Fabrication of γ-Fe2O3 Nanowires from Abundant and Low-cost Fe Plate for Highly Effective Electrocatalytic Water Splitting

Water splitting is thermodynamically uphill reaction, hence it cannot occur easily, and also highly complicated and challenging reaction in chemistry. In electrocatalytic water splitting, the combination of oxygen and hydrogen evolution reactions produces highly clean and sustainable hydrogen energy and which attracts research communities. Also, fabrication of highly active and low cost materials for water splitting is a major challenge. Therefore, in the present study, γ-Fe2O3 nanowires were fabricated from highly available and cost-effective iron plate without any chemical modifications/doping onto the surface of the working electrode with high current density. The fabricated nanowires achieved the current density of 10 mA/cm2 at 1.88 V vs. RHE with the scan rate of 50 mV/sec. Stability measurements of the fabricated Fe2O3 nanowires were monitored up to 3275 sec with the current density of 9.6 mA/cm2 at a constant potential of 1.7 V vs. RHE and scan rate of 50 mV/sec.

In contrast, electrochemical water splitting method has high efficiency, excellent adaptability and flexibility, which can efficiently produce hydrogen with high purity. In electrochemical water splitting, anodic oxygen evolution reaction (OER) and cathodic hydrogen evolution reaction (HER) are two critical half-cell reactions. The standard oxidation potential for OER is 1.23 V vs. RHE and standard reduction potential for HER is 0 V vs RHE for electrochemical water splitting. But in practice, it should be larger than 1.23 V to avoid some unfavorable factors such as activation energy, ion and gas diffusion, electrolyte concentration, wire and electrode resistances, electrolyte diffusion blockage, bubble formation and thermodynamic losses, which leads additional potential over the standard potential 13 . Initially, anodic and cathodic reactions were catalyzed by Ru, Ir and their oxides catalysts 14,15 . However, these noble metals cannot be used for commercial applications due to their high cost and lack of abundance. At this situation transition metals such as Ni, Co, Fe, Mn, W, etc. and their oxides have been extensively used as electrocatalytic materials for effective electrochemical water splitting 16,17 . Among them iron oxides received greater attention to split water efficiently. In addition, it is highly abundant, non-toxic to the environment, cost effective and also stable in aqueous solution 18,19 . Moreover, the demerit like slow reaction kinetics can be solved by fabricating an electrode with nanostructured morphology. Based on this phenomenon wide varieties of nanostructures have been reported so far.
Recently, Sharifi et al. 20 reported the current density of γ-Fe 2 O 3 modified carbon nanotubes as 1 mA/cm 2 at 1.57 V vs RHE in 0.1 M KOH and Tavakkoli et al. 21 reported the current density of γ-Fe 2 O 3 nanoparticle decorated carbon nanotubes as 10 mA/cm 2 at the potential of 1.61 and 1.57 V vs. RHE in 0.1 and 1 M NaOH, respectively. Moreover, Ashwani Kumar et al. 22 reported the current density of NiFe-NC obtained from the composite of NiO and α/γ-Fe 2 O 3 as 10 mA/cm 2 at the potential of 1.67 V vs. RHE in 1 M KOH. Recently, Davodi et al. 23 reported that MWNTs functionalized with nitrogen-rich emeraldine salt (ES-MWNT) as a promising catalyst support to boost the electrocatalytic activity of magnetic maghemite (γ-Fe 2 O 3 ) NPs and the current density of the electrocatalyst (Ni@γ-Fe 2 O 3 /ES-MWNT) was measured as 10 mA/cm 2 at the potential of 1.49 V vs RHE in 1 M NaOH. In addition, Chandrasekaran et al. 24 employed γ-Fe 2 O 3 with reduced graphene oxide for PEC water splitting and the photocurrent density of the RGO/γ-Fe 2 O 3 nanocomposite was reported as 6.74 mA/cm 2 at 1.80 V vs RHE in 1 M NaOH. Even though these reports are available for electrocatalytic water splitting, still it is challenging to attain greater efficiency at low applied potential through highly available and cost-effective materials.
