Promoting Photoelectrochemical Water Oxidation on Ti-Doped Fe 2 O 3 Nanowires Photoanode by O 2 Plasma Treatment

: Surface electron traps on semiconductor photoanodes mediate surface recombination and deteriorate the photoelectrochemical (PEC) water oxidation performance of the photoanode. Developing convenient methods to reduce surface electron traps is therefore essential for high efﬁciency PEC water oxidation on semiconductor photoanodes, particularly for nanostructured photoanodes with large surface area. Herein, we employ a O 2 plasma treatment to boost the PEC water oxidation performance of Ti-doped Fe 2 O 3 (Ti-Fe 2 O 3 ) nanowires photoanodes, aiming to reduce surface oxygen vacancies, the dominant electron traps on Ti-Fe 2 O 3 surface. X-ray diffraction (XRD), scanning electron microscopy and spectroscopic analyses show that the oxygen plasma treatment changes the structural, morphological and optical properties negligibly, but it does reduce the content of surface oxygen vacancies, as estimated from O1s X-ray photoelectron spectroscopy spectra. An optimal O 2 plasma treatment (200 W, 70 s) increases the photocurrent density of the Ti-Fe 2 O 3 nanowire photoanode to 2.14 mA · cm − 2 (1.23 V vs. RHE) under air mass 1.5G simulated solar light, which is 1.95 times higher than the pristine Ti-Fe 2 O 3 nanowire photoanode. The surface hole transfer efﬁciency is also improved by 1.66 times due to the reduced surface recombination. The work suggests that O 2 plasma treatment is a convenient but effective method to boost the PEC water oxidation performance of Ti-Fe 2 O 3 photoanode and might be applicable to other semiconducting oxide photoanodes for high efﬁciency PEC water splitting.


Introduction
Photoelectrochemical (PEC) water splitting into hydrogen and oxygen with a semiconductor photoanode is regarded as a promising strategy to provide clean and sustainable energy since it was first reported by Fujishima and Honda in 1972 [1], in which a highly efficient photoanode plays a decisive role. The nanostructured Fe 2 O 3 photoanode has attracted a large number of investigations due to its suitable photon absorption (2.0-2.2 eV), shortened minority diffusion distance, and large contact area with the electrolyte for surface reactions [2][3][4][5]. However, the large surface area of nanostructural Fe 2 O 3 usually introduces many surface states, such as surface oxygen vacancies and Fe 2+ sites, which mediate the unwanted surface charge recombination [6][7][8][9]. Thus, it is an essential issue to reduce the surface states of the Fe 2 O 3 photoanode for high efficiency PEC water splitting.
Surface modification of the passivation layer and loading of the cocatalyst are usually adopted to reduce the surface recombination mediated by surface oxygen vacancies and improve the PEC performance of Fe 2 O 3 photoanodes [10][11][12][13][14]. For example, Zou et al. demonstrated that the ion-impermeable Al 2 O 3 layer on Ti-Fe 2 O 3 nanorods passivated the surface states and suppressed the back reaction, which improved the photocurrent onset potential [14]. The result proved the necessity of surface states treatment of Ti-doped Fe 2 O 3 nanowire photoanodes. However, the methods of surface loading the passivation layer and cocatalyst have some drawbacks, introducing a barrier in the interface or reducing the absorption of photons. In addition, Yu et al. reported that the most significant issue in suppressing the surface states of Fe 2 O 3 was to reduce the oxygen vacancies on the surface of the electrode [15]. Therefore, it is still necessary to develop a convenient and effective method to controllably tune the surface oxygen vacancy states of Fe 2 O 3 photoanodes.
