/or post-treatment, low toleration versus toxic elements, and sudden shocks [11].Hence, the removal of DNP from wastewater represents an emerging environmental concern.A number of studies elicited that conventional biological treatment methods are not effective enough to remove DNP in high level concentration [12], so that interest in photo-catalysis [13], electro-catalysis [14], Fenton-assisted oxidation processes [15] and chemical ultrasonic [16] methods have been intensified in recent years.Employing titania (TiO 2 ) as a photocatalysis with interesting physicochemical characteristics including costeffectiveness, high reactivity, biological and chemical inertness, chemical stability and availability is recognised as an applicable approach to degrade the organic contaminants [17].Wang et al. [18] have investigated the photocatalytic degradation of DNP using MWCNTs/TiO 2 and reported MWCNTs/TiO 2 was effective in removing DNP from solution at acidic conditions.In another study, Chand et al. [19] have reported an influential role of TiO 2 in DNP photo-degradation within 1 h of time.Furthermore, recent studies also confirmed that the degradation process is suitable technique for DNP removing from the aquatic environment [20].However, difficult separation (collection) and reusability of TiO 2 from solution has been taken into account as a major problem [21].To solve the aforementioned problem, Fe 3 O 4 NPs have been employed to magnetise the catalyst in recent years [22].The benefits of this technique include improving the catalyst recyclability, reducing recombination, and enhancing the catalyst efficiency through ferrous ions based on the following equations [23]: On the other hand, adsorption of intermediate products generated from the degradation process on the porous surface of Fe 3 O 4 and their concentrate at or near the surface of TiO 2 to continue the degradation process is another important point about Fe 3 O 4 [24].When Fe 3 O 4 @TiO 2 as a photocatalyst is used on a large scale (industrial wastewater treatment), there are two major problems: (1) lack of contact between UV radiation and catalytic agent due to the dark nature of industrial wastewater which consequently prevents the production of free radicals, and (2) agglomeration Fe 3 O 4 nanoparticles and a reduction in surface reactive sites for contaminant adsorption [25].High-intensity UV irradiation can sometimes alleviate the former problem, although it requires energy and expensive equipment.Applying the ultrasonic irradiation instead to UV radiation is suggested as an alternative to solve both problems [26].Ultrasonic irradiations produce a considerable number of electron-hole pairs and OH radicals.They also lead to acceleration in the transfer of electrons in the TiO 2 crystal structure [23,27].Furthermore, ultrasonic irradiation contributes to the generation of microbubbles in aquatic solutions, which therefore influences the production of OH radicals and consequently improves the contaminant degradation process [28].
Given the issues introduced here, this study was therefore developed to (I) synthesise magnetic TiO 2 nanocomposite (Fe 3 O 4 @TiO 2 ) and characterise its morphology, (II) examine the efficiency of FTU in DNP degradation by considering the effects of operating parameters including pH, Fe 3 O 4 @TiO 2 dose (Cat.amount), DNP initial concentration (DNP Con.) and US power using response surface methodology (RSM) as a statistical approach, (III) predict the DNP degradation results with RSM-CCD method and compare with actual results, (IV) optimise the parameters influencing degradation in terms of Desirability Function (DF) and compare with experimental results, (V) determine the intermediates, by-products and pathway of the degradation process using GC-MS, and (VI) study biodegradability and mineralisation of process.

Chemicals and instruments
Chemicals and Instruments used for the study were given in Supplementary data, Context 1.

Synthesis of catalysts and its characterisation
The detailed information concerning the procedure of prepared catalysts and their characterisation in current research is provided in Supplementary data, Context 2.

Degradation experiments
The degradation capacity of DNP was determined by putting a specified quantity of the catalyst in different concentrations of the pollutant.The solution of DNP and catalyst was stirred by a mechanical stirrer for 30 minutes before being irradiated with ultrasonic waves to achieve the adsorption-desorption equilibrium on the catalyst's surface and to eliminate the influence of adsorption during the sonocatalytic process.All of the tests were carried out in a sealed batch reactor.Fig S1 , shows a schematic of the planned reactor for this investigation.The pH, catalyst amount, DNP concentration, and ultrasonic power used in the degradation studies were all in the predetermined values.The ultrasonic wave was irradiated at specific time intervals.After that, a magnet was used to separate the catalyst from the solution and the residual concentration of DNP in the solution was quantified by HPLC.The following equation was used to describe the ability of the Fe 3 O 4 @TiO 2 -US (FTU) system to degrade DNP: Where C 0 and C e are the initial and residual concentrations of DNP (mg/L), respectively.
After the optimisation of the degradation conditions, the kinetics of the process were studied and reported.Because most studies utilise first-order kinetics to describe the degradation of organic and inorganic pollutants, the rate of degradation of DNP was investigated using first-order kinetics in this work.The first-order kinetic equation is as follows: Where K obs represents the rate of the reaction (min −1 ), C 0 and C e are the initial and residual concentrations of the contaminants (mg/L), and t is the time (min).

