Optimization of α-Fe2O3@Fe3O4 incorporated N-TiO2 as super effective photocatalysts under visible light irradiation

Accepted Manuscript Title: Optimization of ␣-Fe2 O3 @Fe3 O4 Incorporated N-TiO2 as Super Effective Photocatalysts Under Visible Light Irradiation Author: Mohamed Mokhtar Mohamed T.Y. Mansour El-Ashkar W.A. Bayoumy M.E. Goher M.H. Abdo PII: DOI: Reference:

S0169-4332(17)30892-9 http://dx.doi.org/doi:10.1016/j.apsusc.2017.03.200 APSUSC 35576

To appear in:

APSUSC

Received date: Revised date: Accepted date:

10-2-2017 10-3-2017 22-3-2017

Please cite this article as: M.M. Mohamed, T.Y.M. El-Ashkar, W.A. Bayoumy, M.E. Goher, M.H. Abdo, Optimization of rmalpha-Fe2 O3 @Fe3 O4 Incorporated N-TiO2 as Super Effective Photocatalysts Under Visible Light Irradiation, Applied Surface Science (2017), http://dx.doi.org/10.1016/j.apsusc.2017.03.200 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Optimization of α-Fe2O3@Fe3O4 Incorporated N-TiO2 as Super Effective Photocatalysts Under Visible Light Irradiation Mohamed Mokhtar Mohameda*, T. Y. Mansour El-Ashkarb, W.A. Bayoumya , M. E. Goherb, M. H. Abdob a

Benha University, Faculty of Science, Chemistry Dept., Benha, Egypt. National Institute of Oceanography & Fisheries, Environmental Chemistry, Cairo, Egypt.

ip t

b

cr

*Corresponding author; e-mail:[email protected]

us

Graphical abstract

M

an

Highlights The α-Fe2O3/Fe3O4 doped n-TiO 2 was synthesized via deposition-self assembly technique. The photocatalyst 1% α-Fe2O3 /Fe3O4/n-TiO2 show a remarkable performance while MB degradation. The strong interaction between α-Fe2O3/Fe3O4 and n-TiO 2 plays an important role. It exhibits a unique textural, optical and charge transfer properties.

Ac ce pt e

d

Abstract Well dispersed α-Fe2O3@Fe3O4 nanoparticles (7 nm) supported on mesoporous nitrogen doped titanium dioxide (N-TiO 2) are synthesized by deposition self-assembly route and their performances as photocatalysts toward methylene blue (MB) degradation are evaluated. The results illustrate that the spherical yolk-shell structure of α-Fe2O3@Fe3O4@N-TiO2 at the loading of 1%; of excellent SBET (187 m2/g) and pore volume (0.50 cm3/g), achieved high photocatalytic performance for the MB degradation (20 ppm, λ>420 nm, lamp power= 160 W) under visible light illumination (k= 0.059 min-1). The influence of the interface formation between α-Fe2O3@Fe3O4 and n-TiO2 affects severely the charges separation efficiency and enhances the electron transfer to keep on the existence of Fe3+ /Fe2+ moieties; those take significant role in the reaction mechanism. The existence of the latter junction is affirmed via XRD, TEMSAED, Raman and FTIR techniques whereas, the photogenerated charges, their separation together with their transport and recombination rates are depicted via photoluminescence, electrical conductivity, incident photon to current efficiency (IPCE), cyclic voltammetry (CV) and impedance (EIS) measurements. The catalyst loading, zero point charge, pH variation, total organic carbon (TOC%) and the effect of lamps power are thoroughly investigated. The 1%α-Fe2O3@Fe3O4@N-TiO 2 photocatalyst also indicated high activity as a Fenton-like reagent accomplishing the MB degradation (100% removal) in 35 min with a rate of 0.07 min-1 at H2O2 concentration of 0.4 mM. The obtained results demonstrate that the heterojunction nanoscaled materials possess superior visible-light driven photocatalytic activitywith appreciable recyclability and promising utilization as a supercapcitor (426 F g -1 at scan rate of 5 mV s-1) device. 1 Page 1 of 22

Keywords: Photocatalysis; α-Fe2 O3@Fe3O 4@ N-TiO2; Visible light; Electrochemical property; Charge transfer

1.

Introduction

Ac ce pt e

d

M

an

us

cr

ip t

Incorporation of nitrogen into TiO2 crystal lattice has stimulated the generation of extra allowed energy levels and thus induced reduction in the energy band gap values [1, 2]. This indeed promotes the visible light absorption and causes an enhancement during the photocatalytic reactions [3, 4]. However, such nitrogen implantation and its induced shift into the visible light was somehow marginal in most cases. This basically depends on the exact chemical state of the introduced nitrogen, its location in titania lattice (substitutional and/or interstitial) as well as its limited inside the TiO 2 lattice [5]. One of the key factors affecting the visible light responsive photocatalytic activity of nitrogen doped TiO2 is associated with the creation of oxygen vacancies connected to Ti3+ species, which act as recombination centers for the photoinduced electrons and holes. This indeed nullifies the visible-light activity originated from the N 2p levels nearby the valence band [6]. Accordingly, surface nitrogen-doping of TiO2 are far from attaining the desired improvement in visible light driven photocatalytic reactions. Therefore, different approaches have been explored to improve the visible light TiO2 photocatalysis including the addition of multi-metal oxides [7-9]; which can help to tune both optical and electronic properties, other than adding Fenton based oxidation compounds via utilizing Fe2+ and H2O2 components [10]. It was also found that carbon is utilized within the mentioned systems to improve the separation of photogenerated charge carriers and to facilitate the adsorption of the pollutant molecules at the carbon/catalyst interface, as in the work of Hou et al. [11] reported for the Bi12TiO20 photocatalyst. In same essence, the microspheres Fe3O4@C@F-TiO2 prepared by solvothermal method have shown superior visible-light driven photocatalytic activity in Rhodamine-B degradation [12]. The photo-Fenton catalyst Fe3O4@void@TiO2 of nanospheres shape demonstrated high degradation activity of tetracycline exceeded those of Fe3O4@TiO 2, hollow TiO2 and Fe3O4 [13]. The magnetically separable hollow spherical Fe3O4/TiO2 photocatalysts prepared via a polystyrene-acrylic acid template exhibited good photocatalytic activity for RhB degradation under UV light irradiation [14]. However, these systems suffer problems while syntheses and performances including complicated multi-step procedures, time-consumption, necessity of pH control while performing the reactions, large H2O2 outflows and troubles concerning the sludge generation [14-16]. Some are also used UV irradiation (4-5% of sunlight) and thus, limits the possibility of performing the organics degradation under the wide visible light illumination range [14, 17]. Accordingly, great efforts are still necessitated to the catalyst design and modification. Most importantly, increasing the oxidation potential of •OH even at high pH values as well as alleviating and/or removing the usage of H2 O2 is mandatory. Accordingly, the creation of •OH radicals within the used photocatalytic system becomes an 3+ 2+ essential task. In addition, the reduction of Fe to Fe by the photo-induced electron transfer

[18–19] with enhanced charge separation and transportation must be accomplished in a facile way. Accordingly, based on the advantages of N-doped TiO 2 in improving its capacity, electronic and chemical properties it’s used as a support for the designated hetero-architectures α-Fe2O3/Fe 3O4 for well addressing photocatalytic remediation of an organic pollutant. It has been acknowledged that Fe 3O4 @TiO2 systems suffer dramatically from reduced photocatalytic activities owing to the electron-hole recombination caused by the electronic heterojunction involved among the core-shell Fe3O4@TiO 2 construction [20, 21] and rather they only act under UV light irradiation [22–23]. An alternative strategy to develop a solution of this issue is coating Fe3O4 with α-Fe2O3 of low energy band gap (Eg=2.2 eV) and excellent chemical stability to 2 Page 2 of 22

Experimental Materials

Ac ce pt e

2. 2.1.

d

M

an

us

cr

ip t

control delaying the electron-hole recombination while irradiating α-Fe2O3 and N-TiO 2 via the developed interface of Fe3O4. Thus, the increase in the charge transfer at the αFe2O3 and N-TiO2 surfaces via the superparamagnetic Fe3O4 is expected to cause an increase in activity aided by the acknowledged unique electrical properties of the latter moiety [24]. Apart from the mentioned deficiencies, α-Fe2O3 is considered as a great photocatalyst with an appreciable visible light absorptivity and with a flat band potential of 0.32 V vs. normal NHE (pH=0) [25-26]. Thus, it can form a junction with Fe3O4 due to the acknowledged difference in the work function between them. That heterojunction structure will improve the separation of photogenerated e--h+ pairs, suppressing the charge recombination and indeed expect to increase the charges lifetime. In addition, depositing the α-Fe2O3/Fe 3O4 moieties on the mesoporous N-TiO 2 of magnificent surface texturing properties will indeed diminish any possibility of their agglomeration and passivation problems. Accordingly, the as-synthesized αFe2O3@Fe3O4@N-TiO2 photocatalysts via deposition-self assembly technique were thoroughly characterized using XRD, TEM-SAED, Raman, UV-diffuse reflectance, photoluminescence, N2 adsorption and then investigated in the oxidative degradation of the methylene blue (MB) dye; under visible light illumination. The kinetics of MB degradation together with total organic carbons, effect of pH, zero point charge and catalyst loadings were thoroughly evaluated. Estimation of the reactive species responsible for the MB degradation as well as the incident photon to current efficiency (IPCE) was also studied. Investigating the different oxidation species on the catalyst surfaces, their role and charges transfer obstruction were also determined via cyclic voltammetric and impedance measurements together with evaluating the photocatalysts recyclability. The role of charge densities of the various samples and their function in the activity at room temperature were determined via electrical conductivity measurements.