Herein, we employed oxygen plasma treatment to reduce the surface oxygen vacancy of Ti-doped Fe 2 O 3 (Ti-Fe 2 O 3 ) nanowire photoanodes. Oxygen plasma, consisting of highly reactive oxygen ions with large concentration and kinetic energy, is expected to oxidize the surface of Ti-Fe 2 O 3 in a convenient way without introducing impurities. The effect of oxygen plasma on the surface states and PEC performance of Ti-Fe 2 O 3 photoanodes were investigated. At an optimal treatment condition (200 W, 70 s), the PEC performance of Ti-Fe 2 O 3 photoanodes was remarkably improved, exhibiting a 1.95 times higher photocurrent than the pristine Ti-Fe 2 O 3 photoanode at the potential of 1.23 V vs. reversible hydrogen electrode (RHE). The surface hole transfer efficiency of the plasma-treated Ti-Fe 2 O 3 (OPT-Ti-Fe 2 O 3 ) photoanode reached 78%, 1.66-fold higher than the pristine one. O1s X-ray photoelectron spectroscopy (XPS) analysis showed that the integrated area ratio of oxygen vacancy to lattice oxygen at the surface of the OPT-200W-70s electrode was reduced by 6.7%. The work suggests that O 2 plasma treatment is a convenient but effective method for tuning the surface defect states of semiconducting oxides photoanode for enhanced PEC water splitting. Crystal structure of the photoanodes was characterized by Raman spectroscopy and X-ray diffraction (XRD). The Raman spectra of the photoanodes are depicted in Figure 2a. Four phonon lines at 222, 288, 405, and 604 cm −1 appeared in the three spectra of the photoanodes, which corresponded to the A 1g (1), E g (2), E g (3), and E g (4) modes of hematite [16]. There were no obvious differences between the photoanodes with or without O 2 plasma treatment. This demonstrated that plasma treatment has little effect on the crystal structure of the films. XRD patterns of the photoanodes are shown in Figure 2b. Five distinct diffraction peaks of hematite could be observed in the spectra, and these peaks corresponded to the (012), (104), (110), (024), and (330) facets of Fe 2 O 3 [17]. There were some other diffraction peaks originated from the Sb:SnO 2 substrate, and no peak related to the Ti dopant was observed. The result demonstrated that the diffraction peaks of Ti-Fe 2 O 3 photoanodes had no obvious change after plasma treatment.

Results and Discussion
We also compared the absorption spectra of the Ti-Fe 2 O 3 photoanodes with and without O 2 plasma treatment. The absorption spectra (Figure 2c) of the three films were almost the same. The absorption threshold of the films all appeared at around 600 nm, which corresponded to the bandgap of 2.1 eV [18]. This was consistent with the bandgap of hematite. Based on these results, it can be concluded that the morphology, structure, and optical properties of Ti-doped Fe 2 O 3 electrodes change negligibly after O 2 plasma treatment. Crystal structure of the photoanodes was characterized by Raman spectroscopy and X-ray diffraction (XRD). The Raman spectra of the photoanodes are depicted in Figure 2a. Four phonon lines at 222, 288, 405, and 604 cm −1 appeared in the three spectra of the photoanodes, which corresponded to the A1g(1), Eg(2), Eg(3), and Eg(4) modes of hematite [16]. There were no obvious differences between the photoanodes with or without O2 plasma treatment. This demonstrated that plasma treatment has little effect on the crystal structure of the films. XRD patterns of the photoanodes are shown in Figure 2b. Five distinct diffraction peaks of hematite could be observed in the spectra, and these peaks corresponded to the (012), (104), (110), (024), and (330) facets of Fe2O3 [17]. There were some other diffraction peaks originated from the Sb:SnO2 substrate, and no peak related to the Ti dopant was observed. The result demonstrated that the diffraction peaks of Ti-Fe2O3 photoanodes had no obvious change after plasma treatment.  We also compared the absorption spectra of the Ti-Fe2O3 photoanodes with and without O2 plasma treatment. The absorption spectra (Figure 2c) of the three films were almost the same. The absorption threshold of the films all appeared at around 600 nm, which corresponded to the bandgap of 2.1 eV [18]. This was consistent with the bandgap of hematite. Based on these results, it can be concluded that the morphology, structure, and optical properties of Ti-doped Fe2O3 electrodes change negligibly after O2 plasma treatment.

