The high catalytic activity and strong stability of 3%Fe/AC catalysts for catalytic wet peroxide oxidation of m-cresol:The role of surface functional groups and FeOx particles

Peiwei Han,Chunhua Xu,Yamin Wang,Chenglin Sun,Huangzhao Wei,Haibo Jin,Ying Zhao,*,Lei Ma,*

1 Beijing Key Laboratory of Fuels Cleaning and Advanced Catalytic Emission Reduction Technology/College of Chemical Engineering,Beijing Institute of Petrochemical Technology,Beijing 102617,China

2 Shandong Key Laboratory of Water Pollution Control and Resource Reuse,School of Environmental Science and Engineering,Shandong University,Qingdao 266237,China

3 Dalian Institute of Chemical Physics,Chinese Academy of Sciences,Dalian 116023,China

Keywords:Catalytic wet peroxide oxidation Fe/AC catalyst Surface functional groups Reaction mechanism

ABSTRACT FeOx supported on activated carbon(AC)has been shown to be an ideal catalyst for catalytic wet peroxide oxidation (CWPO) due to its high CWPO reaction activity and stability.Although there have been some studies on the mechanism of Fe/AC catalysis in CWPO,the specific contribution of each component(surface oxygen groups and FeOx on AC)inside an Fe/AC catalyst and their corresponding reaction mechanism remain unclear,and the reaction stability of CWPO catalysts has rarely been discussed.Then the optimal CWPO catalyst in our laboratory,3%Fe/AC,was selected.(1) By removing certain components on the AC through heat treatment,its contribution to the reaction and the corresponding reaction mechanism were investigated.With the aid of temperature-programmed desorption–mass spectrometry (TPD–MS) and the CWPO reaction,the normalized catalytic contributions of components were shown to be:37.3%(carboxylic groups),5.3% (anhydride),19.3% (ether/hydroxyl),-71.4% (carbonyl groups) and 100% (FeOx),respectively.DFT calculation and EPR analysis confirmed that carboxylic groups and Fe2O3 are able to activate the H2O2 to generate ·OH.(2) The catalysts at were characterized at different reaction times(0 h,450 h,900 h,1350 h,and 1800 h)by TPD–MS and M?ssbauer spectroscopy.Results suggested that the number of carboxylic goups gradually increased and the size of paramagnetic Fe2O3 particle crystallites gradually increased as the reactions progressed.The occurrence of strong interactions between metal oxides and AC was also confirmed.Due to these effects,the strong stability of 3%Fe/AC was further improved.Therefore,the reasons for the high activity and strong stability of 3%Fe/AC in CWPO were clearly shown.We believe that this work provides an idea of the removal of cresols from wastewater into the introduction to show the potential applications of CWPO.

1.Introduction

Environmental pollution has long been a topic of widespread concern.One important consideration is industrial wastewater,which is strictly controlled due to the potential for contamination with toxic pollutants [1].One such group of pollutants is cresols,which are often found in industrial wastewater due to their wide range of industrial applications.Cresols have direct or potentially hazardous impacts on humans and so contaminated wastewater must be thoroughly treated to remove them [2–4].Advanced oxidation technologies exhibit excellent performance in the treatment of pollutants in organic wastewater and can be adopted to remove persistent organic pollutants that are not degraded by traditional treatments.Among them,heterogeneous catalytic wet peroxide oxidation (CWPO) has gained particular attention due to its high efficiency and comparatively low cost [5,6].

Currently,two issues limit the industrial use of CWPO.The CWPO reaction takes place under strongly acidic conditions (typically pH 3.0),so considerable quantities of acid are required to adjust the pH of wastewater to 3.0,which may increase the operating costs.Furthermore,under such strongly acidic conditions,corrosion of the catalyst and leaching of active components are inevitable,leading to a shortened service life of the catalyst [7,8].To overcome the above problems,one solution that has been proposed is to develop high-performance catalysts that can extend the optimal reaction conditions to neutral conditions [9].This would not only reduce the operating costs,but also reduce the leaching of active ingredients,and would improve catalyst stability.

Actived carbon (AC) is commonly used as an highly effective catalyst or as a catalytic support.Rochaet al.[10] stated that the catalytic performance of AC is highly dependent on the surface functional groups(SOGs)in the catalytic wet air oxidation(CWAO)process.In CWPO,variations in SOG types and abundances also play an important role in pollutant degradation [11,12].Meanwhile,when AC is used as a catalytic support,its well-developed porous structure is beneficial,since the catalyst’s active components are ideally distributed on the AC surface [13].