Titanium (IV) isopropoxide (TIP), conc. HCl, aqueous ammonia, urea and methanol were obtained from Sigma-Aldrich. Brij-35 C12H25(OCH2CH2)nOH, n~23 of Molecular Weight 1199.54 and methylene blue dye were obtained from Loba Chemie. 2.2: Catalyst fabrication 2.2.1. Synthesis of α-Fe2O3/Fe3O4 NPs α-Fe2O3/Fe3O4 nanoparticles were synthesized by co-precipitation of Fe 2+ and Fe3+ aqueous salt solutions using NH3·H2O as a precipitating agent. Accordingly, the solutions of 3.9 g Mohr's salt (NH4)2Fe(SO4)2 · 6H2O in 100 mL H2O (0.1 M) and 4.8 g NH4Fe(SO4)2. 12H2O in 100 mL H2O (0.1 M) were mixed with a molar ratio of 1:2. Ammonia aqueous solution was then dropped into the mixture slowly until reaching pH value of 9. The complete precipitation of α-Fe2O3/Fe3O4 was then obtained, washed, dried at 100oC and stored in desiccators. 2.2.2. Synthesis of nitrogen doped-mesoporous TiO 2 (N-TiO2)

3 Page 3 of 22

cr

ip t

The N-TiO2 was synthesized according to Evaporation-Induced Self-Assembly route. Typically, 1.5 ml Titanium (IV) isopropoxide (TIP) dispersed in 3 ml conc. HCl; labeled as solution A, was stirred vigorously till the pale yellow color is formed. 190 µL of the non-ionic surfactant Brij 35 (30 % W/V) plus 2 ml distilled water with different weights of urea (0.01 , 0.02 and 0.03 gm ) were added to 35 ml methanol in a 250 ml beaker, and thus labeled as solution B. Solution B was added drop by drop to solution A while moderate stirring. An aqueous ammonia solution (30%) was then dropped wisely till pH 7.5. The mixture was stirred for 1 h then poured to a petri dish and kept static at 40oC for 24 h. The temperature was raised to 50oC then to 60oC for 24 h after that it kept at 100oC for 4 h to ensure the completeness of the condensation process. The solid product was then grinded in the mortar with a pestle then calcined in air at 350oC for 2 h with a heating rate of 2oC /min.

us

2.2.3. Synthesis of α-Fe2O3/Fe3O4 /N- TiO2.

Ac ce pt e

d

2.3. Physical measurements : 2.3.1. X-ray diffraction

M

an

After mixing B over A as mentioned above, different weights percentages of αFe2O3/Fe3O4 to form 1, 3 and 5wt% ratios (0.0032, 0.0096, 0.016 g) relative to N-TiO2 were added via sonication in 20 ml methanol and 5 ml distilled water followed by stirring for 1 h. The weight of urea employed in these preparations was 0.02 g, equivalent to 8wt% of the produced TiO2. An aqueous ammonia (30 %) solution was added drop by drop till pH 7.5. The treatment of the samples was exactly as mentioned above till calcination at 350oC. These were denoted as 1, 3 and 5wt% αFe2O3/Fe3O4/N-TiO2.

X-ray diffraction (XRD) was measured at room temperature by using a Philips diffractometer using Model PW-3710. The patterns were progressed with Ni-filtered copper radiation (λ = 1.5418 Å) at 30 kV and 10 mA with a scanning speed of 2θ = 5◦ /min. The mean crystallites size were calculated using the Debye–Scherrer Eq. (1), in which K is a constant equal 0.9, λis the wave length of the CuKαradiation, βis the half peak width of the diffraction peak in radiant. The different phases were recognized with the help of ASTM powder data files.

2.3.2. FTIR spectroscopy The Fourier transform infrared (FT-IR) spectra were monitored via a single beam Thermo scientific Nicolet iS10 instrument. The samples were grounded with KBr (1:100) to form tablets, and thus confined into the sample holder in the spectrometer cavity to record the measurements in the 4000–400 cm−1 region. 2.3.3. N2 adsorption 4 Page 4 of 22

2.3.4. Ultraviolet–visible diffuse reflectance spectroscopy

ip t

The surface texturing properties namely BET surface area, total pore volume (V p) and mean pore radius (r) were determined from N2 adsorption isotherms measured at 77 K using a conventional volumetric apparatus. The samples were out-gassed at 473 K for 3 h under a reduced pressure of 10−5 Torr before starting the measurement. The total pore volume was taken from the desorption branch of the isotherm at p/p0 = 0.95, assuming complete pore saturation.

an

2.3.5. Transmission electron microscope (TEM)

us

cr

Diffuse reflectance ultraviolet–visible spectroscopy (UV–vis DRS) of the samples was carried out at room temperature using UV–vis JASCO spectrophotometer, V-570, in the range of 200–1000 nm with using BaSO4 as the reflectance standard. The band gap energy (Eg) of the samples is estimated via using the equation: αh=A(h-Eg)n/2 [22] and determined by finding the intercept of the straight line of the plot of (αh)1/2 or (αh)2 against hfor both indirect and direct transitions, respectively.

d

M

TEM micrographs were measured using a JEM-2100 model which achieves high resolution of 0.19 nm at a power of 200 kV. The powder samples were put on carbon foil with a microgrid. TEM images were observed with minimum electron irradiation to prevent damaging of the samples structure. Selected area electron diffraction (SAED) images were also recorded at an accelerating voltage of 200 kV.

Ac ce pt e

2.3.6. Photoluminescence spectroscopy

The photoluminescence (PL) excitation and emission spectra were measured on a thermo scientific fluorescence spectrophotometer model Lumina. The measurements were conducted at room temperature using a He/Cd laser (310 nm) as an excitation source. 2.3.7. Determination of point of zero charge Batch equilibrium technique [27] was employed to determine the pH at the point of zero charge (PZC). Portions of catalyst powder were introduced into an identified volume (20 mL) of 0.1 mol dm−3 KNO3 solution; as an inert electrolyte devoted for adjusting the ionic strength throughout the experiments. Initial pH values (pHinitial) of KNO3 solutions were adjusted from∼4.5 to∼11.5 via addition of 0.1 mol dm−3 HNO3 and/or 0.1 mol dm−3 KOH solutions. Suspensions of different solid to solution ratios (1:100) were allowed to equilibrate for 24 h in a shaker kept at room temperature. The suspensions were then filtered and the pH values (pHfinal) were again determined. 5 Page 5 of 22

2.3.7. Raman spectroscopy Raman spectra were traced using a LABRAM-HR Raman system with a 633 nm He– Ne laser (Horiba, Jobin-Yvon, France) using the 488 nm line of an ArC laser as the excitation beam. The incident power was 50 mW and the step length was 2 cmK1.

ip t

2.4. Photocatalytic reaction

Ac ce pt e

d

M

an

us

cr

The photocatalytic activities of synthesized samples were evaluated by the photodegradation of the methylene blue (MB) dye. Typically, a few mg of a catalyst was suspended in 100 mL aqueous solution of 20 mg L−1 methylene blue (MB) in a cylindrical quartz vessel surrounded with a water circulation. The solution was stirred under dark conditions for 60 min to ensure the accomplishment of an adsorption– desorption equilibrium. 160 W high pressure mercury lamp (Philips) emitting visible light irradiation (λ>420 nm) was placed above the MB solution with a fixed distance of 20 cm. The light was illuminated for a specific time through intervals. Roughly 3 ml of suspension was withdrawn and centrifuged at 4000 rpm for 5 minute to remove the catalyst. Clear samples absorbance was tested using UV-vis spectrophotometer jenway-6800 at 664 nm. In addition, the total organic carbon (TOC) of the mixture was determined via using a high TOC Elementar Analyser system in order to investigate whether the dye is photo-bleached or completely degraded. TOC analysis is attained via taking samples at the beginning, in the middle and at the end of the experiment. The removal efficiency was calculated using the following equation: removal (%) = C0 −Ce /C0 × 100, where C0 and Ce were initial and equilibrium concentrations of MB (mg/L), respectively. For exploring the reactive species might produced in the photocatalytic reaction [28], we used different scavengers including isopropanol (aquencher of •OH), pbenzoquinone (a quencher of •O2−), Na2EDTA (a quencher of h+ ) and carbon tetrachloride (a quencher of e-) at a concentration of 1.0 mM. 2.5. Electrical properties

The electrical properties of the prepared composites were demonstrated via compressing the powder of the sample under a pressure of 5 tons cm-2 to build up pellets. The two equivalent surfaces of the pellets (7 mm diameter and 1mm thickness) were coated with silver paste to ensure good electrical contact. The electrical measurements were carried out at a constant voltage (1 volt), in a frequency range from 1.0 kHz to 300 kHz, at the temperature of 25oC, using a programmable automatic LCR bridge (HIOKI: 3532-50). The dc electrical resistivity was measured with an electrical circuit consists of an electrometer (model 6517, Keithley), voltameter (Keithley, 2182) and 5 kV dc power supply. The dc-conductivity σdc of the material was calculated by the following equation σdc = (l/As). (1/R dc) where Rdc is the sample resistance, l is the length of the sample and A s is the cross-sectional area. 2.6. Electrochemical Measurement. 6 Page 6 of 22