Effect of O2 Plasma on the PEC Performance of Ti-Fe2O3 Photoanodes
The conditions of O2 plasma treatment were first optimized . The linear scan voltammetry (LSV) curves of pristine Ti-Fe2O3 and OPT-Ti-Fe2O3 photoanodes were performed in 1 M NaOH (PH = 13.6) aqueous solution at a scan rate of 10 mVs −1 , and the results are presented in Figure S1. The optimal films of OPT-100W-150s and OPT-200W-70s treated   Figure 3a. It can be seen that the OPT-200W-70s photoanode showed the highest photocurrent density of 2.14 mA·cm −2 at 1.23 V vs. RHE, which was 1.95 times higher than 1.10 mA·cm −2 of the pristine Ti-Fe 2 O 3 . The onset potential of the OPT-200W-70s photoanode exhibited an approximate 50 mV cathodic shift compared with pristine Ti-Fe 2 O 3 , which lowered the applied potential to drive the PEC water oxidation reaction. It was also found that the photocurrent of OPT-200W-70s here showed greater superiority for water oxidation than many reported Ti-Fe 2 O 3 nanowires as depicted in Table S1 [5,[19][20][21]. This all suggests that O 2 plasma treatment is an effective method for promoting the PEC water oxidation kinetics of Ti-Fe 2 O 3 photoanodes. In addition, the stability of the Ti-Fe2O3 photoanodes was investigated, and the I-t curves of pristine Ti-Fe2O3 and OPT-200W-70s were measured at the constant potential of 1.23 V vs. RHE as depicted in Figure S2. The photocurrent density of the pristine Ti-Fe2O3 photoanode was kept at 1.1 mAcm −2 at the duration of 2000 s at 1.23 V vs. RHE. The OPT-200W-70s showed the photocurrent density of 2 mAcm −2 at the moment of light on, and then the photocurrent rapidly dropped in the first 500 s. After that, the photocurrent dropped slowly and almost stabilized at 1.72 mAcm −2 . Although the photocurrent dropped a little during the measurement process, it was still far higher than the pristine Ti-Fe2O3 photoanode at the same constant potential.
To investigate the effect of O2 plasma on the surface charge transfer of Ti-Fe2O3 photoanode, we further calculate the hole transfer efficiency from the I-V curves using H2O2 as hole scavenger. When adding H2O2 in the electrolyte, the photocurrent densities of pristine Ti-Fe2O3, OPT-100W-150s, and OPT-200W-70s tended to be almost the same, as shown in Figure 3b. It can be concluded that the carriers' generation and bulk charge transport in the three electrodes show little difference, because almost 100% of the photogenerated holes are considered to be extracted from the electrodes in the presence of the hole scavenger H2O2 in the electrolyte. The practical charge transfer efficiencies (ηtransfer) of these photoanodes are calculated using the ratio of JH2O/JH2O2, where the JH2O and JH2O2 denote the photocurrent density measured in 1 M NaOH aqueous solution without and with 0.5 M H2O2, respectively. The curves are depicted in Figure 3c [22][23][24]. It can be seen that OPT-200W-70s and OPT-100W-150s showed higher ηtransfer of 78% and 64% than the 47% of pristine Ti-Fe2O3 in the measured potential range, which were 1.66 and 1.36 times that of pristine Ti-Fe2O3. These were very close to the 1.95 and 1.5 times enhancements of photocurrents, correspondingly. The result reveals that O2 plasma treatment significantly increased the surface charge transfer efficiency of the photoanodes.