Statistical design
The information about the procedure of Statistical design is provided in Supplementary data, Context

Samples characterisation
Scanning electron microscope (SEM) images of the Fe 3 O 4 NPs and synthesised Fe 3 O 4 @TiO 2 are presented in Figure S2 (a-b).As obviously observed, synthesised Fe 3 O 4 @TiO 2 are in spherical structures with a size range of 30 −100 nm.The image reveals that the surface of the samples is rough and has a good porosity.This feature (presence of holes) improves the exposed (contact) between the DNP and Fe 3 O 4 @TiO 2 , which resulted in an improvement in adsorption and degradation performance.Furthermore, no clot (agglomeration) is seen in the Fe 3 O 4 @TiO 2 structure, and the Fe 3 O 4 @TiO 2 has a uniform structure.Fe 3 O 4 NPs consists of agglomerated bulks, which show ball-like shapes and sizes around 70 ± 20 nm.Fe 3 O 4 NPs shows relatively uniform particle distribution with a smaller size than Fe 3 O 4 @TiO 2 .Energy Dispersive Spectroscopy (EDS) tests were carried out to examine the element composition of Fe 3 O 4 @TiO 2 and the corresponding results are indicated in Figure S2 (b,  inside).The element peaks of O, Ti, and Fe can be observed in the spectrum of Fe 3 O 4 @TiO 2 which confirms that TiO 2 and Fe 3 O 4 are synthesised and stabilised well in the catalyst structure.A transmission electron microscope (TEM) was applied to investigate the shape and particle sizes of Fe 3 O 4 NPs and Fe 3 O 4 @TiO 2 .The images indicate that both Fe 3 O 4 NPs and Fe 3 O 4 @TiO 2 are with almost a spherical shape.Figure S2 (c) confirms that the size of Fe 3 O 4 NPs is in the range 60 nm. Figure S2 (d) shows that a uniform shell of TiO 2 is coated on the Fe 3 O 4 surface in core shell form.The spherical shape of the cores has a size distribution around 70 ± 9 nm while the average thickness of the shell is about 26 ±6 nm.The Fe 3 O 4 and TiO 2 can be recognised by dark and bright regions, respectively.The electron binding ability of TiO 2 is lower than that of Fe 3 O 4 , so the shells identified as the lighter region (bright regions) are compared to the cores.DLS technique was employed to determine the particle size distribution data: the obtained average sizes of Fe 3 O 4 NPs, TiO 2 and Fe 3 O 4 @TiO 2 were 74 nm, 22 nm and 90 nm, respectively.The Fourier-transform infrared spectroscopy (FTIR) of Fe 3 O 4 NPs and synthesised Fe 3 O 4 @TiO 2 in the range of 4,000-400 cm −1 are presented in Figure S2 (e).In both spectra, the O-H (hydroxyl) group on the surface of synthesised NPs appears to be in a broad band around 3,600 cm −1 .The presence of strong bands at 590 cm −1 and 1,600 cm −1 in all samples is attributed to the Fe-O vibration of Fe 3 O 4 NPs and H-O-H bending in water molecules on the surface, respectively.The peaks in the Fe 3 O 4 @TiO 2 spectrum at 2,360 cm −1 , 1,340 cm −1 and 1,118 cm −1 are attributable to the C = O from the air, the Ti-O-Ti (Ti-O) and Fe-O-Ti bonds, respectively.In addition to the above-mentioned information, the combination of Ti-O-Ti stretching vibration and the Fe-O bond contributes to formation a broad band in range of 500 and 700 cm −1 , confirming the TiO 2 -coated Fe 3 O 4 structure.
X-ray diffraction (XRD) patterns of Fe 3 O 4 NPs and synthesised Fe 3 O 4 @TiO 2 in the range of 2θ = 10-80 ᵒ are represented in Figure S3 (a).The diffraction peaks at 2θ equal 30.22 ᵒ for (220), 35.65 ᵒ for (311), 43.36 ᵒ for (400), 53.73 ᵒ for (422), 57.58 ᵒ for (511), 63.13 ᵒ for (440) and 75.18 ᵒ for (533) according to JCPDS no.19-629 standard correspond to one lattice plane of Fe 3 O 4 NPs.In Fe 3 O 4 @TiO 2 , the diffraction peaks at 25.16 ᵒ for (101), 37.82 ᵒ for (004), 48.06 ᵒ for (200), 53.97 ᵒ for (105), 55.17  S3 (c-d).The isotherms are identified as type V with H1 typical hysteresis loops, which are the characteristic isotherm of mesoporous materials according to IUPAC classification.In addition, H1 typical hysteresis associated with porous materials consisting of well-defined cylindrical-like pore channels or agglomerates of approximately uniform spheres [29].Detailed information on other specifications of the samples including values of BET surface area, BJH cumulative pore volume and average pore size are presented in Table S2.The pore size distribution data indicates that average pore diameters of the TiO 2 , Fe 3 O 4 NPs and synthesised Fe 3 O 4 @TiO 2 are 6.9, 9 and 12.2 nm, respectively.The BET surface area of the TiO 2 , Fe 3 O 4 NPs and Fe 3 O 4 @TiO 2 was calculated to be 76, 138 and 127 m 2 /g.This decrease in the surface area indicates that the titania species are successfully deposited on the surface and inside the pores of the Fe 3 O 4 .Magnetic measurements and separation of Fe 3 O 4 and Fe 3 O 4 @TiO 2 were taken with a vibrating sample magnetometer (VSM) at room temperature, as presented in Figure S3 (e).This analysis was performed in a magnetic field of ±10 KOe and saturation magnetisations around ± 100 emu/g.The saturation magnetisations (Ms) of Fe 3 O 4 were observed around 76 emu/g, representing high magnetic properties and the separability of this compound from aquatic solutions.The Ms parameter for synthesised composite was about 51 emu/g.The lower saturation magnification of Fe 3 O 4 @TiO 2 was attributed to the presence of titanium dioxide in the structure of synthesised Fe 3 O 4 @TiO 2 , having no magnetic property.However, from Figure S3 (e, inside), the complete separation of Fe 3 O 4 @TiO 2 and subsequently transparency of the solution in less than 30 seconds is observed, which proves the excellent magnetic properties and reusability of the synthesised catalysis.