Usually, highly available and low-cost materials are sluggish in nature which needs some modifications, but the materials and methods used to modify the electrodes are highly expensive. Hence, we intended to use a pure, low cost and highly obtainable material without any additional modification with greater efficiency even at very low applied potential. Moreover, it has been reported that facile thermal oxidation is a simple and effective method to grow nanostructured materials including iron oxides [25][26][27] . Therefore, in the present study, we fabricated γ-Fe 2 O 3 nanowire array from low cost and easily available Fe plate by adopting simple thermal treatment. The formation of Fe 2 O 3 nanowires and its structural morphology was investigated by microscopic and spectroscopic techniques viz., Field emission scanning electron microscopy (FESEM), Energy-dispersive X-ray spectroscopy (EDS), Fourier-transform infrared spectroscopy (FTIR), X-ray diffraction (XRD), Transmission electron microscopy (TEM), Selected area electron diffraction pattern (SAED) and X-ray photoelectron spectroscopy (XPS). Finally, the current density of the obtained Fe 2 O 3 nanowires was measured by applying the materials as working electrode for electrocatalytic water splitting and the current density has reached as 10 mA/cm 2 at 1.88 V vs. RHE. The value obtained from bare Fe 2 O 3 is comparable with previous results, where chemical modifications or doping is necessary [20][21][22][23][24] .

Results and Discussion
Morphologies and structural characterizations. After thermal treatment, the morphology of Fe plate with scratched surface was compared with unscratched Fe plate using FESEM analysis (Fig. S1, Supporting Information). From the images it is observed that the scratched surface ( Fig. S1(a)) showed substantial growth of nanowires than unscratched surface ( Fig. S1(b)) and the corresponding mechanisms ( Fig. S2) for nanowire formation has drawn on the basis of experimental results and previous reports 11,12 . A reducing environment or straining the surface by applying external force has already been employed to prepare γ-Fe 2 O 3 [28][29][30] . Here, the surface of the Fe plates has been strained by uniform scratching, as a result, the effect of strained surface generates applied stresses. This applied stresses fasten the grain boundaries efficiently from the very beginning of the surface oxidation process. The strained surface encourages the initial compressive stress on the surface of the Fe plate, which clogged the volume expansion of oxide layer. Consequently, the vertical stress gradient occurred and expedites the iron ion diffusion and followed to improve the nanowire growth. Moreover, the strained surface can enhance the surface roughness of the Fe plate and thus expand the volume of Fe 2 O 3 layer. The enhanced volume expansion of Fe 2 O 3 layer can increase the tensile strength of the Fe plate and thus facilitates the driving force of iron ion diffusion. In unscratched Fe 2 O 3 , grain boundaries have been created by the effect of internal stresses at a respective temperature. The obtained grain boundaries provides route to iron ion diffusion. However, the inadequate internal stress generated in unscratched Fe 2 O 3 produces lesser grain boundaries for iron ion diffusion and thus indicates fewer growths of nanowires in Figs. S1(b) and S2(b).
The schematic diagram for nanowire formation is shown in Fig. 1 Fig. S8. In Fig. 2, the obtained images of Fe 2 O 3 -1 showed highly dense and ordered growth of Fe 2 O 3 nanowires on iron plate than Fe 2 O 3 -2. The specific reason for the growth of nanowire is atomic diffusion of Fe atom due to the existence of compressive stress. Also, the continuous supply of Fe and O at higher temperature for prolonged duration enhances the growth of nanowires 31,32 .