Electrochemical impedance spectra (EIS) of pristine and OPT-Ti-Fe2O3 photoanodes were also measured to further study the charge transport and transfer dynamics at a bias of 1.23 V vs. RHE under illumination conditions. The impedance spectra of the photoanodes are illustrated in Figure 4a, in which the impedance of Ti-Fe2O3 treated with In addition, the stability of the Ti-Fe 2 O 3 photoanodes was investigated, and the I-t curves of pristine Ti-Fe 2 O 3 and OPT-200W-70s were measured at the constant potential of 1.23 V vs. RHE as depicted in Figure S2. The photocurrent density of the pristine Ti-Fe 2 O 3 photoanode was kept at 1.1 mAcm −2 at the duration of 2000 s at 1.23 V vs. RHE. The OPT-200W-70s showed the photocurrent density of 2 mAcm −2 at the moment of light on, and then the photocurrent rapidly dropped in the first 500 s. After that, the photocurrent dropped slowly and almost stabilized at 1.72 mAcm −2 . Although the photocurrent dropped a little during the measurement process, it was still far higher than the pristine Ti-Fe 2 O 3 photoanode at the same constant potential.
To investigate the effect of O 2 plasma on the surface charge transfer of Ti-Fe 2 O 3 photoanode, we further calculate the hole transfer efficiency from the I-V curves using H 2 O 2 as hole scavenger. When adding H 2 O 2 in the electrolyte, the photocurrent densities of pristine Ti-Fe 2 O 3 , OPT-100W-150s, and OPT-200W-70s tended to be almost the same, as shown in Figure 3b. It can be concluded that the carriers' generation and bulk charge transport in the three electrodes show little difference, because almost 100% of the photogenerated holes are considered to be extracted from the electrodes in the presence of the hole scavenger H 2 O 2 in the electrolyte. The practical charge transfer efficiencies (η transfer ) of these photoanodes are calculated using the ratio of J H 2 O /J H 2 O 2 , where the J H 2 O and J H 2 O 2 denote the photocurrent density measured in 1 M NaOH aqueous solution without and with 0.5 M H 2 O 2 , respectively. The curves are depicted in Figure 3c [22][23][24]. It can be seen that OPT-200W-70s and OPT-100W-150s showed higher η transfer of 78% and 64% than the 47% of pristine Ti-Fe 2 O 3 in the measured potential range, which were 1.66 and 1.36 times that of pristine Ti-Fe 2 O 3 . These were very close to the 1.95 and 1.5 times enhancements of photocurrents, correspondingly. The result reveals that O 2 plasma treatment significantly increased the surface charge transfer efficiency of the photoanodes.
Electrochemical impedance spectra (EIS) of pristine and OPT-Ti-Fe 2 O 3 photoanodes were also measured to further study the charge transport and transfer dynamics at a bias of 1.23 V vs. RHE under illumination conditions. The impedance spectra of the photoanodes are illustrated in Figure 4a, in which the impedance of Ti-Fe 2 O 3 treated with plasma is smaller than the pristine one. The impedance of OPT-200W-70s was the smallest. By fitting the EIS data according to the equivalent circuit shown in the inset of Figure 4a, the data of the series resistance (R s ), charge transfer resistance (R ct ), the bulk capacitance of hematite (C bulk ), and the capacitance of the space charge layer at the interface of Fe 2 O 3 /electrolyte (C sc ) for the electrodes were obtained. The relevant data are listed in Table 1   As the carrier density is an important factor for PEC performance, the Mott-Schottk measurement was taken to understand the effect of O2 plasma treatment on the carri densities of Ti-Fe2O3 photoanodes. Mott-Schottky curves of the electrodes were all mea ured at 1 kHz in a dark condition as illustrated in Figure 4b. The effective carrier densiti of the electrodes could be calculated according to the equation [25,26]:

Circuit Element R s (Ω) R ct (Ω) C bulk (F) C sc (F)