Effect of operational parameters and kinetic study
Since all the sonocatalytic reactions were commenced after adsorption equilibrium, the subsequent decrease in DNP concentration be attributed to degradation process.However, the effect of operating parameters on the FTU system requires further investigation, so several experiments were performed to examine the effect of reaction parameters on DNP degradation with Fe 3 O 4 @TiO 2 as the catalyst.

Effect of pH
The pH of the solution is considered as important operating parameter in degradation process; pH influences the surface charge of the catalyst and contaminant structure.The effect of initial pH solution (3)(4)(5)(6)(7)(8)(9)(10)(11) on DNP removal efficiency in FTU system was evaluated (See Figure 1 and Figure S4 (a)).Within 30 min of reaction time, the degradation efficiencies were decreased about ~32% by increasing pH values from 3 to 11.In fact, acidic conditions are more desirable for DNP degradation by Fe 3 O 4 @TiO 2 .Meanwhile, the degradation rate (k obs ) of the DNP decreased from 0.0705 to 0.0112 min −1 , as the pH of solution were increased from 3 to 11 (Figure not shown).As seen from figure, the degradation has an upward trend with an increase in pH ranging from 3 to 5 and then takes a downward trend at pH values over 5. Low degradation rates in strong acidic conditions (pH values 3 ±1) can be attributed to existence of excess hydrogen ions (H + ), which are able to scavenge free radicals [30].This is while, at pH 5, hydroxyl and superoxide radicals generated can strongly react with DNP molecules, which results in increased in decomposition (65.5%) and pollutant oxidation rate (0.0925 min −1 ).In the reaction mixture, OH radicals on the Fe 3 O 4 @TiO 2 can be formed through two routes: the reaction of gaps in the valence band (h vb + ) with adsorbed H 2 O molecules or OH − of surface titanol groups (TiOH).
O 2 •-was produced by reacting between sono-generated electrons (e cb − ) and O 2 .
Furthermore, the superoxide radicals further produce HO  [31].The equations addressed below clearly show the above-mentioned routes for radicals' formation.However, the substantial drop in yield at pH > 5 can be due to the formation of iron(III) oxide-hydroxide complexes following precipitation of iron ions, transformation of O À � 2 and OH � radicals into species with lower oxidation potential based on , and lower generation rate of free radicals because of electrostatic repulsion between catalyst and oxidant [32].
The electrostatic interactions and pH-dependent trend between the surface of catalyst and negative charge of DNP investigated by pH zpc are as follows: Figure S5 shows that the zero point of charge (zpc) for Fe 3 O 4 @TiO 2 was 6.2; catalyst surface is positively charged in a solution with pH <6.2 and negatively charged at pH> pHzpc.Electrostatic interactions between the protonated surfaces of Fe 3 O 4 @TiO 2 and DNP (negatively charged due to the presence of phenolic compounds) leads to adsorption and degradation of the corresponding contaminant in acidic pH (see Equations (16)(17)).Conversely, at pH> pHzpc, the surface charge of synthesised catalysis moves to a more negative value, causing a reduction in DNP degradation.In fact, repulsions between similarly-charged Fe 3 O 4 @TiO 2 surfaces and DNP prevent the agglomeration (adsorption) and subsequent efficient decomposition of DNP (see Equations (18)(19)), therefore, pH = 5 was considered as an optimal condition for performing subsequent experiments.Wang et al. investigated the degradation of DNP in aqueous solution under multi-walled carbon nanotubes (MWCNTs)/TiO 2 system and found that optimum pH for DNP removal was acidic which it is in line with the present work.Similarly, Guo and et al. reported that increasing pH from 3 to 9 cause decreases in NDP removal efficiency from 97.2% to 32.4% and low pH favoured the ultrasonic degradation of DNP [33].

Effect of DNP concentration.
To assess the influence of initial concentration of DNP on degradation performance using the FTU system, a series of experiments was performed in given contaminant concentration (5-45 mg/L).According to Figure 1 and Fig S4 (b), almost complete removal of DNP was achieved at <5 mg/L concentration during 30 min of reaction.However, removal efficiencies were decreased from 70% to 58% in cases of DNP concentrations from 5 to 45 mg/L.Furthermore, a 5-fold increase in DNP concentrations resulted in the reaction rate constant correspondingly falling from 0.1547 to 0.01938 min −1 .Reducing the efficiency can be explained by the competition of DNP molecules for adsorption on the catalyst surface (resulting in reduction in the rate of the adsorption process), restriction in US irradiation performance at higher DNP molecules and a decrease in the formation of reactive species, fixed concentrations of Fe 3 O 4 @TiO 2 vs. an increase in DNP concentrations, and generation of by-products in a large extent, which are competitors to parent compounds in reaction with oxidising radicals [34].Similar trends have been reported by other researchers for degradation of the DNP by combining sonolysis and different additives [33].

