Enhanced photoelectrochemical and photocatalytic behaviors of MFe2O4 (M = Ni, Co, Zn and Sr) modified TiO2 nanorod arrays

Modified TiO2 nanomaterials are considered to be promising in energy conversion and ferrites modification may be one of the most efficient modifications. In this research, various ferrites, incorporated with various cations (MFe2O4, M = Ni, Co, Zn, and Sr), are utilized to modify the well aligned TiO2 nanorod arrays (NRAs), which is synthesized by hydrothermal method. It is found that all MFe2O4/TiO2 NRAs show obvious red shift into the visible light region compared with the TiO2 NRAs. In particular, NiFe2O4 modification is demonstrated to be the best way to enhance the photoelectrochemical and photocatalytic activity of TiO2 NRAs. Furthermore, the separation and transfer of charge carriers after MFe2O4 modification are clarified by electrochemical impedance spectroscopy measurements. Finally, the underlying mechanism accounting for the enhanced photocatalytic activity of MFe2O4/TiO2 NRAs is proposed. Through comparison among different transition metals modified TiO2 with the same synthesis process and under the same evaluating condition, this work may provide new insight in designing modified TiO2 nanomaterials as visible light active photocatalysts.

magnetic property of transition metal ferrites [13][14][15] . It is scarce on study of the visible responsiveness of MFe 2 O 4 to increase utilization of solar energy as well as to enhance the photoelectrochemical and photocatalytic performance of TiO 2 .
ZnFe 2 O 4 , with a relatively small band-gap (ca. 1.9 eV) 16 , is the most frequently studied to modify TiO 2 in enhancing the photoelectrochemical capacity. Yuan et.al observed that the ZnFe 2 O 4 /TiO 2 nanocomposite is more effective as a photocatalyst in the phenol degradation than pure TiO 2 . However the mechanism of the enhanced photoactivity of the ZnFe 2 O 4 /TiO 2 composite is still needed to be further understood 17 . Furthermore, the following researches proposed similar theory to explain the role of ZnFe 2 O 4 in enhancing photoactivity of TiO 2 18,19 , that is, the adoption of ZnFe 2 O 4 makes the ZnFe 2 O 4 /TiO 2 composite could use visible light, and the good match of band edges between ZnFe 2 O 4 and TiO 2 is in favor of charge carriers separating effectively. Reports about other MFe 2 O 4 modified TiO 2 , such as NiFe 2 O 4 /TiO 2 , Mn 0.5 Zn 0.5 Fe 2 O 4 /TiO 2 , MgFe 2 O 4 /TiO 2 and CuFe 2 O 4 /TiO 2 all show higher photoelectrochemical and photocatalytic performance [20][21][22][23] . It seems that MFe 2 O 4 is such a promising material to improve the photoelectrochemical and photocatalytic performance of TiO 2 . However, the comparison among the photoelectrochemical and photocatalytic performances of transition metal ferrites modified TiO 2 reported in the literatures is extremely difficult, because the experimental conditions were very different, such as catalysts synthesis process, light irradiation wavelength, reactor geometric configuration, catalyst loading and so on. Moreover, the origin and the crystalline structure of TiO 2 , which strongly affect its electronic and photoactivity, are also different. Therefore, the same condition should be taken into consideration when assessing the real effect of transition metal ferrites on the photoactivity of TiO 2 , such as using the same bare TiO 2 as the starting material, taking the same procedure to modify TiO 2 by transition metal ferrites, and finally evaluating their performances with unified standards. NiFe 2 O 4 , CoFe 2 O 4 , ZnFe 2 O 4 and SrFe 2 O 4 are four common transition metal ferrites which have been frequently studied with their magnetism, but except for ZnFe 2 O 4 , the other three are not common in modifying TiO 2 to enhance its photoactivity, therefore, we select the four as research objects, making a comparision between the common one (ZnFe 2 O 4 /TiO 2 ) and the uncommon ones (CoFe 2 O 4 , ZnFe 2 O 4 and SrFe 2 O 4 modified TiO 2 ).