M

3. Results and discussion

an

us

cr

ip t

The electrochemical behaviour of as-synthesized samples was thoroughly examined; at room temperature, in a three-electrode cell system named PGSTAT204 equipped with Nova 1.11 software for data calculation. The working electrodes were made-up via mixing the synthesized catalysts (85 wt%) with 10 wt% acetylene black and 5 wt% PTFE binder. The FTO conductive glass sheets of dimensions equal 1 cm2 and resistance of 14 Ωcm–2 were well cleaned and dried before depositing the nanocomposites. The nanocomposite layer was heated to 200oC for 3 h then left to cool down to room temperature. The latter deposited working electrodes were transferred to the electrochemical cell containing an electrolyte (KCl of 2.0 M conc.) and a platinum electrode; as the counter electrode, alongside a saturated Hg/HgSO4 electrode, as the reference electrode. CVs were performed between the potential of (-2) to (+1) V and at scan rates 5.0, 10.0, 20.0 and 50.0 mV/s. Electrochemical impedance spectroscopy (EIS) measurements were conducted via using the same mentioned apparatus at conditions of current ranging from 10 μA to 100 mA, frequency margin of 0.1 Hz-100 kHz and at constant potential equal 10 mV. The validation of the impedance spectra was determined based on the Kramers-Kronig transformation. Specific capacitance was also measured by galvanostatic charge-discharge via using chronopotentiometry using Digi-Ivy 2116 B-USA, equipped with DY2100B software for data calculation. The prepared working electrodes were measured at step current of 0.05, 0.2, 0.5, 1 and 2 mA.

d

3.1. The structure and morphology of α-Fe2O3/Fe3O 4/N-TiO2 Catalysts

Ac ce pt e

Fig. 1 displayed the XRD patterns of Fe3O4 and Brij assisted N-TiO 2 synthesis together with 1, 3 and 5wt% Fe3O4/n-TiO2 annealed at 350oC for 2 h. The XRD pattern of Fe3O4 shows peaks at 2θ=36.8o, 47.3o, 54.2o and 61.4 o corresponding respectively to the crystal planes 311, 422, 511 and 440 (JCPDS No. 19-0629). The latter pattern confirms that it’s not a pure form since it exhibits an additional significant peak at 2θ =25.2o ascribed to the 012 plane of the α-Fe2O3 phase (matched with PDF No-01079-007). The latter phase is found to constitute 40% of the Fe3O4 phase based on correlating the intensities of the latter plane to that of the 311 of Fe3O4. It seems that Brij has stimulated the existence of only anatase in the nitrogen doped TiO 2 via elaborating the crystal planes 101, 004, 200, 211, 204, 220 and 215. Interestingly, the latter diffraction peaks are also observed in the nitrogen free TiO2 sample (see inset in Fig. 1) comprehending no phase changes followed the nitrogen incorporation. The pattern of 1wt% α-Fe2O3/ Fe3O4 /TiO2 has shown weak intensities for all N-TiO2 peaks compared to both pure N-TiO 2 as well as the mixed form of α-Fe2O3/Fe3O4, confirming a decrease in crystallite size of this sample and rather postulates the well dispersion of different Fe moieties on N-TiO2 surfaces. This indeed proposes strong interaction between them. The crystallite size of the 101 plane determined via Scherer equation indicates 7 nm for 1wt% α-Fe2 O3/Fe3 O4/n-TiO2 versus 12 nm for the pristine N-TiO2 . On the contrary, 3 and 5wt% α-Fe2O3/Fe 3O4/n-TiO2 patterns exhibit an increase in intensity implying crystallites size enhancement via giving average sizes comprised of 12 and 14 nm, respectively. Revealing an increase in intensity of the 101 plane in latter samples compared to pristine N-TiO 2 points to the appreciable presence of α-Fe2O3 due to coincidence of its plane 012 with that of the 101of N-TiO2 anatase. This may give a 7 Page 7 of 22

ip t

hint about the possibility of substituting Fe3+ (0.65Å) for Ti4+ (0.61 Å) in anatase structure. However, an increase in 2θvalues rather than a decrease was monitored in latter samples with sufficient decrease in their lattice constants unlike that exhibited in 1wt% α-Fe2O3/Fe 3O4/n-TiO2. This indeed emphasized that at low loading (1wt%αFe2O3/Fe3 O4), cationic lattice substitution is obtained where at high loading an interstitial addition is perceived probably due to size effects. Indeed, that possible substitution to occur in the 1wt%Fe/N-TiO 2 sample is expected to create oxygen ion vacancy positions [29].

Ac ce pt e

d

M

an

us

cr

Figure 2a shows that the morphology of the as-prepared N-TiO2 has indicated polyhedral structures composed mainly of hexagonal and rhombus shapes. Their average diameters are ~ 12 nm. The TEM image of the 1% α-Fe2O3/Fe3O4/N-TiO2 sample (Fig.2b) indicates different morphological structure other than that of N-TiO2 and thus exposes a spherical shape of an average diameter equal 10 nm. The N-TiO2 of spherical shape was found to expose Fe moieties; like an egg shell shape, of diameter equal 7 nm. Such encountered structural changes confirm the strong interaction between Fe moieties and N-TiO2 structure. The selected-area electron diffraction (SEAD) pattern (inset in Fig. 2b) exhibits polycrystalline rings correspond to the (101) and (400) planes of anatase N-TiO2 along with (311) and (012) planes of Fe3O4 and αFe2O3 phases, respectively. This indeed confirms the coexistence of Fe3O4 and α-Fe2O3 phases superimposing that of N-TiO2 in the sequence of α-Fe2O3@Fe3O4@N-TiO2. The TEM image of 3%α-Fe2O3/Fe 3O4/ N-TiO 2 shows nanocrystalline structure with a rhombus shape (Figure 2c) and of an average diameter of 20 nm, demonstrating the consistency of the N-TiO 2 shape even at such Fe loading. Retaining the shape of the NTiO2 after Fe addition evokes the weak interaction between the moieties forming this sample. Clear atomic planes with a lattice spacing equal 0.352 nm; corresponding to the (101) lattice planes of anatase TiO2, is noticed together with lattice spacing 0.295 nm corresponding to Fe3O4 (2 2 0) plane. The TEM image of 5% α-Fe2O3/Fe3O4/nTiO2 (Fig.2d) reveals that the obtained N-TiO2 of the spherical shape of an average diameter equal 15 nm exposes Fe NPs (Fig. 1a) of 10 nm. A hollow yolk-shell structure of the α-Fe2O3-Fe3O4 @ N-TiO2 (Fig. 2d inset) sample was also shown. 3.2. N2 adsorption

To further study the porous structure of the as-synthesized α-Fe2 O3/Fe 3O4@N-TiO2 catalysts as well as their pristine N-TiO2 and α-Fe2O3/Fe3O4, the adsorption-desorption isotherms and pore size distribution curves were measured and shown in Fig. 3. As shown, the isotherms belong to type IV with H3 hysteresis loops of mesoporous character. The BET surface area of mesoporous N-TiO2 together with pore volume and pore radius (Table 1) indicated values equal 157 m2 g-1, 0.43 cm3 g-1 and 54.7Å, respectively exceeded those comparable to α-Fe2O3/Fe3O4 (85 m2 g-1, 0.22 cm3 g-1, and 51.2 Å). This proposes widening of the pores related to the N-TiO2 sample, comparatively. The desorption isotherms show major capillary condensation steps at the relative pressure of 0.2 for N-TiO 2 and 0.3 for α-Fe2O3/Fe3O4. However, for the isotherms of Fe incorporated N-TiO2 , the desorption branches were depicted at relative pressure range from 0.4 to 0.6 suggesting pore widening following the Fe incorporation [30]. Interestingly, Table 1 also shows an enhancement in the S BET of 1%α-Fe2O3/Fe 3O4@ N-TiO2 into 187 m 2 g-1 and the pore volume into 0.50 cm 3 g-1; 8 Page 8 of 22

while keeping the pore radius constant compared to pristine N-TiO 2 anatase sample. This indeed predicts that Fe moieties were enforced inside the pores and well dispersed inside it. Increasing the Fe loadings into 3 and 5% decreases the SBET values to be lower than that of the pure N-TiO2 sample. The same behaviour was attained for the pore volume suggesting pore blocking that by its turn caused a decrease in SBET values. This latter blocking affects severely the pore radius in the 5% Fe sample to exhibit the highest decrease in all measured texturing parameters (SBET, r and Vp).