Pristine As the carrier density is an important factor for PEC performance, the Mott-Schottky measurement was taken to understand the effect of O 2 plasma treatment on the carrier densities of Ti-Fe 2 O 3 photoanodes. Mott-Schottky curves of the electrodes were all measured at 1 kHz in a dark condition as illustrated in Figure 4b. The effective carrier densities of the electrodes could be calculated according to the equation [25,26]: where ε, ε 0 , e, N d , V, V FB in the equation represented the permittivity of Fe 2 O 3 (ε = 80) [27,28], the permittivity of vacuum (ε 0 = 8.8542 × 10 −12 F/m), fundamental charge constant (e = 1.602 × 10 −19 C), carrier density, applied potential, and flat band potential, respectively [25,29]. The flat band of pristine

Mechanism of O 2 Plasma Treatment on Promoting PEC Performance
To explore the impact of plasma on the improved PEC performance, we carried out XPS analysis on the films, because the plasma was a surface treatment strategy that could strongly affect the chemical environment near the surface of the electrode. The Fe 2p spectra of the electrodes are summarized in Figure 5a, in which the peaks appear around 724.78 eV and 711.3 eV, indicating a typical spectrum for Fe 2p 1/2 and Fe 2p 3/2 of Ti-doped Fe 2 O 3 , respectively [30,31]. For Fe 2p 3/2 peaks, the peak at 711.3 eV and the shoulder at 710.3 eV were attributed to Fe(III)-O and Fe(II)-O, which correspond to Fe 3+ and Fe 2+ chemical states, respectively. To further clarify the effect of O 2 plasma treatment, XPS difference spectra were calculated as shown in Figure S3. The negative peaks at about 710 eV in the difference spectra suggest the decreased intensity of Fe(II)-O peak after O 2 plasma treatment. For OPT-200W-70s, the peak of Fe 2p 1/2 and Fe 2p 3/2 significantly shifted to the higher binding energy of 725 eV and 711.5 eV, respectively, as listed in Table 2. The Ti 2p XPS spectra of pristine Ti-Fe 2 O 3 (Figure 5b) presented the peaks at the binding energy of 458.39 eV and 464.23 eV that were associated with Ti 2p 3/2 and Ti 2p 1/2 , respectively [32]. After O 2 plasma treatment, the peaks both gradually shifted to higher binding energy with the increase of the power, and the values are listed in Table 2  To explore the impact of plasma on the improved PEC performance, we carried out XPS analysis on the films, because the plasma was a surface treatment strategy that could strongly affect the chemical environment near the surface of the electrode. The Fe 2p spectra of the electrodes are summarized in Figure 5a, in which the peaks appear around 724.78 eV and 711.3 eV, indicating a typical spectrum for Fe 2p1/2 and Fe 2p3/2 of Ti-doped Fe2O3, respectively [30,31]. For Fe 2p3/2 peaks, the peak at 711.3 eV and the shoulder at 710.3 eV were attributed to Fe(III)-O and Fe(II)-O, which correspond to Fe 3+ and Fe 2+ chemical states, respectively. To further clarify the effect of O2 plasma treatment, XPS difference spectra were calculated as shown in Figure S3. The negative peaks at about 710 eV in the difference spectra suggest the decreased intensity of Fe(II)-O peak after O2 plasma treatment. For OPT-200W-70s, the peak of Fe 2p1/2 and Fe 2p3/2 significantly shifted to the higher binding energy of 725 eV and 711.5 eV, respectively, as listed in Table 2. The Ti 2p XPS spectra of pristine Ti-Fe2O3 (Figure 5b) presented the peaks at the binding energy of 458.39 eV and 464.23 eV that were associated with Ti 2p3/2 and Ti 2p1/2, respectively [32]. After O2 plasma treatment, the peaks both gradually shifted to higher binding energy with the increase of the power, and the values are listed in Table 2. OPT-200W-70s exhibited a maximum shift of 0.36 eV for 2p3/2 and 0.3 eV for 2p1/2. The result suggests that the O2 plasma treatment has a great influence on the chemical state of the sample surface.    