In addition to incorporate other materials to modify TiO 2 , structural design is another important method to enhance the photoactivity of TiO 2 . One-dimensional (1D) nanostructure such as nanowire, nanotube, nanorod have attracted lots of attention due to the unique physical and chemical properties. 1D TiO 2 nanomaterials possess all the typical features of TiO 2 nanoparticles 24 . Electron diffusion length (up to ~100 μm) can be prolonged by using vertically aligned 1D nanostructures and excited electrons can easily pass along 1D nanostructure to the transparent conducting oxide electrode 25,26 , which facilitate charge transfer and promote charge separating efficiently 20,27 . However, the relatively low specific surface area on a smooth surface of 1D nanostructures may decrease the absorption ability and a single crystal phase of 1D nanostructures may pose certain constraints on the photoelectrochemical performance 24,28 . Fortunately, these disadvantages can be surmouned by introducing the second phase, i.e., doping metals/nonmetals or forming heterjunctions. Among1D nanostructures, 1D nanorod arrays with large area can be easily obtained by hydrothermal method, which is facile, economic and controllable 29 . Therefore, coupling the traits of one-dimensional TiO 2 nanorods (TiO 2 NRAs) and visible light responsive MFe 2 O 4 nanoparticles seems to be a promising way to enhance the solar energy conversion efficiency of TiO 2 .
To the best of our knowledge, there is few systematic research on the photoelectrochemical and photocatalytic capacity of various MFe 2 O 4 modified one-dimensional TiO 2 NRAs so far. In this study, large area uniform TiO 2 NRAs were synthesized hydrothermally and ferrites containing vaious cations (MFe 2 O 4 , M = Ni, Co, Zn, and Sr) were utilized to modify the as-prepared TiO 2 NRAs. The morphology, crystalline structures and optical properties as well as photoelectrochemical performances of TiO 2 NRAs and MFe 2 O 4 /TiO 2 NRAs were investigated. Moreover, the photocatalytic activities of the MFe 2 O 4 /TiO 2 NRAs were evaluated in the degradation of Cr(VI) aqueous solution under visible light irradiation. Finally, the underlying photcatalytic mechanism was discussed. In order to further examine the phase composition of the samples and confirm the existence of MFe 2 O 4 , Raman spectroscopy was employed. As is shown clearly in Fig. 1(b), there are three Raman peaks at 241.4, 445.6 and 609.5 cm −1 for all samples, which are assigned to the Raman active modes of rutile 30 . This result indicates that the rutile phase dominates the crystalline structure of the samples, which is in accordance with the XRD result. However, the Raman peak corresponding to MFe 2 O 4 is not discernable due to the low content of MFe 2 O 4 . The peak at 117 cm −1 is due to plasma emission of the Ar + laser 31 .

Results and Discussion
In order to further confirm the existence of MFe 2 O 4 , XPS measurement was carried out. The XPS survey spectra are shown in Fig. 2 As is shown in Fig. 2 34,35 . Furthermore, the manganese valences were determined by the position of the multiplet splitting of Sr 2p peaks, the positions of Sr 2p3/2 and Sr 2p1/2 were all assigned to Sr 2+ . As for high-resolution XPS spectra of Fe 2p in Fig. 2(d), one can see that the peaks at ca. 711.6 eV and ca. 724.9 eV can be attributed to Fe 2p3/2 and Fe 2p1/2 for Fe 3+ , respectively, which reveals the oxidation state of Fe 3+ in the MFe 2 O 4 /TiO 2 heterostructure 33,36 .
The high resolution XPS spectra of Ti 2p, O 1s, and C 1s are shown in Fig. 3. The Ti 2p spectra, as presented in Fig. 3(a), all show the main peak located at ca. 458.5 eV and ca. 464.2 eV, which can be attributed to Ti 2p3/2 and Ti 2p1/2 in TiO 2 , respectively 37 . It is clear that the O 1s spectra of these MFe 2 O 4 /TiO 2 NRAs samples can be deconvoluted into two components centered at ca. 529.8 eV and ca. 531.4 eV using two Gaussian curve fittings {Fig. 3(b)}, The components at the lower and higher binding energy side can be assigned to the crystal lattice oxygen of TiO 2 and MFe 2 O 4 and chemisorbed oxygen in a defective lattice site (i.e.-OH), respectively 32,[38][39][40][41] . It is suggested that the hydroxyl group can capture the photogenerated holes and form highly reactive hydroxyl free radicals, which plays an important role in enhancing photocatalytic activity 18 . The high resolution XPS spectrum of C 1s is shown in Fig. 3(c). The primary peak located at ca. 284.6 eV is assigned to C-C/C-H bonds from adventitious carbon 42 , while the peaks at ca. 286.2 eV and ca. 288.4 eV can be attributed to the formation of carbonate species, resulting mainly from CO 2 adsorption 38,43-45 . Especially, the peak at 288.4 eV can be ascribed to the Ti-O-C structure in carbon doped TiO 2 by substituting some of the lattice titanium atoms [46][47][48] . Interestingly, carbon doping is beneficial to light absorption capability as well as absorption of organic molecules to some extent 24,25 .