Based on the fact that AC is commonly used as a catalyst,we developed an AC-supported iron catalyst(3%Fe/AC),which exhibits a high level of catalytic activity even under neutral conditions.The leaching of iron oxide,the active component,was minimal under neutral conditions.Industrial applications have already been identified for this catalyst.As shown in Fig.S2,a pilot study in the laboratory has also demonstrated that the catalyst possesses outstanding capacity for the catalytic degradation ofm-cresol by CWPO [14].

A number of studies have shown that the high catalytic capacity of Fe/AC is related to the surface properties (namely,SOGs) and iron oxide distribution in AC.However,the specific contribution of each of these two aspects to the degradation ofm-cresol is unknown.In addition,the reaction stability of CWPO catalysts has rarely been discussed.Therefore,the catalytic mechanism of iron-supported AC was investigated in this study.In addition,the performance of AC-based materials in catalytic degradation by CWPO was evaluated in a batch reactor to demonstrate the separate degradation contributions of SOGs and active components to activated carbon.

2.Material and Methods

2.1.Catalyst preparation

Commercial activated carbon derived from coconut shell (product code TM-L106)was purchased from Anshan Senxin Activated Carbon Factory.This untreated activated carbon is designated as AC.Samples of the AC were subjected to heat treatment under a flow of N2gas at 200 ml?min-1for 90 min at 553 K,773 K,1023 K,and 1173 K respectively,and then allowed to cool to room temperature;the treated samples were designated AC-553,AC-773,AC-1023 and AC-1173,respectively.

To prepare the iron-loaded catalysts,3%Fe/AC was preparedviathe equivalent-volume impregnation method.Subsequently,the samples were dried at 383 K for 3 h in an oven,and then calcined at 723 K with a heating rate of 3 K?min-1under a continuous flow of N2gas for 3 h to obtain the final catalysts.

2.2.Catalyst characterization

The surface appearance of each AC sample was analyzed by scanning electron microscopy (SEM,Quanta 200F,FEI Company).The SEM and transmission electron microscopy (TEM) of 3%Fe/AC was conducted on related equipment (Quanta 200F,FEI Company,FEI TECNAI F20 G2)supported by focused ion beam(FIB).The function of the FIB was to strip the surface carbon atoms in order to better observe the surface topography of 3%Fe/AC.

The specific surface area of each sample was obtained by the Brunauer-Emmett-Teller (BET) method using QUADRASORB SI4 apparatus produced by Quantachrome Instruments.The SOGs were analyzed by temperature-programmed desorption coupled with TPD–MS.The elemental content was determined by X-ray fluorescence analysis(XRF)using Magix 601 equipment produced by PANalytical.The carbon content was calculated from the difference in mass before and after the calcination of catalysts.The AC and 3%Fe/AC were characterized by X-ray diffraction (XRD) (X’pert Pro machine with a Cu Kα target and scan angle of 10–80 degrees) to identify the crystal structure.57Fe M?ssbauer spectroscopy at room temperature (Topolpgic 500A,Japan) was used as a local microstructure probe for the iron species in catalysts.The pH slurry of the samples was measured with a 0.5 g sample in 10 ml distilled water,with continuous stirring until the pH of the slurry was stabilized.

A ReactIR15 instrument manufactured by METTLER TOLEDO was used for in-situ monitoring utilizing an attenuated total reflection (ATR) sensor at the tip of the probe to determine changes in the composition of the reaction through observing changes in characteristic peaks.

An EPR(electron paramagnetic resonance)test for·OH was conducted as follows.200 μl of 1 mol?L-1DMPO solution was added into a mixture of 200 μl sample solution with a catalyst concentration of 1000 mg?L-1.Next,500 μl of buffer solution (pH 4,2 mmol?L-1HAC/NaAC)and 100 μl of 30%hydrogen peroxide were also added into the mixture.After 2 min of reaction time,the sample was subjected to the·OH test.