The deconvoluted peaks for O 1s are depicted in Figure 6. The observed main peak at 530.2 eV and the shoulder peak at approximate 531.8 eV were attributed to the lattice oxygen (denoted as O 2-) and the defect sites with low oxygen coordination, i.e., oxygen vacancies on sample surface (denoted as O v ) [33][34][35][36]. The main peak at 530.2 eV also shifted toward higher binding energy after O 2 plasma treatment, and the maximum shift was about 0.22 eV, as depicted in Table 1. Furthermore, the integrated area of O v /O 2of the three films was calculated to understand the change in the oxygen-related surface states under different treatment conditions. For pristine Ti-Fe 2 O 3 , the ratio of O v /O 2was calculated to be 53.4%. After treatment with O 2 plasma, the ratios of OPT-100W-150s and OPT-200W-70s were decreased to 51.6% and 49.8%, respectively. The values are illustrated in Table 3. The results strongly suggest that O 2 plasma treatment reduced the oxygen vacancies on the surface of Ti-Fe 2 O 3 films. Thus, the electron density around Ti and Fe was decreased due to the increase in lattice oxygen, which resulted in the decreased shielding ability and the binding energy moving towards high energy. Combined with the previous discussion, it is believed that the enhanced PEC performance of Ti-Fe 2 O 3 was attributed to the reduction of oxygen vacancies at the surface of electrode, rather than the variation in properties such as light absorption, morphology, and crystal structure. vacancies on sample surface (denoted as Ov) [33][34][35][36]. The main peak at 530.2 eV also shifted toward higher binding energy after O2 plasma treatment, and the maximum shift was about 0.22 eV, as depicted in Table 1. Furthermore, the integrated area of Ov/O 2-of the three films was calculated to understand the change in the oxygen-related surface states under different treatment conditions. For pristine Ti-Fe2O3, the ratio of Ov/O 2-was calculated to be 53.4%. After treatment with O2 plasma, the ratios of OPT-100W-150s and OPT-200W-70s were decreased to 51.6% and 49.8%, respectively. The values are illustrated in Table 3. The results strongly suggest that O2 plasma treatment reduced the oxygen vacancies on the surface of Ti-Fe2O3 films. Thus, the electron density around Ti and Fe was decreased due to the increase in lattice oxygen, which resulted in the decreased shielding ability and the binding energy moving towards high energy. Combined with the previous discussion, it is believed that the enhanced PEC performance of Ti-Fe2O3 was attributed to the reduction of oxygen vacancies at the surface of electrode, rather than the variation in properties such as light absorption, morphology, and crystal structure.  Finally, let us discuss the effect of O2 plasma treatment on the enhanced PEC perfor-   [37]. When the concentration of oxygen vacancy is too high, Fe 2+ sites may cause the recombination of photogenerated holes [38], which decreases surface charge transfer properties, and therefore retards the kinetics of hole reaction [15,39]. Furthermore, the oxygen vacancy is difficult to remove as an intrinsic defect. However, O 2 plasma has high reactivity to fill up the oxygen vacancies at the photoanode surface. Thus, the ratio of O v /O 2is reduced when highly reactive oxygen atoms collide with the photoanodes' surface, which in turn causes the reduction in donor density of the Ti-Fe 2 O 3 photoanode. The filling of the oxygen vacancy also reduced the amount of Fe 2+ sites, which was consistent with the decreased intensity of Fe(II)-O peaks in XPS spectra. Therefore, the surface recombination mediated by oxygen vacancies was lowered, and the charge transfer properties were enhanced, leading to a remarkably improved PEC performance, as illustrated in Scheme 1. Furthermore, the PEC performances had little change under different processing time when the plasma treatment was set as 100 W. This might be because the kinetic energy of oxygen plasma is small, and only a small thickness of the surface layer was affected by the plasma. For the power of 200 W with a greater oxygen plasma kinetic energy, the treatment's depth became deeper. More oxygen vacancies were remedied by extending the treatment time; hence, the PEC performance became better. However, when the time is further increased, the smaller carrier density will result in a smaller surface electric field. This will decrease the charge separation and deteriorate the PEC performance of the photoanodes.