Effect of catalyst dosage
Figure 1 and Fig S4 (c), show the influence of different Fe 3 O 4 @TiO 2 dosages on the degradation of DNP under the FTU system, which was performed in 0.1 to 1 g/L within a 30 min reaction time.It can be seen the rate of DNP degradation and its efficiencies were improved by increasing the catalyst dosage.In fact, the degradation process approached its maximum value of 77.76% when catalyst dosage was 1 g/L.The apparent rate constants of DNP degradation in terms of Fe 3 O 4 @TiO 2 dosage at 0.1, 0.325, 0.55, 0.775 and 1 g/L were calculated as 0.0499, 0.0622, 0.0703, 0.1196 and 0.1277 min −1 , respectively.Adequate loading of the catalyst increases the generation of active sites, surface cavitation and enhance rate of electron/hole pairs to improve further pollutants degradation [35].Rani and et al [2].reported that increase in catalyst dosage to a certain amount has a positive effect on increasing the degradation process, but a further increase in catalyst results in lower efficiency [2].

Effect of ultrasonic irradiation power
A range of 25-100 W/L was considered to evaluate the effect of US power on the degradation of DNP; the corresponding results are shown in Figure 1 and Fig S4 (d).As pointed out in the figures, the degradation efficiency and rate of process (K obs ) increased from 45% to 75.76% and 0.0549 to 0.1146 min −1 when US power rose from 25 to 100 W/L.Increased OH radicals, improved emission and mass transmission, accelerated cavitation effect and ameliorated surface cleaning of the catalyst can be reasons to explain the obtained results.It should also be stated that some researchers believe that occurrence of microstreaming along with the ultrasound irradiation plays paramount role to clean the surfaces of MNPs, which accordingly improve degradation rate following more formation of reactive species [36].

Statistical analysis of the operating parameters
The following statistical formula presents the relationship between the influencing parameters and degradation efficiency of DNP: The positive and negative signs in the above equation show that the type of independent parameters influence the response.The positive sign shows the direct effect and the negative sign indicates the inverse effect of the parameters on degradation efficiency.The results of ANOVA analysis considering the influence of the operating parameters in DNP degradation are summarised in Table 1.
The values of two statistical parameters, R-square (0.990) and R-square adjusted (0.977), confirm that the DNP degradation process by the FTU system is fitted with the model designed for the experiments.Adequate precision measures the signal to noise ratio.A ratio >4 is required and confirms the appropriate relationship between the actual and predicted results.In the present study, this measurement was found to be 26.348.The effect of key parameters was investigated through ANOVA analysis.The results indicated that parameters pH, Cat.amount, DNP Con., US Power, pH 2 , US Power [2], pH*Cat.amount, pH* DNP Con., pH* US Power, Cat.amount* DNP Con., Cat.amount* US Power were found to be have significant impacts on DNP degradation.The values of the Effect factor ranked in order: US Power > Cat.amount > pH > DNP Con., represent US Power and DNP Con.as the highest and lowest influencing parameters on DNP degradation.Lack of fit as another important indicator to evaluate the model was found to be insignificant (6.3612e-01), representing the precision and adequacy of the model.
Investigating the normality of the studied data and their residuals are two important assumptions for using the statistical model.The normal probability diagram represents the normal distribution of data around the mean, and the linear form of this graph shows the normalisation of test data.The normality and residual plots of outputs as validation indicators of the model are shown in Figure S6.R-squared = 0.942 was obtained by a normal probability diagram.The residual distribution plot confirms the normality of the data, whilst the p-value in the analysis of the linear normality of the residues is close to zero, meaning that the residues are normal.