All calculations were performed with spin-polarized density functional theory under the generalized gradient approximation with the Perdew-Burke-Ernzerhof functional [15–17].An energy cut-off of 380 eV was adopted,and the k-space was sampled with a grid of 1×1×1 under the Monkhorst-Pack scheme,due to large supercell sizes.The van der Waals interaction was allowed for using the DFT-D3scheme [18,19].The molecular structures ofmcresol and the various intermediates were optimized using quantum chemistry methods by application of density functional theory(DFT).The detailed DFT calculation scheme for the quantum chemical descriptor is described below.The B3LYP method was used with large basis sets(for example 6-311++G(2df,2p))on the reactants.In all calculations,the CPCM model was used to address the solvent effect of water.The calculations were performed as implemented in the Gaussian 09 program package using an Intel(R)Xeon(R) E5520/2.27 GHz processor.In addition,the quantum chemical descriptorEHOMOwas generated.Gaussview 5.0 was used to visualization analysis the computed structure and electronic cloud model.

2.3.Reaction procedures

The CWPO batch reaction ofm-cresol over the abovementioned carbon samples was performed in a 1000 ml glass conical beaker fitted with a stopper.This beaker was shaken in a thermostatic bath with a constant stirring velocity of approximately 150 r?min-1.Two experiments were carried simultaneously:(i)adsorption ofm-cresol and (ii) oxidation ofm-cresolviaheterogeneous CWPO.For the adsorption,500 ml of 100 mg?L-1m-cresol was mixed with 0.3 g?L-1of the sample powder at 303.15 K.The initial aqueous pH was set to (7.0 ± 0.1) and adjusted with 3 mol?L-1H2SO4or NaOH.For the oxidation,541 mg?L-1H2O2was also introduced into the system,which theoretically corresponded to the total mineralization of 100 mg?L-1m-cresol.Samples were taken at pre-determined time intervals and filtered with a 0.45 μm film to remove all catalyst particles.

The continuous experiments were carried out in a glass upflow fixed bed reactor with an inside diameter of 21 mm and a height of 170 mm that was equipped with an electronic diaphragmatic metering pump(Fig.S1).The test was performed at a temperature of 303.15 K and an LHSV of 1 h-1.The initial pH value of themcresol solution was approximately 7.0.The catalysts (50 ml) were saturated withm-cresol prior to being placed into the CWPO reaction.Subsequently,541 mg?L-1H2O2was fed into the reactor along with 100 mg?L-1m-cresol,and the catalytic experiments were initiated as 0 h.In the continuous reaction,the catalyst was directly placed into the reactor with an irregular sheet morphology.

2.4.Analytical methods

High-performance liquid chromatography (HPLC) was employed to analyze them-cresol concentration using an HPLC-p 1201 (Dalian Elite Analytic Instruments Co,Ltd).The total organic carbon (TOC) content of the solution was measured using a TOCVCPNTOC analyzer (Shimadzu,Japan),while the and pH was measured with a PHS-3C pH meter(Rex Instrument Factory,Shanghai,China),respectively.

The intermediates in the treated solutions were identified by gas chromatography–mass spectrometry(GC–MS)using an Agilent 7890A with a FFAP capillary column of size 30 m × 0.25 mm × 0.25 μm,and a TraceMS2000(Finnigan,USA)with a scan range ofm/z30–300.

3.Results and Discussion

3.1.Characterization of 3%Fe/AC

Fig.1 shows the morphology of the pristine catalyst obtained by SEM and TEM.The AC shows a regular appearance and retains the fibrous structure of the plants from which it is derived,with obvious macropores (Fig.1(a)).For 3%Fe/AC,the active constituents(ferric oxides) are mainly stored in the AC macropores (Fig.1(b)).This may be beneficial in controlling ferric oxide leaching.FIB(Fig.1(c) and (d)) was used to obtain an extremely thin sample of 3%Fe/AC for TEM analysis (Fig.1(d)–(f)).To prepare the 3%Fe/AC samples using FIB,first a nanomanipulator was used to extract a portion of 3%Fe/AC sample (the area enclosed by white dotted line in Fig.1(b)) from the entire sample (Fig.1(b)).Next,the stripped sample shown in Fig.1(c)was carefully cut into very thin slices(Fig.1(d)).The TEM image(Fig.1(e)and(f))shows dispersed nanoparticles with an average size of～2–3 nm and a preferential crystallographic orientation corresponding to the (311) plane of crystalline γ-Fe2O3,which was further substantiated by XRD analysis of γ-Fe2O3particles (Fig.S3(a)).Meanwhile,Fig.1(e) shows that ferric oxides are mainly present in the forepart of the macropores that favor the diffusion of pollutant molecules on the surface of γ-Fe2O3particles.