In briefly, the as-received Fe-sheet was subjected to pretreatment using distilled water and ethanol and then individually employed for annealing process at 480 and 550 °C for 2 h and thus obtained Fe 2 O 3 -1 and Fe 2 O 3 -2, respectively. The thickness of the as-prepared samples viz., Fe 2 O 3 -1 and Fe 2 O 3 -2 was measured by FESEM analysis. Moreover, to measure the thickness of each layer obtained after thermal treatment, the side view of the cross-sectional FESEM images were taken by cutting the samples as shown in Fig 511) and (440) crystal planes of γ-Fe 2 O 3 , respectively. In addition, the peak observed at 44.3° for the crystal plane of (110) in Fe 2 O 3 -1 belongs to iron plate. In Fe 2 O 3 -2, the new peak observed at 65.0° was relevant to the crystal plane of (300) of α-Fe 2 O 3 . These peaks and its planes are agreed well with the standard XRD database of γand α-Fe 2 O 3 (JCPDS no: 39-1356 and 33-0664) 34 . Hence, the disappearance of Fe peak and the appearance of new peak for α-phase clearly indicate the phase transition of iron oxide (γ to α) from lower temperature to higher temperature. Usually, Fe has various oxidation states but the common oxidation state of Fe is +2 and +3. Initially Fe donate 2e − to oxygen to attain Fe 2+ oxidation state and which produce FeO, when increasing the temperature and duration, FeO replaced by Fe 3 O 4 in such case the oxidation state of Fe in Fe 3 O 4 (Magnetite) is +2 and +3, and it is ferrimagnetic in nature, further heating leads β/γ-Fe 2 O 3 , but β and γ phases are intermediate phases. Since β-phase is highly unstable, not easy to separate, but there is some possibility of getting γ-Fe 2 O 3 (Maghemite), and it is also ferrimagnetic in nature but the oxidation state is +3. At higher temperature, γ-phase is obviously replaced by the stable α-phase (hematite). The phase transition temperature can vary from materials to materials based on the thickness and purity of the Fe-precursor which have been chosen for analysis and the equipment (oven/furnace/ceramic heater/CVD) used for annealing process.
The morphology, size and shape of single nanowire were investigated by TEM analysis. As shown in Fig. 5(a,c) the morphology of the nanowires were found smooth. The diameter and length of Fe 2 O 3 -1 was measured as 397 nm and 1.14 µm (Fig. 5(a)). Whereas, the TEM image of Fe 2 O 3 -2 exhibit different sized nanowires, which are due to the fact that not all nanowires were nucleated at the same time. In Fig. 5(c), the diameter and length of the longer one was found as 10 nm and 1.21 µm, respectively, and the shorter one showed the diameter and length of 19 and 640 nm, respectively. It is suggest that the thinner nanowires are longer than thicker one at a specific oxidation temperature and time 35 . On comparing the TEM results of Fe 2 O 3 -1 and Fe 2 O 3 -2, the nanowires found in Fe 2 O 3 -2 was thinner and longer than Fe 2 O 3 -1 nanowire. The reason is considered to be due to the increased compressive stress with the increase of temperature from 480 to 550 °C which influenced the growth of longer nanowires observed in Fe 2 O 3 -2 than that of Fe 2 O 3 -1 31,32 . The crystal planes and crystalline structure of single Fe 2 O 3 nanowire was further confirmed by SAED pattern. As shown in Fig. 5(b,d), the diffraction spots obtained from Fe 2 O 3 -1 and Fe 2 O 3 -2 were indexed as (220), (311) and (440) planes by measuring the distance of the spots using its appropriate d-spacing, which are corresponding to the cubic crystal cells of γ-Fe 2 O 3 . Further, the obtained results clearly indicate that the fabricated γ-Fe 2 O 3 nanowires were existed in single crystalline structure and the results are agreed well with previous reports 36,37 .