Prediction and optimisation of the operating parameters
Table S1 and Figure S7(a and b) show the actual and predicted results of 21 experimental designed runs for DNP degradation by Design-Expert (RSM) software.A strong correlation was observed between the actual and predicted values, indicating the ability of applied RSM-CCD model to predict the results.As pointed out in the figure, R-square (0.911) obtained from the regression model was very close to 1 and low values of SSE and RMSE (2.491 and 1.145, respectively) statistically confirm that the selected model can predict the test data in the best way.
For further research, desirability function (DF) was employed to optimise the influencing factors on DNP degradation.To optimise DNP degradation, all probable errors were reduced as much as possible and the range of all influencing parameters and responses were considered at in range and maximum level, respectively.The findings obtained from DNP degradation are summarised in Table 1.According to results, pH = 4.47, Cat.amount = 0.164 g/L, DNP Con.= 5.964 mg/L and US Power = 26.576W/L were found to be optimum conditions for DNP degradation by the FTU system (see Figure S8 and Figure S9).Another parameter that is used to evaluate the best degradation conditions is the desirability factor.In the present study, this factor was very close to 1 at the optimum values of the parameters, indicating conditions are favourable for obtaining the highest DNP removal.Then, in order to ensure the reliability of model performance under optimum conditions, the results of process optimisation were repeated five times in laboratory scale (real conditions).As listed in Table 1, the lab results were found to be 88.34% ± 2.28.The good consistency of the DF result with real conditions (91.45 ± 1.45) and its higher accuracy revealed that applied model can excellently optimise the degradation process.Accordingly, the result recorded by DF was used to continue the research.According to the mentioned figures, it can be found that after 30 min, DNP degradation efficiency shows an almost constant trend, therefore, a 30-min time trial was considered in the following steps.DNP removal efficiencies by US were very low after 45 min (35.4%),showing that DNP cannot be degraded enough by US alone.However, higher removal efficiencies were observed when US was coupled with each other's i.e.TiO 2 , Fe 3 O 4 , Fe 3 O 4 @TiO 2 .These higher removal efficiencies can be explained by the fact that formation of reactive species in binary and ternary systems plays supreme roles in degradation of organic compounds through oxidant catalysis process [37].The removal efficiency recorded under the US-TiO 2 system (68%) was more than US alone.The removal efficiency under the US-Fe 3 O 4 system reached 61.2%, which is mainly due to the Fenton reaction.However, when DNP was exposed to the FTU system, the removal efficiency increased significantly to 90%.The most important reason for the better performance of FTU system compare other processes can be due to greater catalytic activity of Fe 3 O 4 @TiO 2 in the creation of the cavitation phenomenon and reactive species in the presence of ultrasound.Furthermore, the 'hot spot' and 'sonoluminescence' mechanisms can denote the enhanced degradation of DNP under FTU system towards the US alone.The kinetic study of DNP degradation over various systems i.e.US, US-TiO 2 , US-Fe 3 O 4 and FTU were obtained by the first order model (-ln (C/C0) Vs. time) and their results depicted in Figure 2(c).The results showed that all studied systems are in good agreement with the first-order kinetic model with R 2 > 0.9 and this model can describe the experimental data well.On the other side, the values of Kobs (degradation constant) in different systems are calculated and shown in Figure 2(d).As seen, the rate constant for the removal of DNP under FTU system was 0.0700 min −1 while these values were 0.0130, 0.0369 and 0.0304 min −1 for US, US-TiO 2 and US-Fe 3 O 4 , respectively.These findings presented that the DNP degradation rate constant using FTU system was 5.384, 1.897 and 2.302 times higher than of US, US-TiO 2 , US-Fe 3 O 4 processes, respectively.These results prove that coupling US and Fe 3 O 4 @TiO 2 can be considered as decisive technique taking advantage of high potential in the reduction of DNP toxicity in comparison with other studied systems, confirming the synergistic effect between applied agents in this process [20].

Mineralisation and biodegradability
TOC and COD analyses were performed to determine the mineralisation degree of DNP by the FTU system under optimised conditions (Figure 3(a)).DNP degradation efficiency was found to be 88.34%, but the amount of COD and TOC decreased by 73% and 64.2%, after 30 min of reaction.Although a significant amount of DNP under the FTU system decomposed quickly within about 30.0 min, the reduction of COD and TOC was slow and seemed to take a few hours.This phenomenon is ascribed to the fact that intermediates or byproducts such as aliphatic and aldehyde compounds which are emerged during the degradation process are more difficult to further mineralise than DNP molecules as main compounds, and complete mineralisation may proceed at a much slower reaction rate.As such, the consumption of free radicals by DNP and other products resulting from degradation led to the removal of TOC and COD being less than the degradation efficiency (%) of DNP under the same experimental conditions.Next, average oxidation state (AOS) and carbon oxidation state (COS) were investigated to determine the biodegradability of DNP solution by the FTU system.AOS and COS values before and after DNP degradation are shown in Figure 3(a).The values of AOS and COS parameters have been increased from 2.5 to 3.29 and 2.55 to 3.74 during 30 min of reaction.The results indicate that the FTU system can improve the bioavailability of DNP.Furthermore, this observation revealed the adequate oxidation of intermediates to biocompatible aliphatic compounds, so the FTU system can provide promising conditions for the biological process.

Reusability and stability of composite
To investigate the stability and economic feasibility of Fe 3 O 4 @TiO 2 in the FTU system, the catalyst was recycled without any physical and chemical modifications for five cycles.In the end of each run, the Fe 3 O 4 @TiO 2 was collected, dried at 90°C for 60 min, and then employed for the succeeding cycle.As observed from Figure 3(b), in the first run and considering optimum conditions, the degradation (%) and leaching of Ti and Fe were 88.34%, 0.03 and 0.158 mg/L, respectively.The degradation efficiencies experienced small changes between the second and third runs from 86.2458% to 83.2995, and reached 79.3711 and 77.4069% in fourth and fifth cycles, respectively, representing a small decline in the performance of the FTU system under recycled times.This minimal drop can be attributed to either Fe 3 O 4 @TiO 2 mass loss on the catalyst surface or a reduction in the adsorption capacity of the catalyst.The leached titanium and iron concentrations were 0.024 and 0.147 mg/L, 0.019 and 0.133 mg/L, and 0.015 and 0.123 mg/L in the second, third and fourth cycles, which dropped to 0.006 and 0.11 mg/L in the fifth cycle.Leaching iron and titanium in small amount in the repeated reactions is rationale to support the high stability and durability of Fe 3 O 4 @TiO 2 after 150 min of reaction, confirming the high efficiency of this system in treating DNP.