Based on the XRD diagrams (Fig.S3(a)),since the diffraction spectrum corresponds to the structure of ferric oxide,the modified crystallographic information file (CIF) and the octahedral stereoscopic structure image of Fe2O3(Fig.S3(b))were obtained.The fitting cell parameters werea=8.287 pm,b=8.287 pm,c=25.144 pm,α=90°,β=90°,and γ=90°.

The elemental compositions of the 3%Fe/AC and the AC were measured using the XRF technique (Fig.S4).The AC was shown to have a maximum carbon content of 97.1% (mass),while the maximum carbon content of iron was 0.07%(mass).The XRF result of 3%Fe/AC demonstrated that 3.63% iron was successfully loaded onto the AC.

3.2.The reaction mechanism and contribution of SOGs and Fe2O3 to 3%Fe/AC

The surface chemistry of AC was characterized using the TPD–MS technique in helium flow,a common thermal analysis method for detecting SOGs on the surface of carbon materials [20,21].As shown in Fig.2(a),based on the information obtained from the literature,the CO and CO2spectra of AC clearly show peaks corresponding to carboxylic group (500–530 K) on the CO2curve,anhydride group (670–700 K) on both the CO and CO2curves,hydroxyl/ether groups on the CO curve,and carbonyl group(1100 K) on the CO curve.A number of functional groups,which play crucial roles in the decomposition of oxidants in CWPO,have been studied in detail[22,23].By removing specific components on the AC through heat treatment,its contribution to the reaction and the corresponding reaction mechanism could be investigated.The different AC samples possessed different SOGs,as shown in Fig.2(b).It was possible to eliminate the carboxylic groups,anhydride,ether/hydroxyl groups,and carbonyls on the AC at different treatment temperatures.

As reported by Noorjahanet al.and Huet al.[9,24],the adsorption capacity of AC has aa significant effect on the elimination of pollutants.The adsorption capacity of the samples (AC,AC-573,AC-773,AC-1023,AC-1173 and 3%Fe/AC) is shown in Fig.3(a),where after an absorption time of 5 h,～47.8% of them-cresol was adsorbed into the AC,which has a developed microporous structure and very high adsorption capacity.The conversion ofm-cresol and the removal of TOC by absorption occurred at almost the same rate,indicating that while the adsorption equilibrium ofm-cresol can be achieved instantly,m-cresol is not significantly decomposed or converted into other organics under the experimental conditions without addition of H2O2[25,26].The results of this study were in agreement with those obtained in the literature [27],indicating that slight changes in textural properties do not have a significant effect on adsorption or catalytic capacity.

Fig.1.(a) SEM image of AC and (b–g) SEM and TEM images of 3%Fe/AC.

Fig.2.(a)TPD–MS analysis of AC and(b)SOGs of samples AC,AC-573,AC-773,AC-1023 and AC-1173.

As mentioned above,SOGs on the AC surface contributed to the major catalytic capacity of carbon catalysts.Subsequently,batch CWPO experiments were run in parallel with the adsorption experiments.In this study,the real value ofm-cresol decomposition by CWPO,which was obtained by subtracting the adsorption value,was set as the grey region shown in Fig.3(b).With the addition of H2O2(stoichiometric amount),the total conversion ofm-cresol by AC exceeded 72.3% and the total TOC removal reached 62.0%after a reaction time of 5 h.As a result,the contributions of CWPO degradation were 23.4%form-cresol conversion and 15.0%for TOC removal.Following calcination at 553 and 773 K,the catalytic contributions of samples AC-553 and AC-773 form-cresol conversion decreased by～10.1% and 11.5%,respectively,and the corresponding TOC removal contributions decreased by 3.7%and 6.4%.Therefore,the results indicated that both carboxylic and anhydride groups were beneficial for CWPO degradation ofm-cresol under the governing conditions.The presence of these acid groups with electron-withdrawing capacity stabilized the electrons on the carbon surface,and thus reduced the activity of electron-rich regions.Therefore,the ability of catalysts to convert H2O2to radical species was restricted in AC-553[28].In AC-1023,the conversion of pollutants(designated as the catalytic contribution)was less than 17.1%of the original,untreated AC,but slightly above that of AC-773(11.9%).The results indicated that the two SOGs (ether/hydroxyl groups) adversely affected organic matter removal.For sample AC-1173,which had the highest adsorption capacity,93.7%conversion and 81.5%TOC removal were achieved after a reaction time of 5 h.Excluding the adsorption contribution,for AC-1173,the catalytic contribution was 36.5% for conversion and 26.2% for TOC removal,giving a clear indication that the carbonyl groups do not have a role in this reaction.Table S1 shows several structural defects and specific surface areas for excluding the properties that may possibly contribute to the AC activity.However,no direct relation was demonstrated between the structural properties and catalytic activity;this suggests that the activity of thermally treated samples in the CWPO reaction is mainly produced by the SOGs.