The SEM images of the bare TiO 2 NRAs and MFe 2 O 4 /TiO 2 NRAs are shown in Fig. 4. It is noteworthy that, after MFe 2 O 4 modification as shown in Fig. 4(c-f) from the top view images, the samples have no obvious changes in morphology compared with the bare TiO 2 NRAs in Fig. 4(a), which indicates that the deposited MFe 2 O 4 nanoparticles are of extremely fine size. The vertically or slantingly aligned TiO 2 nanorods arrays, with diameter of 60~120 nm and length of 2.2 μm, are grown homogeneously on FTO substrate with rectangular cross section. In order to measure the content of MFe 2 O 4 in MFe 2 O 4 /TiO 2 NRAs heterjunction, energy dispersive x-ray spectrum (EDS) analysis was carried out. The results, shown in Fig. 4(g~j), are obtained from collecting the EDS data in red square region of the MFe 2 O 4 /TiO 2 NRAs in Fig. 4(c-f) Figure 5(a) shows the TEM image of the bare TiO 2 nanorod. Essentially, the diameter of the bare TiO 2 nanorod under TEM observation is consistent with the SEM result. It can be seen clearly that the bare TiO 2 nanorod is very  Supplementary Fig. S1), and all show the same morphology. i.e., the smooth surface of TiO 2 nanorod become rough after MFe 2 O 4 modification. The corresponding lattice fringes of CoFe 2 O 4 , ZnFe 2 O 4 and SrFe 2 O 4 are shown in Figure S1(b,d,f), respectively.
The optical absorption spectra of TiO 2 NRAs and MFe 2 O 4 /TiO 2 NRAs are shown in Fig. 6. All samples exhibit typical UV absorption (λ < 380 nm). It is noteworthy that, compared with bare TiO 2 NRAs, all MFe 2 O 4 /TiO 2 samples exhibit strong light absorption in a wide region from 380 nm to 900 nm, which can be attributed to the intrinsic band gap absorption of MFe 2 O 4 . However, unlike other pure TiO 2 , tiny absorption of the as-prepared TiO 2 sample in the visible light range can be observed. There are two reasons accounting for this abnormal phenomenon, one is the scattering of light caused by the nanorod arrays, and the other is the impurity doping during the hydrothermal and sintering process [49][50][51]   Though all MFe 2 O 4 modified TiO 2 NRAs samples exhibit a broader and stronger absorption than the bare TiO 2 NRAs (see Fig. 6), only NiFe 2 O 4 /TiO 2 NRAs possesses a significant enhancement in PEC performance. Very limited improvement for CoFe 2 O 4 modification may result from the inefficient separation of photoexcited charge carriers. This phenomenon is due to the fact that the conduction band (CB) of CoFe 2 O 4 is more positive than that of TiO 2 , while the valence band (VB) of CoFe 2 O 4 is more negative than that of TiO 2 41,57 , which is not favour in carriers separating.
To investigate the photocatalytic capacity of the MFe 2 O 4 /TiO 2 NRAs, experiments were carried out for Cr(VI) photoreduction under visible light irradiation. The concentration changes are detected by the absorption peak (365 nm) of Cr(VI) in the UV-vis spectrum. The photodegradation results are shown in Fig. 7(c). After irradiation for 180 minutes, little Cr(VI) was reduced without catalyst (the reduction rate is only 3.8%). Under the same condition, only 45.1% of Cr(VI) was reduced when bare TiO 2 NRAs was used as a photocatalyst. However, the potoreduction capacity of NiFe 2 O 4 , ZnFe 2 O 4 and SrFe 2 O 4 modified TiO 2 NRAs are enhanced greatly (94.18%, 94.086% and 92.39%, respectively), reaching the same level. This may be attributed to the function of citric acid serving as a sacrificial electron donator to quickly consume the photogenerated holes 19 , thus greatly promote charge separation and further improv photocatalytic reactions. Unfortunately, CoFe 2 O 4 modification makes the photocatalytic degradation rate of Cr(VI) even lower. The following reason may account for this abnormal phenomenon. Eventhough CoFe 2 O 4 modified TiO 2 NRAs can be excited more easily under visible light irradiation, and then generates more charge carriers, the recombination rate of CoFe 2 O 4 /TiO 2 NRAs seems to be higher than that of the bare TiO 2 NRAs which can be deduced from the Voc changes, thus leading to the lower photocatalytic capacity of CoFe 2 O 4 /TiO 2 NRAs.