ip t

3.3. Optical properties

Ac ce pt e

d

M

an

us

cr

The UV-vis absorption spectra of N-TiO 2, Fe3O4 /α-Fe2O3 as well as different Fe loadings doped N-TiO2 samples are depicted in Fig. 4a. The Fe3O4/α-Fe2O3 sample exhibits the lowest light absorption both in ultraviolet and visible light regions. Whereas, N-TiO2 shows higher absorption capacity in the wavelength range from 200 to 350 nm followed by a significant decrease in the visible light range. This sample indicates an absorption edge around 410 nm exceeding most of the as-synthesized TiO2 nanomaterials (used to be at ~ 380 nm) [31]. Both α-Fe2O3/Fe3O4 and N-TiO2 samples show direct charge transitions brought about via UV absorption and assigned to the ligand-metal charge transfer of O2 2p→Fe 3+(Ti4+ ) 3d [32]. The 1wt% αFe2O3/Fe3 O4/N-TiO 2 sample absorbed more light in the 400–800 nm region than rest of the samples where it presented relatively lower light absorption in the UV region than 3 and 5wt% α-Fe2O3/Fe3O4/N-TiO2 samples. Shifting the absorption of the former sample towards visible light region compared to latter samples is explained in view of the strong interaction exhibited between Fe and Ti moieties, as devoted from XRD, TEM-SAED and surface texturing results. The 3 and 5wt% α-Fe2O3/Fe3 O4 doped materials exhibit similar spectral profiles to that of the 1wt% α-Fe2O3/Fe3O4 one, with inferior light absorption in the visible region from 400 to 550 nm. This demonstrates that at high loadings, Fe moieties could cover large parts of TiO2 surfaces prohibiting appropriate light absorption. Additionally, Fe moieties at such high loadings own large crystallite sizes that induce by their turn the poor light absorption property. The spectra of incident photon to current conversion efficiency (IPCE) were measured and presented in Fig. 4b. The 1wt% α-Fe2 O3/Fe 3O4/ N-TiO 2 sample shows the highest IPCE comprised of 22%. This latter value exceeds that of pristine N-TiO2 by two times. However, α-Fe2 O3/Fe3 O4 and 3%α-Fe2O3/Fe 3O4/N-TiO 2 electrodes presented increase in the 250-450 nm margin followed by marked decrease afterwards in the 500800 nm margin comprehending their lower dye loading capacity, unlike that indicated for the 1wt% α-Fe2O3/Fe3O4/N-TiO2 electrode. In general, the trend of the IPCE spectra (Fig. 4b) of α-Fe2O3/Fe3O4, 3 and 5wt% α-Fe2O3/Fe3O4/N-TiO2 is somehow consistent with that of the UV-vis absorption spectra however, it deviates in N-TiO 2 and 1wt% α-Fe2O3/Fe3O4/N-TiO2. These latter samples show an increase in the IPCE in the wavelength range of 500-800 nm unlike that of UV-vis spectra probably due to their considerable decreased densities [32]. The difference in the IPCE enhancement between UV and visible illumination might be attributed to changes in the reactivity of the holes generated from N-2p and O-2p levels in UV and visible illumination. Moreover, the species with an oxidation potential more negative than the N-2p level is possibly can be oxidized by holes in this inter-band gap status thus performing an increase in the measured IPCE% [32]. This explains the increase in IPCE of N-TiO2 and rather assumes small reorganization energies upon doping with 1wt% α9

Page 9 of 22

Fe2O3/Fe3 O4, facilitating the charge transfer reactions as confirmed via the electronic conductivity results (as will be seen latter). Indeed at high Fe loadings, the cross sectional scattering phenomenon will let most of the light will be dissipated but at low loading (1%), the absorption efficiencies dependent on small sizes will be predominate. Since the dielectric constant is the one responsible for the adsorption based on the electron band structure, it is therefore possible that if the dimensions become too small, the dielectric constant and the band gaps alter resulting in the observed behaviour [32].

an

us

cr

ip t

The Tauc plots of (αh)2 versus photon energy of N-TiO2 sample indicates an indirect transition band gap of 2.9 eV comparable to the previously reported results (Fig. 4c) [30-31]. Addition of 1% α-Fe2O3/Fe3O4 to N-TiO2 decreases the band gap value into 2.65 eV exceeding that estimated to the pristine α-Fe2O3/Fe3O4 (2.15 eV). This latter value is in accordance with the band gap of hematite and lesser than some of magnetites [33–35]. However, increasing the Fe loadings into 3 and 5% led to increasing the band gap into 3.2 and 3.7 eV, respectively. Decreasing the band gap of the sample at the loading of 1%α-Fe2O3/Fe3O4 compared to latter samples is very much correlated to the strong interaction exhibited between Fe and Ti components. Also, oxygen vacancies devoted from substitution of Fe3+ to Ti4+ are likely responsible for the strong shift into the red margin.

Ac ce pt e

d

M

For monitoring the migration, transfer as well as the possibility of electron–hole pairs in α-Fe2O3/Fe3O4 doped N-TiO2 materials, the photoluminescence (PL) emission spectra are traced and shown in Fig. 5. The peak positions of all the samples are basically the same, representing that varying the loading of α-Fe2O3 /Fe3O4 onto anatase N-TiO2 has never induced new photoluminescence modifications apart from those existed at 464 nm (2.67 eV) and 530 nm (2.33 eV) for N-TiO 2. After doping N-TiO2 with α-Fe2O3/Fe3O4, the intensities of the PL emission peaks were all decreased and the maximum diminish was for the 1% sample that showed a shift to the peak at 530 nm (2.33 eV) into 590 nm (2.1 eV). This also explains the increase of the IPCE of this sample in the visible light wavelength region. This red luminescence shift is likely due to oxygen vacancy capable of forming specific deep donor levels [35]; as has been affirmed previously using XRD and UV-vis results. These vacancies are more probable on the surface boundaries of the components. Lowering the peak intensities of 1wt% α-Fe2O3/Fe 3O4/n-TiO2 than rest of all doped materials elaborates that doping at latter loading inhibits efficiently the recombination of electrons and holes. This sample is amenable for accomplishing higher photocatalysis performance. This hindered radiative recombination rate of electrons and holes explains that Fe ions work as harvesting centers. Reversing the latter behaviour with high Fe loadings advocates the Fe loading optimality that beyond it the suppression ability is reduced. This declares the formation of α-Fe2 O3.Fe3O/N-TiO2 interface through which electrons and holes transfer takes place. Interestingly, the emission spectrum of α-Fe2O3/Fe3O4 ions shows the lowest recombination rate of electrons and holes explaining not only it’s appropriateness as a delayer for charges recombination; when doped with N-TiO2, but also indicative to lowering the density of charges produced upon light excitation. This was in harmony with the UV-vis results that indicated that this latter sample owned the lowest amount of light absorbed in the whole wavelength range. 3.4. Raman and FTIR spectroscopy 10 Page 10 of 22

d

M

an

us

cr

ip t

To distinguish the phases and to have an idea about the group vibrations, Raman spectra were traced for the samples and presented in Fig. 6. together with the spectrum of α-Fe2O3/Fe 3O4 as inset. In harmony with XRD results, the inset Fig. shows the presence of α-Fe2O3 via existence of bands at 216 and 276 cm-1 whereas Fe3O4 displays through the appearance of the strong bands at 570 and 699 cm-1 [36]. The Raman spectrum of N-TiO2 indicates major bands at 149, 396 and 633 cm-1 associated respectively to Eg, B1g and Eg modes of anatase TiO2. This confirms the purity of our synthesized anatase phase, in conformity with elaborated XRD results. No change was perceived in the 1wt% α-Fe2O3/Fe3O4/N-TiO2 spectrum except an increase in intensity for the 149 cm-1 band, together with exhibiting a small shift of the 633 cm-1 band to 635 cm-1. Increasing the Fe loadings confirmed the existence of a new band at 518 cm -1 correlated to the B1g mode of anatase, notifying the existence of a new symmetry mode as a function of Fe loadings. Besides, the 149 cm-1 band showed a little shift into lower wavenumbers (147 cm-1) unlike the 633 cm -1 band that displayed a gradual increase into 638 cm-1 due to increasing the force constants [33]. The former decrease in wavenumbers of the Eg mode with an increase in intensity could be an indicative to the increase in particles size; in conformity with XRD and TEM results. When the grain size increase, the vibrational properties of nanomaterials are mainly influenced via volume expansion leading to decrease in force constants due to increase in interatomic distances. The sudden increase in scattering intensity as for the Eg mode may be due to the strengthening of long-range translational crystal symmetry caused due to Fe incorporation in non-defect sites. It is evident that both the undoped and Fe-doped NTiO2 samples are in anatase phase confirming the absence of any impurity-related modes, in agreement with XRD results.

Ac ce pt e

The FT-IR spectra of all synthesized catalysts are shown in Fig. S1 to confirm the oxides formation and the possible interaction between them. The spectrum of αFe2O3/Fe3 O4 indicates peaks at 3421, 1638, 881, 574 and 496 cm -1 characterizing respectively the stretching vibration of O-H, bending vibration of O-H, C-O, Fe-O in Fe3O4 and Fe-O in α-Fe2O3 [37]. Mesoporous N-TiO 2 provides peaks at 3342, 1644 and 1407 cm-1 besides, a broad peak extended from 850 to 450 cm-1 corresponding to the Ti-O and Ti–O–Ti vibrations. The peak at 1407 cm-1 is attributed to the Ti-O-N network [38] vibrations. The doped samples provide typical functional groups to that exposed on the N-TiO2 spectrum. In addition, the significant wide band ranging from 450 cm−1 to 850 cm−1 can be assigned to the overlapping peak between the Ti–O–Ti modes and Fe–O vibrations [38]. This indicates the co-existence of α-Fe2O3/Fe3O4 and N-TiO2 phases in the α-Fe2O3/Fe 3O4/N-TiO 2 hetero-junction as established previously using XRD, PL, Raman and TEM-SAED images. Shifting the OH groups in αFe2O3/Fe3 O4 into lower wavenumbers in all Fe doped N-TiO2 indicates more surfaceadsorbed water and hydroxyl groups on the latter, advocating that they are expected to play significant roles in the photocatalytic oxidation reactions. In addition, such heterostructures can capture the photo-induced holes (h+) when irradiated with light to form hydroxyl radicals those by their turn work as centers for O2 adsorption that eventually can form hydroxyl radicals [38]. 3.5. Photocatalytic properties of the α-Fe 2O3 /Fe3O4/N-TiO2 catalysts