Quenching agents
To ascertain the dominant radical in the FTU system, quenching (scavenging) experiments were performed in optimised condition.Methanol, benzoquinone, potassium iodide and tert-Butyl alcohol, in 400 mM concentrations, were used as radical scavengers to quench the SO 4 •-, O 2 •-, h + , and HO • radicals, respectively [38].It can be seen from Figure 3(c) that, with the addition of benzoquinone (BQ) and tert-Butyl alcohol (TBA), degradation efficiency showed a significant decrease compared to potassium iodide (KI), while the addition of methanol (MEOH) had no effect on degradation efficiency.These indicated that O 2 •-and OH • participate more in the degradation process.By comparing the effect of quenching agents on reducing efficiency, it can be found that BQ offers an inhibiting effect much stronger compared to TBA.These results clearly confirmed that O 2 •-radicals are dominant radical species in the FTU system and play more important roles in the degradation of DNP.

EPR studies
EPR spectroscopy with 5, 5-dimethyl-1-pyrroline N-oxide (DMPO) as a spin-trapping agent was employed so as to further confirm the generation of O 2 •-and OH • species.•-spin adducts were also detected in the system.
These suggested that both of O 2 •-and OH • have a serious role in the destruction of DNP.
However, the spectrum intensity of DMPO-O 2 •-was stronger than that of DMPO-OH • .This phenomenon is explained by the fact that the DMPO-OH • signal diminished quickly over time due to the high reaction rate [39].

Degradation mechanism
Based on the obtained results through quenching (scavenging) and EPR experiments, a hypothesis was proposed to explain the type of mechanism involved in the degradation of DNP are as follows: Acoustic cavitation's, which produced from a liquid in the presence of US irradiation, can generate short-lived (lifetimes of a few microseconds), localised 'hot spots' along with temperatures of roughly 5000°C and pressures of about 500 atmospheres in aqueous solution [40].The extreme conditions will lead to the cleavage of dioxygen and water molecules to produce hydroxyl radicals (OH • ), other radical species, and H 2 O 2 , that can attack and oxidise the DNP molecules [41] (Equations (21)(22)(23)(24)).
Additionally, the presence of Fe 3 O 4 may be the other reason.Fe 3 O 4 provides Fe(II) for the reaction system.H 2 O 2 is generated through Equation ( 23), followed by Equations (1-4), which could represent the degradation path in the presence of Fe(II) and H 2 O 2 .So, a potential Fenton-like system is established without additional H 2 O 2 .The Fenton reaction destroys DNP in a mixed solution containing H 2 O 2 and Fe(II) [41].Indeed, Fe(II) is oxidised by H 2 O 2 molecules, and hydroxyl radicals and other oxidising groups which can attack organic substrates are then generated [42].Moreover, a small amount of Fe 3 O 4 can promote the adsorption of DNP on the surface of the catalyst for the ultrasound process, this conclusion was proposed by Mrowetz et al. [43].Aside from the above items, Hasan et al. [44] reported that surface adsorbed oxygen(O 2 ) molecules in contact with these photogenerated electrons transformed into superoxide radical anions (O 2 •-) (Equations (4)-) which plays an important role in the destruction of DNP.
Based on the evidences, it can be reported that resulting reactive species i.e.O 2 •-, OH • , and h + have a major role in degradation of DNP.The possible mechanism for degradation of DNP under FTU system is illustrated in Figure S10.

Identification of degradation intermediates and pathway
Based on GC-MS analysis (Figure S11), the main products formed as a result of DNP degradation by the FTU system were included in 2-nitrohydroquinone (2-NHQ), 4-nitrocatechol (4-NCC), catechol (CC), hydroquinone (HQ), benzoquinone (BQ), and a mixture of carboxylic acid.aromatic compounds brings about phenyl ring cleavage and the formation of aliphatic carboxylic acids, including maleic acid (MA) and fumaric acid (FuA), as well.Accordingly MA and FuA were oxidised and converted to oxalic acid (OA), and formic acid (FoA).
According to identified products, the proposed DNP degradation pathway was suggested in Figure 4.

Effect of inorganic anions and humic acid on DNP degradation
The effects of some inorganic anions Cl − , NO 3 − , SO 4 2-and CO 3 2-and humic acid (HA) that commonly found in wastewater on activity of catalyst were investigated.i.e.As such, the effect of these anions with concentrations of 1, 10 and 100 mM on DNP degradation through FTU system was evaluated under optimal operating conditions.The influence of CO 3 2− on the oxidation of DNP is shown in Figure S13.The results reveal that the CO 3 2− in lower concentration (0 to 1 and 10 mM) partially inhibited the oxidation process, However, when CO 3 2− concentration continued to increase (from 10 to 100 mM) cause a large decrease in removal efficiency According to reaction Equation (31), CO 3 2− can react with OH • to form CO 3 •-radicals.As can be seen in the reaction rate constant of the aforementioned equations, the transformation of OH • with high oxidation potential for CO 3 •-with less reactivity occurs at a relatively fast rate (∼10 6 M −1 s −1 ), which subsequently leads to a reduction in the overall oxidation strength of the system [45].
The existence of NO 3 − at different concentrations under the FTU reaction system, did not show any inhibition effect during DNP degradation.Under the influence of ultrasound waves, NO₃‾ could produce OH • in water through reaction Equation (32) as a result more OH •-oxidised DNP.
On the other hand, degradation efficiency in terms of increase in SO 4 2− concentration from 0 to 100 mM, was slightly enhanced.The increase in degradation efficiency can be explained by the generation of highly reactive sulphate (SO 4 •-) radicals through reaction Equations (33 and 34).
The effect of Cl − on contaminant degradation was more complicated.The presence of chloride at all concentrations had a slightly negative effect on DNP degradation, which can be related to the generation of Cl  , which accordingly reduces the overall oxidation strength, so the presence of Cl − did not greatly inhibit DNP degradation, and the higher concentration of Cl − decreased the inhibiting effect of chloride on DNP degradation [46].