To further understand how each type of SOG contributed to activity of the AC,a summary is provided in Fig.3(c) and(d).Heat treatment at different AC roasting temperatures gradually removed the corresponding functional groups and yielded different the AC samples (AC-553,AC-773,AC-1023 and AC-1173) (Fig.2(b)).The differences inm-cresol conversion between the two AC samples near the calcination temperature was divided by the conversion from 3%Fe/AC(51.0%)to calculate the relative contribution of each functional group to the catalytic activity of the AC.Next,the normalized catalytic contributions of all functional groups and Fe2O3were calculated(Fig.3(c)).Similarly,the catalytic contributions of the different functional groups and of Fe2O3to TOC removal were also calculated(Fig.3(d)).The relative values of catalytic contributions of the carboxylic groups,anhydride group,and ether/hydroxyl groups of AC were 37.3%,5.3% and 19.3%,respectively,while the carbonyl groups showed significant negative effects (-71.4%).As shown in Table S1,the results were partially consistent with the contribution of SOGs to pHslurryof the samples.The acidic SOGs favored the generation of ·OH in the CWPO process.For metal ions (3%Fe/AC),it was possible to achievemcresol conversion of 54.5%,after allowing for the adsorption ofm-cresol.Therefore,metal oxides played an important role in the catalytic contribution.For TOC removal(Fig.3(d)),the contribution of certain SOGs of AC showed the same trend asm-cresol conversion (Fig.3(c)).

In addition,unlike the adsorption process,catalytic oxidation proceeded more slowly.A strong correlation was found between the presence of SOGs and catalytic activity during liquid-phase oxidation processes[29].Therefore,it may be possible to use AC samples in long-term experiments.

To investigate the reaction mechanism,EPR analysis was conducted.According to the intensity analysis of the six EPR spectrum signals (Fig.4),the hydroxyl radicals were not produced in the hydrogen peroxide,3%Fe/AC,or AC system alone.Of the other three oxidation systems,the Fe2++H2O2catalytic system exhibited the weakest signal intensity,followed by the AC+H2O2catalytic system,while the 3%Fe/SAC+H2O2catalytic system exhibited the strongest signal.This indicates that in the presence of H2O2,the catalytic activity sequence is 3%Fe/SAC >SAC >Fe2+.In the presence of iron ions,the reduction of H2O2by Fe(II) leads to the generation of ·OH radicals,which can then react with another H2O2molecule to produce the peroxide radical (·OOH) and Fe(III)in situ simultaneously.Finally,the ·OOH radical transfers one electron to the Fe(III)species and regenerates Fe(II)in a cyclic reaction,leading to the continuous presence of active metals for the generation of ·OH radicals.H2O2itself has no catalytic activity.

3.3.The DFT calculations section

Fig.3.Batch experiment reaction results of m-cresol conversion (a) by ACs and 3%Fe/AC catalysts after 5 h,(b) the contribution of adsorption and CWPO on the m-cresol conversion(column C)and TOC removal(column T),(reaction conditions:Cm-cresol=100 mg?L-1,CH2O2=541 mg?L-1,Ccat=0.3 g?L-1,T=343 K,initial aqueous pH=7.0)(c)contribution of SOGs and metal oxides to AC and 3%Fe/AC catalysts for m-cresol conversion,and (d) TOC removal.