In order to clarify the enhancement in the phototelectochemical and photocatalytic capacity of TiO 2 NRAs after MFe 2 O 4 modification, it is important to figure out the separating and transferring efficiency of the charge carriers, so electrochemical impedance spectroscopy (EIS) measurements were conducted. As shown in Fig. 7(d), except for CoFe 2 O 4 /TiO 2 NRAs, other MFe 2 O 4 modified TiO 2 NRAs samples all have a smaller arc radius compared with that of the bare TiO 2 NRAs. It is generally assumed that the smaller arc radius on the EIS Nyquist plot suggests a more effective separation of the photogenerated electron-hole pairs and a faster interfacial charge transfer 38,58 . From the EIS spectra, it can be seen clearly that      Synthesis of MFe 2 O 4 modified TiO 2 nanorod arrays. Aligned TiO 2 NRAs were vertically grown on transparent fluorine-doped tin oxide (FTO) substrates by the hydrothermal method. Deionized water (DI, 10 mL) was mixed with hydrochloric acid (36.8 wt%, 10 mL) and stirred for 10 min before tetrabutyl titanate (98%, 0.4 mL) was added. When the solution was stirred until clear clarification, it was transferred to a Teflon-lined stainless steel autoclave. Clean FTO substrates were immersed with the conducting side face down. The autoclave was put in an oven at a temperature of 150 °C and was taken out from the oven after 5 h. After the autoclave was cooled to room temperature, the FTO substrate was rinsed with DI water and dried naturally at room temperature. The final area of the nanorod arrays was approximately 4.5 cm 2 .
For the preparation of ZnFe 2 O 4 /TiO 2 NRAs, briefly, zinc nitrate and iron nitrate were dissolved in DI water at room temperature to form a mixture, the as-prepared TiO 2 NRAs were soaked in the Fe(NO 3 ) 3 and Zn(NO 3 ) 2 mixed solution (with concentrations of 0.25 M and 0.125 M, respectively) for 1 h, followed by dipping in DI water for 5s. Afterwards the nanorod arrays were dried in air for 24 h and then annealed at 500 °C in air for 2 h with heating and cooling rates of 5 °C·min −1 . The MFe 2 O 4 /TiO 2 NRs (M = Ni, Co and Sr) were prepared using the same method by replacing the zinc nitrate with other nitrate.
Characterization. The surface morphology was obtained with a scanning electron microscopy (SEM, VEDAIIXMUINCN) equipped with an energy dispersive X-ray spectroscopy (EDS) system. The film microstructure was further characterized by transmission electron microscopy (TEM). X-ray diffraction (XRD, PANalytical) with Cu-Ka (λ = 0.15401 nm) was operated at 40 kV and 40 mA in a 2θ range of 20-80° at a scanning speed of 5° min −1 to characterize the crystal structure. Raman spectra were recorded at room temperature using a inVia Reflex Raman spectrometer under Ar + (532 nm) laser excitation. The optical properties were probed by a UV-vis spectrophotometer (UV1800, Shimadzu) with a FTO substrate as a blank. X-ray photoelectron spectroscopy (XPS) was obtained using a ESCALAB 250Xi (The binding energy of the XPS spectra was calibrated with the reference to the C 1s peak at 284.8 eV.) Photoelectrochemical and photocatalytic measurement. photoelectrochemical measurements were performed in a 250 mL quartz cell using a three-electrode configuration, including the prepared samples as working electrode, a Pt foil as counter electrode, a saturated Ag/AgCl as reference electrode, and 0.5 M Na 2 SO 4 aqueous solution as an electrolyte. The working electrode was illuminated within an area of about 1.5 cm 2 at zero bias voltage versus the Ag/AgCl electrode under solar-simulated (AM 1.5 G filtered, 100 mW·cm −2 , CEL-HXF300) light sources with a UV cutoff filter (providing visible light with λ ≥ 420 nm). The electrochemical impedance spectroscopy (EIS) measurements were recorded by employing an AC voltage of 5 mV amplitude with the initial potencial at 0.4 v (vs. Ag/AgCl) over the frequency range from 100 kHz to 100 mHz without light illumination.
The Cr(VI) photoreduction was performed in a quartz cell. In the photoreduction experiments, 15 mL of aqueous solution containing 20 mg·L −1 of K 2 Cr 2 O 7 and 85 mg·L −1 of citric acid was used. The citric acid served as a sacrificial electron donator. Prior to irradiation, the photocatalyst (area about 6 cm 2 ) was immersed into the Cr(VI) solution in the dark for 30 minutes to establish an adsorption/desorption equilibrium. The relative concentration of Cr(VI) in the solution was derived by comparing its UV-vis absorption intensity with that of the initial Cr(VI) solution at 365 nm. The light source was a 300 W xenon lamp with visible light illumination of 26.5 mW·cm −2 .