11 Page 11 of 22

Ac ce pt e

d

M

an

us

cr

ip t

For examining the photocatalytic capabilities of the as-synthesized nanomaterials, MB absorption was first left for 60 min over the catalysts (0.250 g L-1) to affirm equilibration followed by exposure to the visible light irradiation (λ>420 nm, power =160 W) for 300 min (Fig. 7). Of particular interest, the α-Fe2 O3/Fe3 O4 composite has indicated no activity unlike the pristine N-TiO 2 that indicated appreciable activity comprised with 50% degradation. This suggests evoking of Ti3+ ions together with the amount of N doping and oxygen vacancies, those make contributions to the visible light absorption via the Ti-O-N linkages, affirmed via the FTIR results [38]. Both 3 and 5% α-Fe2O3/Fe3O4/N-TiO2 catalysts have designated similar activity comprised of 30% degradation unlike the 1%α-Fe2O3/Fe 3O4/N-TiO 2 photocatalyst that accomplished the greatest degradation via achieving 100% in 300 min reaction time and at a loading of 0.250 g L-1. Accordingly, the latter photocatalyst encountered the highest degradation rate (K=0.01 min-1 ) that followed pseudo-first-order and exceeded those of the other samples (inset in Fig. 7). The degradation rate constant of 1%αFe2O3/Fe3 O4/N-TiO 2 was found to be 3.7 and 8.3 times higher than N-TiO2 and 3 (5) %α-Fe2O3/Fe 3O4/N-TiO 2 photocatalysts, respectively. This concludes the facile degradation efficiency of the former photocatalyst, comparatively. This encountered enhancement in the activity of 1%α-Fe2O3/Fe3O4/N-TiO2 compared to rest of the samples is indeed correlated to the increase in SBET and pore volume values, high absorption capability in the visible light range, highest IPCE% as well as attaining the lowest band gap value. Specifically, increasing the pore volume value of the 1%αFe2O3/Fe3 O4/N-TiO 2 sample was responsible for enhancing the dark adsorption that reaches 20%, undeniably with increasing the SBET value as well. Contrarily, the significant decrease shown for 3 and 5% α-Fe2O3/Fe 3O4/N-TiO 2 photocatalysts is primarily correlated to lowering the visible light absorption and the consequence thereof concerning the decrease in IPCE and the increase in the Eg values. Besides, the encountered increase in the particles size that affected the surface area values to be the lowest comparatively.

It is interesting testifying that the excellent photocatalytic degradation of 1% αFe2O3/Fe3 O4/N-TiO 2 was performed in the absence of H2O2 signifying the necessity of demonstrating the different reactive species could be displayed during the reaction progressing. The effects of the addition of benzoquinone; BQ, isopropanol; IPA, carbon tetrachloride; CCl4; and sodium EDTA; Na2EDTA, on the catalytic oxidation of the MB dye over the photocatalyst 1%α-Fe2O3/Fe3O4/N-TiO2 have been investigated under identical experimental conditions (Fig. 8). It has been shown that the reaction rate was relatively retained upon using BQ and Na2EDTA reflecting the negligible effects of •O2- and holes species. Experiment with IPA and CCl 4 scavengers has indicated a significant decrease proposing the influential effect of •OH and electrons. Accordingly, adding small concentration of H2O2 (0.5 ml of 0.2 mM) increases the degradation rate to 3 times that without H 2O2 (0.03 min -1 vs. 0.01 min-1) to accomplish the efficacy of our material to work as a photo-Fenton- catalyst, and rather enhanced into 0.07 min-1 upon adding 1.0 ml H2O2 at a concentration of 0.4 mM (Fig.S2). This indicates that no •OH scavenger and no water and oxygen is attained even at high H2O2 concentration reflecting the efficacy of our catalyst also as a Fenton photocatalyst. This result also indicates that the H2O2 decomposition proceeds simultaneously with the MB degradation to produce •OH during the Fenton progressing.More correlations concerning the H2O2 decomposition at different conditions is beyond the scope of this work that shows tremendous activity in the absence of any oxidizing agent. This suggests that the intimate contact between α-Fe2 O3/Fe3 O4 and N-TiO 2; at the 1% 12 Page 12 of 22

us

cr

ip t

loading, as devoted from HRTEM-SAED, Raman and FTIR, is beneficial for the production of •OH and e- as main reactive species retained on the catalyst surface to perform degradation as a result of visible light absorption (scheme 1). Thus, under visible light illumination, the photogenerated electrons from N-TiO2 can be captured by Fe3+ to keep on the existence of Fe2+ based on the potential difference between them. Indeed, such electron trapping phenomenon delays its recombination with holes and thus enlarge the lifetime of holes. Fe2+; on the other hand, can react with •OH to form Fe3+ and OH- to keep on the high conversion efficiency of the Fe 3+/Fe2+ cycle. The holes in N-TiO2 migrate to the surface and can react straight forward with the dye to degrade it. Also, the photogenerated electrons in N-TiO2 or Fe3+ /Fe2+ photocatalysts surface are likely react with adsorbed O2 on the surface, resulting in the generation of O2• that by its turn can transfer into •OH, to increase the overall degradation efficiency. Holes are also capable of generating •OH based on the interaction of Fe2+ with H2O/OH- and thus increases the possibility of •OHs production; which are the most influential reactive species as scavenger study confirms.

an

3.6.1. Effect of catalyst loading, TOC analysis and stability

Ac ce pt e

d

M

Catalyst loading is an important factor that can strongly influence the dye degradation in the photocatalysis processes. Experiments carried out with different amounts of the photocatalyst 1wt%α-Fe2O3/Fe3O4/N-TiO2 showed that the photodegradation efficiency increases with an increase in the amount up to 1.0 g L-1 at which a complete degradation is attained in only 85 min (Fig. 9). The pseudo-first order rate constants are found to be 0.0026, 0.0029, 0.033, 0.051, 0.01 and 0.0592 min-1 for 1wt% αFe2O3/Fe3 O4/N-TiO 2 at the catalyst loading of 0.1, 0.15, 0.2, 0.225, 0.250 and 1.0 g L– 1 , respectively. This observation is explained in terms of accessibility of the active sites on the catalyst surface as well as the facile penetration of the visible light into the mixture suspension. Indeed, the total active surface area increases with increasing the photocatalyst quantity. Binding of water of this catalyst specifically; as devoted from FTIR results, initiates the facile formation of OH groups that takes a part in the reaction. At certain reaction intervals, the UV–vis absorption spectra of the MB dye were undertaken while tracing its photocatalytic degradation using the nanocomposite 1wt% α-Fe2O3/Fe 3O4/N-TiO 2. As shown in Fig. S3, two absorption peak maxima at 615 and 664 nm are observed characterizing the MB typical peaks [39]. These absorption peaks diminish with time and the solution turns colourless gradually within 85 min when using a catalyst quantity of 1.0 g L-1 and it takes 300 min upon using 0.250 g L-1 . The TOC% of the same sample taken after the 85 min reaction time; shown as inset in Fig. S3, indicates 82% elemental carbon representing the photodegradation ratio of the dye organic carbons. Whereas, after 300 min the TOC indicates a value comprised of 77%. Accordingly, the difference between C/Co and TOC% values; typical of 18% and 23% for both samples, are mostly correlated to the presence of non-degradable intermediates which are not amenable to degradation at the mentioned conditions. Extending the time of 85 min to 100 min and that at 300 min to 350 min has accomplished the 100% TOC degradation, verifying the complete transformation of organic carbons into elemental ones. Interestingly, increasing the TOC% at the 85 min than that at 300 min, reflects the efficiency of our catalyst at the 1.0 g L-1 loading to 13 Page 13 of 22

work at such little time to achieve higher TOC% value. This same photocatalyst at the 0.250 g L-1 loading was tested for recyclability (Fig.S4). After six recycles, the 1wt% α-Fe2O3/Fe 3O4/N-TiO 2 photocatalyst retains 80% after 1800 min without performing any treatment among the runs. This indeed indicates the excellent stability of the catalyst as well as its potential towards applicability. 3.6.2. Surface properties and Point of zero charge effect

Ac ce pt e

d

M

an

us

cr

ip t

In order to configure the effect of pH on the MB adsorption and the degradation consequences thereof upon illumination onto different weight percentages of αFe2O3/Fe3 O4 doped N-TiO2 , the point of zero charge (PZC) was measured and presented in Fig. 10. Accordingly, it’s interesting notifying that the pristine N-TiO2 indicates PZC of 9.7 lower than that of α-Fe2O3/Fe3O4 (11.2) exploring that they own great preference to be hydroxylated. The doped samples indicate lower PZC than former samples and assigned in the order: 1% α-Fe2O3 /Fe3O4 /N-TiO2 (8.6)> 3% αFe2O3/Fe3 O4/N-TiO 2 (7.5)>5% α-Fe2O3/Fe3O4/N-TiO2 (6.7). The encountered increase in the MB degradation rate monitored at pH 11; and exceeded that obtained at 8, on 1% α-Fe2O3/Fe 3O4/N-TiO 2 (Fig. 11) was basically due to varying the adsorbent surface properties as well as the degree of ionization of the MB dye as a function of pH. Accordingly, an increase in the pH into 11 induces high negative charge densities on the surfaces of 1% α-Fe2O3/Fe3O4/N-TiO2 and thus facilitates the exposure of the leuco MB (neutral structure) and its uptake attaining the maximum adsorption value. Accordingly, that latter sample attained the highest degradation rate with 170 min reaction time instead of 300 min upon using a catalyst quantity of 0.250 g L-1. Accomplishing 100% removal at pH = 8 for this sample although it acquires partial +ve charges; pH420 nm on the activity of the photocatalyst 1% α-Fe2O3/Fe3O4/N-TiO2 at similar reaction conditions as mentioned above. Obviously, the 160 W power lamp indicated the 14