DNP degradation by FTU + oxidant systems
The performance of the FTU system in the degradation of DNP was evaluated given various oxidants, including S 2 O 8 2-, H 2 O 2 , IO 4 − at optimum conditions (for a better understanding about the differences in the performance of oxidants, the initial DNP concentration was considered to be 5 mg/L).As observed in Figure S14(a and b), S 2 O 8 . Therefore, the oxidation potential (E 0 ) of the studied oxidants in the degradation of DNP followed the order of S 2 O 8 2->IO 4 − > H 2 O 2 .The rate of DNP degradation is described by fitting the firstorder kinetic model to the data recorded by the FTU + oxidant systems (Figure S14(c-d)).In all systems, a high regression coefficient (R 2 > 0.991) showed that DNP degradation kinetics fitted well with the first-order kinetic model.The rate constant (K obs ) for DNP degradation was calculated as 0.2003 min −1 under the FTU+S 2 O 8 2-system, which is higher compared to the values recorded by FTU+H 2 O 2 and FTU+IO 4 − (0.0846 and 0.1191 min −1 , respectively).
The results confirm that oxidant improves the rate of degradation (about 0.1295 min −1 ) and has a positive effect on the removal efficiency of the FTU system alone.

The actual applications of the FTU system
Different environments including tap water, surface runoff, raw wastewater and secondary sedimentation effluent were employed to investigate the performance and ability of the FTU system in real terms specifications of wastewater are listed in Supplementary data, Table S3.The samples were spiked with ~5 mg/L DNP concentration and the experiments were carried out at optimum conditions.Degradation efficiency was found to be 82% for tap water, 74% for surface runoff, 50.39% and 65% for raw and treated wastewater, but all of them are less efficient than DI-water (Figure S15).Inhibiting or quenching free radicals through the intervention of high TDS with free radicals, and creating a competitive mode between the existence of different ions and/or organic compounds in real conditions with DNP for adsorption and reactive oxidising species are two important reasons for these reductions.

Comparison with previous studies
The degradation efficiency (Re, %) of the proposed system for DNP removal was compared with the obtained results from previous studies over the last decade, and the findings are shown in Table 2.It was observed that FTU system has a proper position among the other methods reported in the literature and it had a good ability to compete with similar systems for treatment of DNP Furthermore, the optimum catalyst dosage and reaction experiment's time also have a lower than the majority of studies conducted so far.This observation confirms the key role of the ultrasonic field in the significant increase of the degradation efficiency and reducing the reaction time and catalyst dosage.
ᵒ for (211), and 70.48 ᵒ for (220) based on JCPDS no.21-1272 indicate the presence of TiO 2 with an anatase tetragonal structure.The occurrence of other peaks is attributed to Fe 3 O 4 cores.Figure S3 (b) shows the thermogravimetric analysis (TGA) curve of the Fe 3 O 4 NPs and the synthesised Fe 3 O 4 @TiO 2 .For Fe 3 O 4 NPs, TGA curve indicates three distinguished weight loss phases; first, a steep weight loss in the range of <150°C due to the evaporation of remained water and alcohol from the Fe 3 O 4 NPs surface; the second stage a weight loss in the range of 150°C < T < 350°C, owing to the decomposition of polyethylene glycol as organic substances.The third weight loss between 350 to 750°C is due to the complete dehydration of polyethylene glycol.Fe 3 O 4 @TiO 2 also indicated three distinct weight loss stages.Similar to Fe 3 O 4 NPs, the first and sharp weight loss at 30 to 200°C is attributed to evaporation of residual water and alcohol adsorbed on the samples surface.A minor weight loss by increasing temperature over a wide range (100°C and 500°C) is due to the crystallisation of TiO 2 NPs, indicating coating a thin layer of TiO 2 on the Fe 3 O 4 NPs surface.The last weight loss at >500°C is complete dehydration of polyethylene glycol.These results reflect two outcomes: (1) the high thermal stability of Fe 3 O 4 NPs due to the slight weight loss and (2) the greater thermodynamic stability of the Fe 3 O 4 @TiO 2 compared to Fe 3 O 4 NPs because of the presence of TiO 2 as a shell in Fe 3 O 4 @TiO 2 structure.The N 2 adsorption-desorption isotherms of the TiO 2 , Fe 3 O 4 NPs and synthesised Fe 3 O 4 @TiO 2 are shown in Figure

Figure 1 .
Figure 1.4D response contour plots of variables effect on DNP degradation.