Materials
Sodium sulphate (NaSO4, 99%), ferric chloride (FeCl3·6H2O, 99%), and titanium tetrachloride (TiCl4, 99.5%) were purchased from Aladdin Reagent Company (Shanghai, China). Hydrogen peroxide (H2O2, 30%), and sodium hydroxide (NaOH, 95%) were obtained from Beijing Chemical Reagent Company (Beijing, China). The fluorine-doped SnO2 (SnO2: F) transparent conducting glass (FTO, resistance: 17 Ω/eq) substrates were from Yingkou OPV Tech New Energy Co. Ltd. (Yingkou, China). Furthermore, the PEC performances had little change under different processing time when the plasma treatment was set as 100 W. This might be because the kinetic energy of oxygen plasma is small, and only a small thickness of the surface layer was affected by the plasma. For the power of 200 W with a greater oxygen plasma kinetic energy, the treatment's depth became deeper. More oxygen vacancies were remedied by extending the treatment time; hence, the PEC performance became better. However, when the time is further increased, the smaller carrier density will result in a smaller surface electric field. This will decrease the charge separation and deteriorate the PEC performance of the photoanodes. Ti-doped hematite nanowires array films were prepared on FTO conductive glass by the reported hydrothermal method [40]. FTO substrate (1.1 cm × 2.5 cm) was cleaned by the normal method [41]. A 20 mL mixed solution containing 1 M FeCl 3 ·6H 2 O and 1 M Na 2 SO 4 was prepared first, and then 0.5 mL TiCl 4 ice water with a concentration of 40 mM was added. Then FTO substrate was put into a Teflon-lined stainless steel autoclave that was filled with 2.5 mL solution. The reaction was performed at a temperature of 120 • C for 4 h, and Ti-doped FeOOH nanowires array with yellow color were obtained. After rinsing with deionized water, the films were annealed by rapid thermal annealing at 800 • C for 30 s, and Ti-FeOOH films were transformed to Ti-Fe 2 O 3 nanowires.

O 2 Plasma Treated Ti-Doped Hematite Nanowires
The oxygen gas plasma treatment was performed using a PE-ALD (Jiaxing Kemin electronic equipment Technology Company, KMT-500s, Jiaxing, China) equipped with a radio frequency coil with adjustable power. The

Conclusions
The study demonstrated that the PEC performance of Ti-doped Fe 2 O 3 nanowire photoanodes was remarkably improved after surface treatment with O 2 plasma. O 2 plasma treatment decreased the oxygen vacancies at the surface of Ti-Fe 2 O 3 nanowire films by 6.7% (from 53.4% to 49.8%). The charge transfer of Ti-Fe 2 O 3 nanowire photoanodes was improved by 1.66 times via reducing the losses caused by the oxygen vacancies' related recombination. At 1.23 V vs. RHE, the photocurrent density of O 2 OPT-Ti-Fe 2 O 3 at the power of 200 W for 70 s reached 2.14 mAcm −2 , which was 1.95 times that of the pristine Ti-Fe 2 O 3 . O 2 plasma treatment is demonstrated to be an effective strategy to improve the charge transfer ability at the surface of Ti-doped Fe 2 O 3 photoanode. We believe that further optimizations of the O 2 plasma conditions should lead to even more efficient electrodes, and it can also be applied to tune the surface states and promote the PEC performance of other semiconductor photoelectrodes.  Table S1: The structure, fabrication method, and photocurrent density of some reported works on Ti-doped a-Fe 2 O 3 nanowires photoanodes.
Author Contributions: C.L., experimental investigation, writing-original draft; D.W., writingreview and editing, visualization; J.G., investigation; Y.L., Methodology, Supervision; X.Z., validation, supervision, writing-review and editing. All authors have read and agreed to the published version of the manuscript.