The biding energy, oxidation state and phase transition of Fe 2 O 3 -1 and 2 were further verified by XPS analysis. In Fig. 6(a), the survey spectra of Fe 2 O 3 -1 and Fe 2 O 3 -2 showed the characteristic peaks at 530.3, 710.3 and 724.3 eV corresponding to O 1s, Fe 2p 3/2 and Fe 2p 1/2 , respectively. The deconvoluted XPS spectra of both Fe 2 O 3 -1 and 2 showed the major peaks for Fe 2p 3/2 at 709.9 eV and Fe 2p 1/2 at 723.3 eV (Fig. 6(b)). Further, the charge transfer satellite peak for Fe 2p 3/2 and Fe 2p 1/2 was observed at 717.8 eV and 731.8 eV which are strong evidence for the oxidation state of Fe 3+ in γ-Fe 2 O 3 . This observation was agreed well with previous literatures 38,39 . Also, Wang et al. 40 and Pereira et al. 41 (Fig. 6(c)). This interference was matched with the reported results 35,42,43 . On comparing the results, the nanowires obtained from Fe 2 O 3 -1 and Fe 2 O 3 -2 showed similar characteristic peaks in XPS spectra which indicate that the two samples were exist in same form and also in same phase 35 . electrochemical measurements. The electrochemical water splitting measurements were performed in 1 M NaOH electrolyte solution at pH 13.5 with the help of conventional three electrode system. The fabricated Fe 2 O 3 -1 and Fe 2 O 3 -2 nanowires were applied directly as working electrode without any chemical modification or doping, which is highly advantageous for large scale application in near future and also it can reduce the cost of expensive chemical substances/dopants to modify the electrode surface. As shown in Fig. 7, the linear sweep voltammograms (LSV) of Fe 2 O 3 -1 and Fe 2 O 3 -2 illustrate the enriched current density of 10 mA/cm 2 at the potentials of 1.88 and 1.91 V vs. RHE with 50 mV/sec scan rate. Even though the applied potential is higher than previous results obtained from chemically modified or doped γ-Fe 2 O 3 nanostructures, no one has reported improved performance for bare Fe 2 O 3 electrode (Table 1) [20][21][22][23][24] . Specifically, three important factors such as, (i) highly dense and ordered growth of nanowires, (ii) more abundant of γ-phase and (iii) single crystalline structure played essential role to improve the performance of the as-prepared Fe 2 O 3 electrodes. In detail, the highly dense and ordered growth of Fe 2 O 3 nanowires (obtained after scratching the electrode surface, Figs. S1(a) and S2(a), Supporting Information) produced high surface area. The more abundant γ-phase on Fe 2 O 3 electrode acts as an efficient electrocatalyst for oxygen evolution reaction due to its greater electrical conductivity for rapid electron transfer. Also, γ-Fe 2 O 3 provided substantial and synergetic effect on improving catalysis by producing more catalytic active sites. Therefore, the enormous surface area and the enhanced electrocatalytic active sites offered more capability to adsorb OH − ions (presented in the electrolyte solution) which leads the enhancement in catalysis and fast electron transfer towards oxygen evolution reaction, as a result of the improved performance in Fe 2 O 3 electrode (Fig. 7). Moreover, the sharp edges of the Fe 2 O 3 nanowires highly supported to the charge transfer process of electrode/electrolyte interface i.e. electrons obtained from oxygen evolution reaction were easily moved from electrolyte solution to Fe 2 O 3 electrode. In addition, the single crystalline structure of Fe 2 O 3 electrode (Fig. 5) played vital role to enhance the electron movements from Fe 2 O 3 to Fe layer and which also improved the performance of the as-prepared Fe 2 O 3 electrode.