The optimized geometries of H2O2adsorbed on various sites are shown in Fig.5.According to this,H2O2was stabilized mainly by hydrogen bonds(HBs),which are labelled with green dashed lines.To compare the stability of these geometries,a reference geometry with H2O2at 10.0 × 10-10m from the substrate was chosen and labelled as S----H2O2,in which H2O2has almost no interaction with the carbon-based structures.Therefore,the total energy of H2O2was regarded as zero.In comparison,the calculated energies of other geometries (shown as A–H) were used to evaluate the adsorption energy (AE) of H2O2as listed in Table S2.A,D and F,which show double HBs.As a result,their AEs were smaller than those of the other five(B,C,E,G and H),with E having the highest AE value.Other than HBs,van der Waals interactions also contribute to H2O2adsorption.This is evidenced by geometry C in which H2O2moves out of the pore without forming any HB,resulting in AE of -0.40 eV.Therefore,H2O2was stabilized by van der Waal interactions and HBs.

Fig.4.EPR spectra of adducts formed by DMPO spin-trapping of ·OH radicals.

Interaction with the substrate leads to a change in the geometry of H2O2.Using the bond lengths of S----H2O2geometry as a reference,the O-H and O-O distances were 0.980 × 10-10m and 1.467 × 10-10m,respectively.After weak adsorptions (B,C,G and H),one of the O-H bonds was slightly enlarged to 0.983 × 10-10m (C),0.985 × 10-10m (B),and 0.995 × 10-10m(G and H).However,in the case of strong adsorptions (A,D and F),the O-H bond enlargement was more significant (more than 1.000 × 10-10m).The E adsorption value was the most unique,as it shortened the O-H bond.Unlike the O-H bond,the O-O bond length was not sensitive to the adsorption strength.Based on the above analysis,substrate interaction may activate the O-H bonds in H2O2,and the carbonyl group (E) may inhibit the activation of H2O2by increasing the bond energy in the O-H bonds.

The carbon structures(A–H)can be used as a reference to evaluate H2O2adsorption on the FeO centers of carbon substrates,whose geometries are shown in Fig.5 with AE of-0.56 eV.Clearly,this was a strong adsorption,which can be further supported by short HB lengths (1.572 × 10-10mvs.1.7–1.9 × 10-10m in Fig.6) and enlarged O-H bonds (1.021 × 10-10m).This suggests that FeO actively adsorbed and activated H2O2.When the carboxyl groups were calculated separately,the results (AE=-0.34 eV)showed that H2O2was not readily activated.

Finally,the adsorption energy equation,AE=E(S-H2O2)–E(S----H2O2)suggests that the lower the AE value,the lower the energy of S-H2O2.In other words,the functional group was more likely to activate H2O2.As shown in Table S2,the carbonyl group (E) did not readily activate H2O2and even inhibited H2O2activation by increasing the bond energy of the O-H bond.However,the AE values of the carboxylic group (F),the hydroxyl group (A) and active Fe component in the Fe/AC catalyst were smaller,indicating that they were the main catalyst sites of H2O2in the CWPO reactions.From this,we can conclude that the carboxylic groups and Fe2O3are able to activate the H2O2to generate ·OH.

Fig.5.Single AC-C=O and AC-FeO adsorption morphology.

3.4.The stability discussion of 3%Fe/AC in CWPO

The above results(Section 3.2)suggest that the removal of pollutants was mainly governed by adsorption and oxidation in the presence AC or 3%Fe/AC.It can further be concluded that carboxylic groups and Fe2O3are able to activate the H2O2to generate·OH.Typically,heterogeneous catalysts in CWPO are unstable in the long term because the aqueous acidic pH and the presence of organic acids as byproducts always result in the leaching of active phases.Therefore,a continuous reaction is necessary to evaluate the stability of the carbon catalyst,and thus,the initial aqueous pH was adjusted to 7.0 to avoid metal loss.Wet impregnation with 3% Fe provided significant efficiency in decomposingm-cresol at circumneutral pH,further enhancing the catalytic activity.As shown in Fig.S2,from the beginning of the experiment up to a reaction time of 1800 h,90%m-cresol conversion was achieved.At the same time,a TOC removal of～30% was also achieved,with no significant decline after an 1800 h reaction time (Fig.S2).Iron leaching was also monitored throughout the reaction process to assess the contribution of homogeneous catalysis.As shown in Fig.S2,in the first 200 h,a very small amount of unstable iron particles remained on the AC surface in the acidic effluent at pH=3.5,followed by a slow decrease in leaching.To estimate the total ion leaching during the reaction over 1800 h,～80% of the Fe species remained on the catalyst surface,and the remaining Fe species were very stable in the following CWPO reaction.This was demonstrated by the subsequent results,which showed no Fe species in the effluent after 1800 h,indicating a strong interaction between iron particles and the AC surface.It was understood that the decomposition ofm-cresol occurred mainly at the catalyst surface,not in solution.Therefore,it was concluded that the catalytic stability of 3%Fe/SAC can meet the requirements of industrial application.