Page 14 of 22

highest rate constant (0.01 min-1) followed in sequence by 320 W (0.033 min -1) and 75 W (0.0024 min -1) based on employing the catalyst quantity of 0.250 g L-1. This may be explained in view of the amount of the impinged charges (e—-h+) on the catalyst surface and the possibility of their recombination as in the case of the 320 W lamp because of their high densities. Contrarily, the probability of charge densities in case of the 75 W lamp is likely low enough to catalyze efficiently the degradation of MB.

ip t

3.7. Electrochemical behaviour of synthesized photocatalysts

Ac ce pt e

d

M

an

us

cr

Figure 13 shows the cyclic voltammograms (CVs) of the as-synthesized samples between -2.1 and 0.4 V with a scan rate of 5 mV s -1. The cyclic voltammetry of the αFe2O3/Fe3 O4 electrode exhibits redox peaks that are consistent with the previously reported data of α-Fe2O3 and Fe3O4 [41]. Prevailing pair of redox peaks is also identified in the N-TiO2 electrode. These peaks are attributed to the transformations between Fe(II) and Fe(III) as well as among Ti(III) and Ti (IV), respectively. The current density of the latter electrode [(Ti(III)/Ti (IV)] exceeded that of the former [(Fe(II)/Fe(III))] one and rather showed more shifts into negative potentials, reflecting the facile transfer of the electrolyte into the electrode surface of Ti(III)/Ti(IV). The oxidation peak observed in the 1% electrode and at the potential of -0.5 V possessed higher oxidation current (Fig.13 inset) than rest of the electrodes maximizing the quicker electron transfer from Fe (II) to Fe (III). Significant higher current density is observed on the voltammograms of 1% α-Fe2O3/Fe3O4/N-TiO2 electrode suggesting stronger redox activities compared to α-Fe2O3/Fe3O4 and N-TiO2. This redox behaviour of Fe2+ /Fe3+ as well as Ti 3+ /Ti4+ support the charges transfer role in the photocatalysis mechanism and their great potential to be used as electro-photocatalysts. Interestingly, the 1% α-Fe2O3/Fe3O4/N-TiO2 electrode showed more shifts into negative potential than α-Fe2O3/Fe3O4 and 5% α-Fe2O3/Fe3O4/N-TiO2 those tend to give typical voltammograms to N-TiO2 , with minor shifts. The specific capacitances (C sp) for the investigated electrodes were computed from the CV curves according to the equation (1) [42]. Vc

1 Csp  iVdV vw (V ) Va

(1)

Where ΔV (V) is the potential drop, the scan rate is v (mV s-1), and w (g) is the mass of the electrode. The specific capacitance measured at the scan rate of 5.0 mVs-1 gives values equal 136, 204, 426, 124 and 292 F.g-1 for α-Fe2O3/Fe 3O4, N-TiO2, 1, 3 and 5% α-Fe2O3/Fe 3O4/N-TiO 2, respectively. Increasing the specific capacity of 1%αFe2O3/Fe3 O4/N-TiO 2 into 426 F.g-1 at the scan rate of 5 mVs-1 advocates the enhancement of the mobility as well as the conductivity of the materials forming this electrode. Also, cyclic voltammetry results confirmed that the synergism between TiO2 and Fe moieties at the loading of 1% accelerated the Fe(III)/Fe(II) redox cycle via enhancing electron transfer, and thus facilitate the photocatalysis and Fenton-like performances of this specific catalyst that functioned well in a wide pH range. The 426 F.g-1 value exceeded a lot of work in literatures [42-43] for example that of Cheng et

15 Page 15 of 22

al. [44] reported for the 3D Fe2O3/CNT sponge electrode (∼300 F g−1 at a scan rate of 5 mV s−1) as well as that of xia et al. [45] whom optimized the Fe2O3/FGS composite to reach specific capacitance of 347 F g−1 at a scan rate of 10 mV s−1.

M

an

us

cr

ip t

To confirm the role of the charge carriers of synthesized nanomaterials and their motilities in enhancing the photocatayltic degradation, electrical conductivity was measured (Fig.14). The measured values of the electrical conductivity were in the order, 1% α-Fe2O3 /Fe3O4/N-TiO2; 4.0 x10-6 Ω-1cm-1> N-TiO 2; 3.8 x10-6 Ω-1cm-1> 3% α-Fe2O3/Fe 3O4/ N-TiO 2; 3.5 x10-6 Ω-1cm-1> 5% α-Fe2O3 /Fe3O4 /N-TiO2; 1.0 x10-6 Ω1 cm-1α-Fe2O3/Fe3 O4; 1.0 x10 -6 Ω-1cm-1. In comparison, the 1% loading shows an improved charge carrier density over the non doped N-TiO 2, those exceeded rest of the materials. This was in harmony with the enhanced current density as well as the IPCE accounted for increasing the charge carrier density. It seems that grain boundary scattering of free electrons at high Fe loadings affected the conductivity due to increasing the crystallites size as revealed from TEM results. This indeed explains that the conductivity of 5% α-Fe2O3/Fe3O4/N-TiO2 and α-Fe2O3/Fe3O4 attains the same value probably due to the creation of internal defects and/or surface defects at high Fe concentration and thus affects the electronic transition.

Ac ce pt e

d

The electrochemical impedance spectra (EIS) of all synthesized samples were recorded to acquire information concerning the transport of ions and electrons and the effectiveness of the photocatalyst to separate the photogenerated charges in the frequency range between 0.01 Hz and 5000 Hz (Fig. 15). Well-defined semicircles in the Nyquist plots in the frequency range from 20 to 450 Hz range were depicted for NTiO2 and α-Fe2O3/Fe3O4 electrodes. This Figure indicates lowering of the charge transfer resistance (Rct) between the working electrode and the electrolyte interface in the former than that on the latter. The α-Fe2O3/Fe3O4 electrode indicates a slope along the imaginary axis at low frequency region proposing ionic conductivity as well. Loading α-Fe2O3/Fe3O4 onto N-TiO2 shows that the hemispherical in case of 1 and 5% Fe electrodes is superimposed each other and own much lower resistance than 3% one that suffers lower contact and higher charge transfer resistance. This indeed explains the enhanced photodegdration of the 1% α-Fe2O3/Fe3 O4 material based the facile charge transfer property. However, it seems that it’s not the main reason since the 5% α-Fe2O3/Fe 3O4 indicates lower activity although it owns the same EIS behaviour. This indeed raises that other factors must be dominant as mentioned previously such as particles sizes, Surface area and the separation efficiency between electrons and holes. The electrochemical impedance (EIS) of all nanomaterials data is fitted by an equivalent circuit, which consists of a bulk solution resistance (Rs), a charge-transfer resistance (Rct), a pseudocapacitive element (C) from the redox process of FeII/FeIII and Warburg resistance (W) expected to result from the ions diffusion. This was in parallel with another element Cdl of the double layer capacitance; as shown in the inset of Figure 15. This inset circuit of 1% α-Fe2O3/Fe3O4/N-TiO2 shows the proximity of the circuit with the original data obtained via CVs and most importantly the coexistence of the pseudocapacitive and the double layer capacitance. The typical circuit of 1% α16 Page 16 of 22

cr

ip t

Fe2O3/Fe3 O4/N-TiO 2 indicated 2 of 1.3 that was lower than 3% α-Fe2O3/Fe 3O4 /NTiO2 (5.5) and 5% α-Fe2O3/Fe3O4/N-TiO2 (2.1). The latter electrodes exceed αFe2O3/Fe3 O4 and N-TiO2 ; those showed larger internal resistance. This clearly indicates that 1% α-Fe2O3/Fe 3O4/N-TiO 2 provides easier access (less resistance) for intercalation and deintercalation of charges compared to rest of the electrodes. The calculated time constant evaluated based on the cut-off frequency fmax, τ=1/2πfmax (fmax is the peak frequency)was the lowest for 1% α-Fe2O3/Fe3O4/N-TiO2, illustrating that the shortest time to charge the capacitor was for the latter electrode compared to rest of the materials. The latter observations concerning CVs and impedance measurements indicate the role of Fe incorporation at the 1%loading in increasing the charges density as well as their separation efficiencies. Other parameters such as light absorption, incident photon conversion efficiency, catalyst loading, pH and its influence on the point of zero charge are important factors influencing the reaction rate.

us

4. Conclusions

Ac ce pt e

d

M

an

In this study, α-Fe2O3/Fe3 O4 doped N-TiO2 synthesized using deposition-self assembly technique was used as visible light photocatalysts for the MB degradation. Amongst all the photocatalysts, the 1% α-Fe2O3/Fe3O4/N-TiO2 photocatalyst shows an exceptional rate for MB degradation comprised of 0.059 min-1 at a catalyst loading of 1.0 g L-1 with an acceptable usability. The strong interaction exhibited between Fe moieties and N-TiO2 has affected the reaction performance based on decreasing the band gap (2.65 eV), decreasing particles size (7 nm), increasing SBET (187 m2 g-1) and enhancing the incident photo to current efficiency (IPCE) value. It has been depicted that •OH and electrons are the main reactive species controlling the reaction mechanism. Cyclic voltammetery and EIS results have confirmed the ease of the charge transfer between the redox couple of Fe(III)/Fe(II), whereas the density of charges were affirmed via the electrical conductivity measurements. This result indicates that the α-Fe2O3/Fe3O4/ NTiO2 product especially at the 1% loading is a promising candidate for high performance photocatalyst and as supercapacitor.