On comparing the results obtained from electrochemical measurements, a small potential difference has been noticed between Fe 2 O 3 -1 and Fe 2 O 3 -2 (Fig. 7). This is due to non-uniform and lack of nanowires growth, and increased thickness of Fe 2 O 3 layer found in Fe 2 O 3 -2. The non-uniform and lack of nanowires growth slightly reduced the surface area and catalytic active sites of the electrode, and the thickness slowdown the electron movement from Fe 2 O 3 to Fe layer during water splitting. Therefore, Fe 2 O 3 -2 electrode showed slight increment in   www.nature.com/scientificreports www.nature.com/scientificreports/ respectively 44,45 , and the observation was explained by the following possible reaction mechanism 13 . The OER reaction on γ-Fe 2 O 3 was also supported by previous report 21 . Specifically, catalytic reactions involved in four steps such as (a) adsorption of reactant molecules onto catalyst surface, (b) diffusion of reactants onto catalyst surface, (c) reaction takes place from reactants and which produce products, and finally (d) desorption of products from catalyst surface. The electrocatalytic measurement of the present investigation has been performed under alkaline condition, therefore the OER reaction proceeds by the following reaction steps: (1) Adsorption of 'OH − ' (present in alkaline electrolyte solution) onto γ-Fe 2 O 3 catalyst surface.
Moreover, it is value to mention here that α-Fe 2 O 3 (hematite) offers a favorable combination of good visible light absorption up to 590 nm hence it has been widely used as photoactive material for solar water splitting 46 , but recently few reports are demonstrated the excellent electrocatalytic activity of chemically modified γ-Fe 2 O 3 in electrocatalytic water splitting [20][21][22][23][24] . The morphology of γ-Fe 2 O 3 electrode used for stability measurements was observed by FESEM analysis and the images are shown in Fig. S16 (Supporting Information). This image shows that there was no substantial change in the electrode morphology and the result reveals the best stability of γ-Fe 2 O 3 electrode in alkaline solution and the resistivity against corrosion. The samples after stability measurements were also measured by XPS analysis and the results are shown in Fig. S17 (Supporting Information). These results suggested that γ-Fe 2 O 3 NWs are highly efficient material to achieve excellent current density and stability. Moreover, the single crystalline structure of γ-Fe 2 O 3 is also an important factor to increase the efficiency of electrocatalytic water splitting, however, the preparation of γ-Fe 2 O 3 with single crystalline structure is still a challenging so far 36,37 .
In addition, the electrochemical capacitances have been calculated through double layer capacitance of the fabricated electrode material in 1 M NaOH 22,47,48 . To measure double layer capacitance, cyclic voltammetry measurements were performed at non-faradaic region (i.e., −0.090 to 0.110 V vs. Ag/AgCl is converted as 0.904 to 1.104 V vs. RHE) with different scan rates viz., 5, 10, 20, 50, 100, 200, 500 and 1000 mV/sec as shown in Fig. 10. All measured current at this non-faradaic region was assumed to as double-layer charging current. Then the cathodic and anodic capacitance current was measured at 0.026 V vs Ag/AgCl (i.e. 1.020 V vs. RHE) from each scan rate. The obtained capacitance currents were plotted against scan rates (Fig. 11). Then the double-layer capacitance was calculated from the average absolute value of the both cathodic and anodic slope of the linear fitting of the plot, i.e. double layer capacitance (C DL ) was measured from the slope of charging currents (i c ) as a function of scan rate (υ) as shown in Eq. 1.
The electrochemical double-layer capacitance measured from the scan-rate study for the Fe 2 O 3 -1 catalyst was C DL = 0.119 mF. The electrochemically active surface area (ECAS) of Fe 2 O 3 -1 was measured from the obtained electrochemical double-layer capacitance of the catalytic surface by using Eq. 2.
DL s Where C s is the specific capacitance of the sample or the capacitance of a planar surface per unit area under identical electrolyte conditions. The average specific capacitance of 1 M NaOH was reported as 0.040 mF cm −2 21,47,48 . www.nature.com/scientificreports www.nature.com/scientificreports/ Then, the electrochemically active surface area of Fe 2 O 3 -1 was measured as ECAS = 3 cm 2 by applying the C s value of 1 M NaOH in Eq. 2. Further, the roughness factor (RF) of the Fe 2 O 3 -1 electrode has been calculated by using Eq. 3.