Fe species are known to be active in acidic solutions,leading to an effective Fenton-type reaction by catalysing H2O2[30-34].Therefore,every 450 h,2 ml of 3%Fe/AC was removed from the reaction system and analyzed using TPD–MS and57Fe M?ssbauer spectroscopy.The results are shown in Figs.7 and 8,respectively.The TPD–MS results of the samples after reaction clearly differed from those obtained from the original sample(Fig.7).The amount of carboxylic acids that released CO2at lower temperatures gradually increased with the reactions shown in Fig.7.On the other hand,m-cresol and some aromatic intermediates were adsorbed on the carbon surface from the beginning of the reactions involving CO (about 950 K,Fig.7(a)) and CO2evolution (～925 K,Fig.7(b)).However,their evolution tended to decrease as the reactions proceeded,indicating that these aromatic intermediates were gradually converted to carboxylic acids by the ·OH radicals generated during the oxidation process.As previous discussion in Figs.3–6 showed that carboxylic acid groups were one of the major reaction sites to activate H2O2to generate ·OH,the increase of carboxylic acid groups would ensure the catalytic capacity of 3%Fe/AC in a CWPO reaction over the long-term.

Fig.8 shows the57Fe M?ssbauer spectra of fresh and used catalysts.The corresponding parameters,including isomer shift (IS),electric quadrupole splitting (QS),linewidth (LW) and hyperfine magnetic field at Fe nuclei(H)are listed in Table 1.The M?ssbauer spectrum of 3%Fe/AC consisted of 1) a central doublet with hyperfine parameters of IS=0.33 mm?s-1and QS=0.80 mm?s-1,assigned as superparamagnetic Fe3+species with crystallite size<10 nm,and 2)29.0%sextet(relative area,RA)of larger paramagnetic Fe2O3particles (IS=0.31 mm?s-1,QS=-0.05 mm?s-1,H=47.2 T)[35-38].In the sample studied,a similar superposition of magnetically split sextuplet and a quadruple split doublet was also observed when the superparamagnetic Fe3+was still dominant.However,the RA of the larger iron-oxide crystallites gradually increased from the initial 29.0% to 39.5% after 1800 h of the continuousm-cresol degradation.Most superparamagnetic Fe3+species remained stable and showed a relatively strong interaction between Fe and AC.However,Fe3+also tended to condense into larger particles due to continuous reactions with oxidants and organic molecules.This study also confirmed that the catalytic activity did not always decrease when the degree of condensation of superparamagnetic ions into large particles was low.Fe2O3crystallites with particle size >10 nm had stronger interactions with the AC support compared to superparamagnetic particles.The results suggested that the formation of paramagnetic Fe2O3particles was supported by increasing amounts of Fe within a certain range.

Table 1 M?ssbauer parameters of 3%Fe/SAC before and after CWPO reaction

3.5.Reaction model and decomposition mechanism of m-cresol

To investigate the decomposition mechanism,in situinfrared tests ofm-cresol with and without catalysts were performed.The results are shown in Fig.9.The quantity of C=O produced by the oxidation of H2O2alone increased (Fig.9(a)),indicating that the intermediate products,such as carboxyl groups,were produced gradually.After the addition of catalysts ((Fig.9(b)),the quantity of C=O showed a slight decreasing trend.The quanity of C-OH(carboxylic hydroxyl) initially increased and then decreased,indicating that the hydroxyl radical mechanism was present.

Fig.6.Structural optimization results.

Fig.7.TPD–MS of 3%Fe/AC catalyst during continuous reaction,(a) spectra of CO evolution and (b) spectra of CO2 evolution.

Fig.8.M?ssbauer spectra of 3%Fe/AC catalyst during continuous reaction after (a) 0 h,(b) 450 h,(c) 900 h,(d) 1350 h,and (e) 1800 h.