Fig.S1 Fig.S2 Fig.S3 Fig.S4

References [1] A. Hattori, M. Yamamoto, H. Tada, and S. Ito, A Promoting Effect of NH 4F Addition on the Photocatalytic Activity of Sol-Gel TiO 2 Films. Chemistry Letters 8 (1998) 707. [2] T. Yamaki, T. Sumita, and S. Yamamoto, Formation of TiO2−xFx compounds in fluorine-implanted TiO 2. Journal of Materials Science Letters 21 (2002) 33.

17 Page 17 of 22

Ac ce pt e

d

M

an

us

cr

ip t

[3] H. Irie, S. Washizuka, N. Yoshino, K. Hashimoto, Visible-light induced hydrophilicity on nitrogen-substituted titanium dioxide films. Chemical Communications 9 (2003)1298. [4] R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki, and Y. Taga, Visible-light photocatalysis in nitrogen-doped titanium oxides. Science, 293(2001) 269. [5] B. Viswanathan and K. R. Krishanmurthy, Compositing semiconductor photocatalysts and their microstructure modulation. International Journal of Photoenergy Volume 2012 (2012), Article ID 269654, 10 pages doi:10.1155/2012/269654 [6] J. Wang, D. N. Tafen, J. P. Lewis, Z. Hong, A. Manivannan, M. Zhi, M. Li, N. Wu, Origin of photocatalytic activity of nitrogen-doped TiO2 nanobelts. J. Am. Chem. Soc., 131 (34) (2009) 12290. [7] L. L. Garza-Tovar, L.M. Torres-Martinez, D.B. Rodriguez, R. Gomez, G. del Angel, Photocatalytic degradation of methylene blue on Bi2MNbO7 (M = Al, Fe, In, Sm) sol–gel catalysts. J. Mol. Catal. A – Chem. 247 (2006) 283. [8] V. Subramanian, R.K. Roeder, E.E. Wolf, Synthesis and UV−Visible-Light Photoactivity of Noble-Metal−SrTiO3 Composites. Ind. Eng. Chem. Res. 45 (2006) 2187. [9] S. Murugesan, M. N. Huda, Y. Yan, M.M. Al-Jassim, V. Subramanian, BandEngineered Bismuth Titanate Pyrochlores for Visible Light Photocatalysis . J. Phys. Chem. C 114 (2010) 10598. [10] B. Kakavandi, A. Takdastan, N. Jaafarzadeh, M. Azizi, A. Mirzaei, A. Azari, Application of Fe3O4@C catalyzing heterogeneous UV-Fenton system for tetracycline removal with a focus on optimization by a response surface method. J. Photochem. Photobiol. A: Chem. 314 (2016) 178. [11] J. Hou, S. Jiao, H. Zhu, R.V. Kumar, Carbon-modified bismuth titanate nanorods with enhanced visible-light-driven photocatalytic property. Cryst. Eng. Comm., 13 (2011) 4735. [12] G. Liu, F. Hea, J. Zhang, L. Li, F. Li, L. Chen, Y. Huang, Yolk–shell structured Fe 3 O 4@ C@ F-TiO2 microspheres with surface fluorinated as recyclable visible-light driven. photocatalysts Applied Catalysis B: Environmental 150– 151 (2014) 515. [13] D. Du, W. Shi, L. Wang, J. Zhang, Yolk-shell structured Fe 3O4@ void@ TiO2 as a photo-Fenton-like catalyst for the extremely efficient elimination of tetracycline. Applied Catalysis B: Environmental 200 (2017) 484. [14] S. Xuan, W. Jiang, X. Gong, Y. Hu, Z. Chen, Magnetically separable Fe3O4/TiO2 hollow spheres: fabrication and photocatalytic activity. J. Phys. Chem. C 113(2009) 553. [15] E. Neyens, J. Baeyens, A review of classic Fenton's peroxidation as an advanced oxidation technique. J. Hazard. Mater. 98 (2003) 33. [16] V. Yathindranath, Z. Sun, M. Worden, L. J. Donald, J. A. Thliveris, D.W. Miller, One-pot synthesis of iron oxide nanoparticles with functional silane shells: a versatile general precursor for conjugations and biomedical applications. Langmuir 29(34) (2013) 10850. [17] S. Caudo, G. Centi, C. Genovese, S. Perathoner, Homogeneous versus heterogeneous catalytic reactions to eliminate organics from waste water using H2O2. Top. Catal. 40 (2006) 207. 18

Page 18 of 22

Ac ce pt e

d

M

an

us

cr

ip t

[18] Z. Ai, L. Lu, J. Li, L. Zhang, J. Qiu, M. Wu, Fe@ Fe2O3 core−shell nanowires as iron reagent. 1. Efficient degradation of rhodamine B by a novel sono-Fenton process .J. Phys. Chem. C 111 (2007) 4087. [19] X. Tan, L. Lu, L. Wang, J. Zhang, Facile synthesis of bimodal mesoporous Fe3O4@ SiO2 composite for efficient removal of methylene blue. Eur. J. Inorg. Chem. 2015 (18) (2015) 2928. [20] J. Navio, G. Colon, M. Trillas, J. Peral, X. Domenech, J. J. Testa, J. Padron, D. Rodriguez, M. I. Litter, Heterogeneous photocatalytic reactions of nitrite oxidation and Cr (VI) reduction on iron-doped titania prepared by the wet impregnation method. Appl. Catal. B 16 (1998) 187. [21] K. T. Ranjit, B. Viswanthan, Synthesis, characterization and photocatalytic properties of iron-doped TiO2 catalysts. Catalysts. Photobiol. B 108 (1997) 79. [22] Y, Yang, Y. H. Guo, C. W. Hu, Y. H. Wang, E. B. Wang, Preparation of surface modifications of mesoporous titania with monosubstituted Keggin units and their catalytic performance for organochlorine pesticide and dyes under UV irradiation. Appl. Catal. A 273 (2004) 201. [23] Y.X. Li, M. Zhang, M. Guo, X. D. Wang, Preparation and properties of a nano TiO2/Fe3O 4 composite superparamagnetic photocatalyst. Rare Metals 28(2009) 423. [24] L. Xu, J. Wang, Fenton-like degradation of 2, 4-dichlorophenol using Fe 3O4 magnetic nanoparticles. Appl. Catal. B: Environ. 123–124 (2012) 117. [25] Y. Xu, M.A.A. Schoonen, The absolute energy positions of conduction and valence bands of selected semiconducting minerals. Am. Mineral. 85(2000) 543. [26] X. Wei, T. Xie, L. Peng, W. Fu, J. Chen, Q. Gao, G. Hong, D. Wang, Effect of Heterojunction on the Behavior of Photogenerated Charges in Fe3O4@Fe2O3 Nanoparticle Photocatalysts. Phys. Chem. C 115 (2011) 8637. [27] I. Ibrahim, I. O. Ali, T. M. Salama, A. A. Bahgat, M. M. Mohamed, Synthesis of magnetically recyclable spinel ferrite (MFe2O4, M = Zn, Co, Mn) nanocrystals engineered by sol gel-hydrothermal technology: High catalytic performances for nitroarenes reduction. Applied Catalysis B: Environmental 181 (2016) 389. [28] M. Khairy and M. M. Mohamed, SnO2(β-Bi2O3)/Bi2Sn2 O7 nanohybrids doped with Pt and Pd nanoparticles: applications in visible light photocatalysis, electrical conductivity and dye-sensitized solar cells, Phys. Chem. Chem. Phys., 17 (2015) 21716. [29] M. V. Ganduglia-Pirovano, A. Hofmann, J. Sauer, Oxygen vacancies in transition metal and rare earth oxides: Current state of understanding and remaining challenges. Surface Science Reports 62 (2007) 219. [30] M. M. Mohamed, I. Ibrahim, T. M. Salama, Rational design of manganese ferritegraphene hybrid photocatalysts: Efficient water splitting and effective elimination of organic pollutants. Applied Catalysis A: General 524 (2016) 182. [31] L. G. Devi, R. Kavitha, Review on modified N–TiO 2 for green energy applications under UV/visible light: selected results and reaction mechanisms. RSC Adv., 4 (2014) 28265. [32] Q. Wu, Q. Zheng, R. van de Krol, Creating Oxygen Vacancies as a Novel Strategy To Form Tetrahedrally Coordinated Ti4+ in Fe/TiO2 Nanoparticles. J. Phys. Chem. C, 116 (12) (2012) 7219; P. Spinelli, M. Hebbink, C. van Lare, M. Verschuuren, R. de Waele, A. Polman, “Plasmonic Anti-Reflection Coating for Thin Film Solar Cells. In Optical Nanostructures for Photovoltaics” OSA Technical Digest (CD) (Optical 19