where GSA is geometric surface area, the geometric surface area of the catalytic material used in this study was 0.25 cm 2 . Then, the roughness factor (RF) of the working electrode Fe 2 O 3 -1 was measured as RF = 12. Therefore, it has been cleared that the measured electrochemically active surface area (3 cm 2 ) and surface roughness (12) of Fe 2 O 3 -1 played vital role to contribute to the high activity of the fabricated material. The higher activity of the fabricated Fe 2 O 3 electrocatalyst was corresponding to the existence of γ-phase in Fe 2 O 3 -1. Tavakkoli et al. 21 also reported that γ-Fe 2 O 3 nanoparticles showed higher activity than α-Fe 2 O 3 .

conclusions
In conclusion, we have successfully fabricated highly effective, easily available and low-cost γ-Fe 2 O 3 nanowires by adopting a simple thermal oxidation method. The current density of the newly fabricated nanowires array was measured as 10 mA/cm 2 at 1.88 V vs. RHE with the scan rate of 50 mV/sec in 1 M NaOH solution and the stability of the working electrode were maintained up to 3275 sec and the current density was observed as 9.6 mA/cm 2 . This result is in a similar or higher order with that of other reported chemically modified/doped γ-Fe 2 O 3 nanostructured electrodes. The unique electrocatalytic nature, γ-phase, high density and single crystalline structure of the fabricated γ-Fe 2 O 3 nanowires played a vital role to obtain higher current density and long durability. The obtained simple bare electrode can definitely possess cost effectiveness, high availability, higher energy conversion, compatibility and activity than existing modified electrodes. Hence, the newly fabricated Fe 2 O 3 nanowires obtained without using any co-catalyst, doping agents or any other modification onto electrode materials are expected to be a prominent candidate for electrochemical water splitting in near future compared with existing modified/doped nanostructured materials.

Methods
The schematic representation for the fabrication of γ-Fe 2 O 3 was shown in Fig. 1. Initially, the commercially available Fe plate with 99.5% purity was cut into small pieces (5 mm × 10 mm), and the surface of the Fe plate (front side) which is used for electrolysis was scratched by knife. Then the plate was cleaned by dispersing with ethanol and distilled water separately and dried under room temperature. Afterwards it was introduced for thermal treatment i.e. the pretreated Fe plate was placed on ceramic boat and heated for 2 h at 480 °C by using electric furnace under air. After thermal treatment the steel blue color of the Fe plate changed into dark brown color, and which has been named as Fe 2 O 3 -1. Similarly, Fe 2 O 3 -2 has been prepared by treating the Fe plate for 2 h at 550 °C. The obtained samples were analyzed by spectroscopic and microscopic techniques. All the time three set of samples have been prepared in both experimental condition and employed for spectroscopic and microscopic analysis. The morphology and elemental mapping of the samples were monitored by Field Emission Scanning Electron Microscopy (FESEM) combined with Energy-dispersive X-ray spectroscopy (EDS) using JSM-7200F. Then the different iron oxides group present in the sample was observed by Fourier-transform infrared spectroscopy (JASCO FT/IR-4100). The crystalline properties of the Fe 2 O 3 -1 and Fe 2 O 3 -2 were investigated by X-ray diffraction analysis (XRD, ATX-G, RIGAKU with Cu Kα1). Then, the morphology, structure, size and the corresponding Selected Area Electron Diffraction pattern (SAED) of a single nanostructure was investigated by Transmission Electron Microscopy (TEM) analysis using JEM-2100. The binding affinity of iron and oxygen present in γ-Fe 2 O 3 and the oxidation state of Fe in γ-Fe 2 O 3 was further verified by X-ray Photoelectron Spectroscopic (XPS) analysis (ESCALAB 250). Electrochemical water splitting measurements were measured by using potentiostat (ALS/ Figure 11. Anodic and cathodic capacitance currents measured for Fe 2 O 3 -1 at 0.026 V vs. Ag/AgCl and plotted as a function of scan rate.