Figure 2 (
Figure2(a,b) represents the efficiency of the FTU system in comparison with US, US-Fe 3 O 4 and US-TiO 2 processes under the optimised experimental conditions.According to the mentioned figures, it can be found that after 30 min, DNP degradation efficiency shows an almost constant trend, therefore, a 30-min time trial was considered in the following steps.DNP removal efficiencies by US were very low after 45 min (35.4%),showing that DNP cannot be degraded enough by US alone.However, higher removal efficiencies were observed when US was coupled with each other's i.e.TiO 2 , Fe 3 O 4 , Fe 3 O 4 @TiO 2 .These higher removal efficiencies can be explained by the fact that formation of reactive species in binary and ternary systems plays supreme roles in degradation of organic compounds through oxidant catalysis process[37].The removal efficiency recorded under the US-TiO 2 system (68%) was more than US alone.The removal efficiency under the US-Fe 3 O 4 system reached 61.2%, which is mainly due to the Fenton reaction.However, when DNP was exposed to the FTU system, the removal efficiency increased significantly to 90%.The most important reason for the better performance of FTU system compare other processes can be due to greater catalytic activity of Fe 3 O 4 @TiO 2 in the creation of the cavitation phenomenon and reactive species in the presence of ultrasound.Furthermore, the 'hot spot' and 'sonoluminescence' mechanisms can denote the enhanced degradation of DNP under FTU system towards the US alone.The kinetic study of DNP degradation over various systems i.e.US, US-TiO 2 , US-Fe 3 O 4 and FTU were obtained by the first order model (-ln (C/C0) Vs. time) and their results depicted in Figure2(c).The results showed that all studied systems are in good agreement with the first-order kinetic model with R 2 > 0.9 and this model can describe the experimental data well.On the other side, the values of Kobs (degradation constant) in different systems are calculated and shown in Figure2(d).As seen, the rate constant for the removal of DNP under FTU system was 0.0700 min −1 while these values were 0.0130, 0.0369 and 0.0304 min −1 for US, US-TiO 2 and US-Fe 3 O 4 , respectively.These findings presented that the DNP degradation rate constant using FTU system was 5.384, 1.897 and 2.302 times higher than of US, US-TiO 2 , US-Fe 3 O 4 processes, respectively.These results prove that coupling US and Fe 3 O 4 @TiO 2 can be considered as decisive technique taking advantage of high potential in the reduction of DNP toxicity in comparison with other studied systems, confirming the synergistic effect between applied agents in this process[20].

Figure 2 .
Figure 2. Degradation efficiency, kinetic study and rate constant of DNP degradation over various systems.

Figure 3 .
Figure 3.The changes of DNP, TOC and COD (a), recycling and leaching tests (b) radical scavenger quenching experiment to determine the dominant radical (c) and EPR spectra in the FTU system under DMPO = 0.1 M (d).

Figure 3
(d)    shows the characteristic signal peak of DMPO-OH • and DMPO-O 2•-in FTU system.The characteristic quartet peaks of DMPO-OH • adducts with an intensity ratio of 1:2:2:1 appeared under US irradiation, demonstrating that OH • is generated in the degradation process.The ESR signals of DMPO-O 2 Figure S12 shows the trends of DNP concentrations and their intermediates during different reaction times in the FTU system.As observed, the concentration of DNP decreased by about 88% in 30 min and then continued to decrease gradually.The first degradation intermediate produced is obtained by hydroxyl radicals attacking DNP.In fact, an OH radical joins the DNP molecules and transfers them into 2-NHQ and 4-NCC.Subsequently, hydroxyl groups break the 4-NCC and lead to the formation of CC.In the second possible path, HQ was generated after the nitrate (NO 3 − ) atom was removed from 2-NHQ in the next step.Some studies reported that the nitrate atom located in the para-position on the aromatic ring of DNP was directly attacked by OH• through steric effects.Hence, HQ could be produced by the hydroxylation of DNP.This hydroquinone was subsequently dehydrogenated to give BQ by the electrophilic attack of hydroxyl radicals (OH • ), direct sono-oxidation of DNP by holes h VB + , or direct oxidation of hydroquinone by superoxide radicals (O 2 •-) in the solution.As expected, further oxidation of

Figure 4 .
Figure 4.The suggested pathway for the degradation of DNP in the FTU system.

Table 1 .
Analysis of variance (ANOVA) for CCD modelling and process optimisation results.
2-exhibited higher ability in the degradation of DNP than the others.The mechanism of S 2 O 8 2-in the degradation of DNP is as follows: Possible reactions for H 2 O 2 and IO 4 − , are also shown in the following equations: It was found that HO � andIO � 3 were the predominant radicals in the FTU+H 2 O 2 and FTU+IO 4 − systems, respectively, but OH � and SO �À 4 radicals according to the Equations (43-47) are generated in the FTU+S 2 O 8 2-system.In fact, better performance of sulphate-based oxidants (100% at <20 min reaction time) can be attributed to the ability of S 2 O 8 2-in the simultaneous production of sulphate and hydroxyl radicals.Evidence suggests that FTU was more successful in the activation of S 2 O 8 2-than IO 4 − and H 2 O 2 .On the other hand, H 2 O 2 showed the weakest performance among oxidants, which could be due to the hole scavenging effect of H 2 O 2 in the FTU+H 2 O 2 system as H 2 O 2

Table 2 .
Comparison of DNP degradation reported in the literature with the proposed systems.