Page 19 of 22

Ac ce pt e

d

M

an

us

cr

ip t

Society of America, 2010), paper PWE3; R. Nakamura, T. Tanaka, Y. Nakato, Mechanism for Visible Light Responses in Anodic Photocurrents at N-Doped TiO2 Film Electrodes, J. Phys. Chem. B 108 (2004) 10617. [33] E.L. Miller, D. Paluselli, B. Marsen, R.E. Rocheleau, Low-temperature reactively sputtered iron oxide for thin film devices. Thin Solid Films 466 (2004) 307. [34] N. Ozer, F. Tepehan, Optical and electrochemical characteristics of sol–gel deposited iron oxide films. Sol. Energy Mater. Sol. Cells 56 (1998) 141. [35] S. Khan, J. Akikusa, Photoelectrochemical splitting of water at nanocrystalline nFe2O3 thin-film electrodes. J. Phys. Chem. B 103 (1999) 7184. [36] A. G. Nasibulin , S. Rackauskas, U. Jiang,Y. Tian, P. R. Mudimela,S. D. Shandakov, L. I. Nasibulin, S. J. E. I. Kauppinen, Simple and rapid synthesis of α-Fe2O3 nanowires under ambient conditions. Nano Research 2(2009) 373. [37] T. Wang, G. Yang, J. Liu, B. Yang, S. Dingc, Z. Yang, T. Xiao, Orthogonal synthesis, structural characteristics, and enhanced visible-light photocatalysis of mesoporous Fe2O3/TiO2 heterostructured microspheres. Applied Surface Science 311 (2014) 314. [38] D. Huang, S. Liao, S. Quan, L. Liu, Z. He, J. Wan, W. Zhou, J. Mater. Res. 22 (2007) 2389; G. Yang, Z. Jiang, H. Shi, T. Xiao, Z. Yan, Preparation of highly visiblelight active N-doped TiO2 photocatalyst . J. Mater. Chem., 20 (2010) 5301. [39] M. M. Ibrahim, S. A. Ahmed, K. S. Khairou, M. M. Mohamed, Carbon nanotube/titanium nanotube composites loaded platinum nanoparticles as high performance photocatalysts. Applied catalysis A: 475 (2014) 90. [40] A. A. Hoffmann, S. L. P. Dias, J. R. Rodrigues, F. A. Pavan, E. V. Benvenutti, E. C. Lima, Methylene blue immobilized on cellulose acetate with titanium dioxide: an application as sensor for ascorbic acid. J. Braz. Chem. Soc. 19 no.5 (2008) 943. [41] T. Zhu, J. S. Chen, X. W. Lou, J. Phys. Chem. C 115(2011) 9814-9820; X. Hu, M. Ma, M. Zeng, Y. Sun, L. Chen, Y. Xue, T. Zhang, X. Ai, R. G. Mendes, M. H. Rümmeli, L. Fu, Nanocrystalline iron oxide based electroactive materials in lithium ion batteries: the critical role of crystallite size, morphology, and electrode heterostructure on battery relevant electrochemistry. ACS Appl. Mater. Interfaces 6 (2014) 22527. [42] L. Wang, H. M. Ji, S. S. Wang, L. J. Kong, X. F. Jiang, G. Yang, Preparation of Fe3O4 with high specific surface area and improved capacitance as a supercapacitor. Nanoscale 25 (2013) 3793. [43] P. H. Yang, Y. Ding, Z. Y. Lin, Z. W. Chen, Y. Z. Li, P. F. Qiang, M. Ebrahimi, W. J. Mai, C. P. Wong, Z. L. Wang, Low-cost high-performance solid-state asymmetric supercapacitors based on MnO2 nanowires and Fe2 O3 nanotubes. Nano Lett. 14 (2014) 731. [44] X. P. Cheng, X. C. Gui, Z. Q. Lin, Y. J. Zheng, M. Liu, R. Z. Zhan, Y. Zhu, Z. K. Tang, Three-dimensional α-Fe2O3/carbon nanotube sponges as flexible supercapacitor electrodes. J. Mater. Chem. A 3 (2015) 20927.

20 Page 20 of 22

[45] H. Xia, C. Y. Hong, B. Li, B. Zhao, Z. X. Lin, M. B. Zheng, S. V. Savilov, S. M.

Aldoshin, Facile Synthesis of Hematite Quantum-Dot/Functionalized Graphene-Sheet Composites as Advanced Anode Materials for Asymmetric Supercapacitors. Adv. Funct. Mater. 25 (2015) 627.

Ac ce pt e

d

M

an

us

cr

ip t

Figure 1: XRD patterns of: α-Fe2O3/Fe3O4, N-TiO 2, 1%α-Fe2O3/Fe3O4/N-TiO2, 3%αFe2O3/Fe3 O4/N-TiO 2 and 5%α-Fe2O3/Fe3O4/N-TiO2 together with virgin TiO 2 pattern as inset. Figure 2: TEM micrographs of: a) N-TiO2 b) 1% α-Fe2O3/Fe3O4/n-TiO2 and SAED as inset C) 3%α-Fe2O3/Fe3O4/n-TiO 2 and HRTEM as inset d) 5%α-Fe2O3/Fe 3O4 / NTiO2 with HRTEM as inset. Figure 3: N2 adsorption-desorption isotherms and their corresponding pore size distribution curves of: N-TiO2, α-Fe2O3/Fe 3O4, 1%α-Fe2 O3/Fe3 O4/N-TiO 2, 3%αFe2O3/Fe3 O4/N-TiO 2 and 5%α-Fe2O3/Fe3O4/N-TiO2. Figure 4: a) UV-Vis diffuse reflectance spectra of N-TiO2, α-Fe2O3/Fe3O4, 1%αFe2O3/Fe3 O4/N-TiO 2, 3%α-Fe2O3/Fe 3O4/N-TiO 2 and 5%α-Fe2O3/Fe3O4/N-TiO2 samples b): Incident-photon-conversion efficiencies (IPCE%) of N-TiO 2, αFe2O3/Fe3 O4, 1%α-Fe2O3/Fe3O4/N-TiO2 , 3%α-Fe2O3/Fe 3O4/N-TiO 2 and 5%αFe2O3/Fe3 O4/N-TiO 2 and their corresponding Tauc plots of (αh)2 vs. hν(C). Figure 5: The PL spectra of N-TiO2, α-Fe2O3/Fe3O4, 1%α-Fe2O3/Fe3O4/N-TiO2, 3%αFe2O3/Fe3 O4/N-TiO 2 and 5%α-Fe2O3/Fe3O4/N-TiO2. Figure 6: The Raman spectra of N-TiO2, 1%α-Fe2O3/Fe3O4/N-TiO2, 3%αFe2O3/Fe3 O4/N-TiO 2 and 5%α-Fe2O3/Fe3O4/N-TiO2 samples and that of α-Fe2O3/Fe3O4 as an inset. Figure 7: Photocatalytic degradation of the MB dye by N-TiO 2, α-Fe2O3/Fe3O4, 1%αFe2O3/Fe3 O4/N-TiO 2, 3%α-Fe2O3/Fe 3O4/N-TiO 2 and 5%α-Fe2O3/Fe3O4/N-TiO2 photocatalysts under visible light irradiation: reaction conditions: (lamp power= 160 W, filter λ=420 nm, catalyst weight 0.250 g L-1, dye conc. 20 ppm). Inset is the corresponding kinetic fits for the MB degradation. Figure 8. Effect of the reactive scavengers on the MB degradation activity of the 1%αFe2O3/Fe3 O4/N-TiO 2 photocatalyst under visible light irradiation: the reaction conditions are (lamp power= 160 W, filter λ=420 nm, catalyst weight 0.250 g L-1, dye conc. 20 ppm). Figure 9. Effect of the catalyst weight on the activity of the MB degradation at similar reaction conditions as in Fig. 8. The inset is the corresponding kinetic curves. Figure 10: Iso-electric point determination of N-TiO2 , α-Fe2O3/Fe 3O4, 1%αFe2O3/Fe3 O4/N-TiO 2, 3%α-Fe2O3/Fe 3O4/N-TiO 2 and 5%α-Fe2O3/Fe3O4/N-TiO2 samples. Figure 11: Degradation efficiency of the photocatalyst 1%α-Fe2O3/Fe3 O4/N-TiO 2 as a function of pH. The reaction conditions are lamp power= 160 W, filter λ=420 nm, catalyst weight 0.250 g L-1, dye conc. 20 ppm.

21 Page 21 of 22

us

cr

ip t

Figure 12: Effect of the lamps power on the activity of 1%α-Fe2O3/Fe3O4/N-TiO2 towards degradation of the MB dye at reaction conditions: (filter λ=420 nm, catalyst weight 0.250 g L-1, dye conc. 20 ppm). The inset is the corresponding kinetic curves. Figure 13: CV curves of N-TiO2 , α-Fe2O3/Fe 3O4, 3%α-Fe2O3/Fe 3O4/N-TiO 2 and 5%αFe2O3/Fe3 O4/N-TiO 2 electrodes in 2.0 M KCl at scan rate of 5.0 mV s -1. The inset is the CV curve of 1%α-Fe2O3/Fe3O4/N-TiO2 at the scan rate of 5.0 mV s-1. Figure 14. Electrical Conductivity results of N-TiO 2, α-Fe2O3/Fe3O4, 1%αFe2O3/Fe3 O4/N-TiO 2, 3%α-Fe2O3/Fe 3O4/N-TiO 2 and 5%α-Fe2O3/Fe3O4/N-TiO2 samples measured at room temperature. Figure 15: Nyquist plots in 2.0 M KCl for N-TiO2, α-Fe2 O3/Fe 3O4, 1%αFe2O3/Fe3 O4/N-TiO 2, 3%α-Fe2O3/Fe 3O4/N-TiO 2 and 5%α-Fe2O3/Fe3O4/N-TiO2 electrodes at potential equal 0.01V in the frequency range 0.01 Hz-100 KHz. In-set is the equivalent circuit for the 1%α-Fe2O3/Fe3O4/N-TiO2 electrode. Scheme 1

Vp total (cm3/g)

157.85

54.74

0.43

85.90

51.24

0.22

187.36

53.58

0.50

156.67

53.91

0.42

151.95

47.20

0.35

M

r(Å)

Ac ce pt e

d

SBET (m2/g)

an

Table 1

22 Page 22 of 22