Fig.9. In situ infrared testing of batch reactions (a) without catalysts,(b) with 3%Fe/AC as catalyst,and (c) 3D spectras with 3%Fe/AC,reaction conditions:Cm-cresol=100-mg?L-1,CH2O2=541 mg?L-1, T=343 K,initial pH=7.0.

As shown in Figs.S2 and 8,over 90% ofm-cresol can be completely decomposed through continuous CWPO reactions.However,instead of complete mineralization,most of thism-cresol was converted to poorly soluble organic intermediates[39].Hence,the reaction pathway was studied in detail using GC–MS.The results demonstrated that the ·OH radicals attacked the aromatic ring ofm-cresol by catalyzing 3%Fe/SAC and generating considerable quantities of levulinic acid,malonic acid and oxalic acid.As shown in Fig.S5I,a plausible reaction pathway duringm-cresol destruction was proposed.Note that no cyclic intermediates were present in the GC–MS analytical results in this multistep decomposition study due to the rapid oxidation and accumulation of the generated short-chain organic acids.These short-chain organic acids have lower HOMO energy and are not easily destroyed by·OH radicals.Fig.S5II shows the calculated HOMO energies of the reactants and a series of possible products during the decomposition process.

A plausible reaction pathway duringm-cresol degradation has been proposed,as shown in Fig.S5I.First,·OH attacksm-cresol via direct hydroxylation to generate 2-methylhydroquinone,which is subsequently oxidized to 2-methyl-p-benzoquinone due to its relatively higher HOMO energy (-5.86 eV).Further degradation leads to aromatic ring cleavage,and the cyclic intermediates are transformed into five-carbon and four-carbon compounds(including levulinic acid).Then,malonic acid and some other short-chain acids may be formed and finally converted to oxalic acid,which the GC–MS analysis in this study showed has the highest concentration.The dashed arrow indicates the possible degradation pathway,which could not be detected experimentally due to low concentrations.It is important to note that cyclic intermediates were not observed during the GC–MS analysis in this multistep degradation study,due to its quick oxidation and lack of accumulation that generated short-chain organic acids.

4.Conclusions

With AC as the catalyst support,a highly efficient and stable 3%Fe/AC catalyst was synthesized.This catalyst may be applied to CWPO treatment of wastewater containingm-cresol.After indepth research and systematic analysis,we draw the following conclusions.

(1) SOGs were the necessary factors for enhancing the CWPO reaction with ACs.Among the SOGs,carboxylic groups(37.3%),anhydrides (5.3%),and ether/hydroxyl (19.3%)showed positive effects,while carbonyl groups showed negative effects (-71.4%).

(2) Wet impregnation of 3%Fe/AC showed excellent CWPO capacity.Both SEM and TEM analyses,supported by FIB,indicated that the active components were mainly presentin the forepart of AC macropores.EPR demonstrated that the 3%Fe/SAC+H2O2catalytic system had the highest concentration of ·OH radicals.

(3) During continuous CWPO experiments with 3%Fe/AC as a catalyst,m-cresol conversion and TOC removal reached 90% and 30%,respectively,from the start of the experiment to 1800 h of reaction time(and even after that),with no significant decrease in activity.

(4) It is confirmed by DFT calculation that the carboxyl group,hydroxyl group and active Fe component in the Fe/AC catalyst were the main catalyst sites of H2O2in the CWPO reactions.

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.

Nomenclature

AC activated carbon

AE adsorption energy

ATR attenuated total reflection

BET Brunauer-Emmett-Teller

CIF crystallographic information file

CWAO catalytic wet air oxidation

CWPO catalytic wet peroxide oxidation

DFT density functional theory

EPR electron paramagnetic resonance

FIB focused ion beam

HB hydrogen bonds

IS isomer shift

OH carboxylic hydroxyl

QS quadrupole splitting

SEM scanning electron microscopy

TEM transmission electron microscopy

TOC total organic carbon

XRD X-ray diffraction

XRF X-ray fluorescence

Acknowledgements

This research was funded by the National Natural Science Foundation of China(52100072),the Beijing Natural Science Foundation(8214056),the special fund of Beijing Key Laboratory of Clean Fuels and Efficient Catalytic Emission Reduction Technology,the Strategic Priority Research Program of the Chinese Academy of Sciences(XDA21021101),the National Key Research and Development Program of China (2019YFA0705803),Scientific Research Common Program of Beijing Municipal Commission of Education(KM202010017006).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2021.04.040.