Abstract
An ultrathin FeOOH cocatalyst is deposited on α-Fe2O3 photoanodes in a simple room temperature immersion process for efficient photoelectrochemical (PEC) water splitting. The prepared FeOOH/Fe2O3 photoanode has a photocurrent density of up to 2.4 mA/cm2 at 1.23 V versus reversible hydrogen electrode (RHE), and the photocurrent density is increased by about 160% compared to the bare Fe2O3 of 1.55 mA/cm2. An obvious cathodic shift of the photocurrent onset potential from 0.661 to 0.582 V was also observed, and excellent stability was maintained with almost no deterioration for 5 h. The enhanced PEC performance is attributed to the decrease of the interfacial resistance between electrode and electrolyte and the increase of the injection efficiency of holes in Fe2O3.
1. Introduction
Solar-driven photoelectrochemical (PEC) water splitting has been a promising method to translate solar energy to chemical fuels, which has been widely concerned and studied [1]. In a PEC water splitting cell, the oxidation evolution reaction (OER) involves multiple electrons, resulting in a high overpotential requirement. The OER is thus the rate-determining step for PEC water splitting. Therefore, photoanode materials with broad range of light absorption, efficient charge transfer, low overpotential, and durability are necessary to construct a practical PEC device [2].
So far, lots of promising materials have been explored as photoanodes in a PEC device, such as Fe2O3, WO3, BiVO4, Ta3N5, and TiO2 [3–8]. As a promising candidate material for PEC water splitting, α-Fe2O3 has advantages such as excellent stability, economical, nontoxic, appropriate band gap (~2.1 eV), high theoretical photocurrent value of 12.5 mA/cm2, and nearly 15.8% solar-to-hydrogen efficiency (STH) [9]. However, due to the insufficient properties of α-Fe2O3 photoanode such as high overpotential, short hole diffusion length, and relatively low visible light absorption coefficient, the experimental solar energy conversion efficiency is still far from reaching the theoretical value [10]. Therefore, some effective strategies for improving the PEC performance of the α-Fe2O3 photoanode were explored, such as element doping (such as Mo6+, Nb5+, Ti4+, Al3+, Mg2+, and Zn2+), nanostructure regulation (such as nanobelts, nanorods, and nanowires), heterojunction building, surface passivation (such as Al2O3 and Ga2O3), incorporation of conducting scaffolds (such as ITO and graphitic carbon nitride), and deposition of oxygen evolution cocatalysts (OECs) [2, 11–14]. There is slow oxidation kinetics on the surface of α-Fe2O3, so the overpotential of the α-Fe2O3 electrode is extremely high. OEC deposition can effectively accelerate the oxidation rate of water, thereby reducing the onset potential of α-Fe2O3. Co-Pi and noble metal oxides (such as IrO2 and RuO2) are representative traditional materials applied as OECs [15]. The loading VIII metal (Fe, Co, and Ni) oxide or (oxy) hydroxide widely used as cocatalysts on photoanodes for water oxidation have proven to be an effective method [16, 17]. Among them, Fe is less toxic than Co and Ni and the most sufficient transition metal in the earth’s crust [18, 19]. Meanwhile, oxyhydroxides are one of the most efficient cocatalysts in recent years for some semiconductor thin film photoanodes [20].
In this work, an FeOOH cocatalyst deposit onto Fe2O3 photoanodes (abbreviated as FeOOH/Fe2O3) was realized via a simple method of room temperature immersion method. Urea was added to the precursor that prepared FeOOH/Fe2O3. The urea slowly hydrolysed and formed NH3 to enable the formation of FeOOH in the presence of Fe3+. In this way, an FeOOH nanolayer can be rapidly grown on the surface of an Fe2O3 nanorod. The formed FeOOH cocatalyst effectively decreases the interfacial resistance between the electrode and the electrolyte of Fe2O3 photoanodes and increases the injection efficiency of holes. This allows the holes to travel faster from the Fe2O3 photoanode surface to the electrolyte, thus the efficient water oxidation reaction occurs at the interface of electrode and electrolyte. Compared with the bare Fe2O3 photoanodes, the photocurrent density of the FeOOH/Fe2O3 photoanodes is revealed to be considerably improved, with an increase from 1.55 to 2.23 mA/cm2 at 1.23 V vs. RHE. Simultaneously, a cathodic shift of onset potential is about 80 mV from 0.66 to 0.58 V and excellent photostability in 5 h.
2. Experimental
2.1. Sample Preparation
The Fe2O3 nanorod array grows on substrate of F-doped tin oxide (FTO) transparent conductive glass, which is prepared by a simple hydrothermal method. The prepared FTO glass is ultrasonically cleaned with ultrapure water, acetone, and ethanol, respectively, for no less than 30 min. A stainless steel autoclave contains the FTO glass and 25 ml ultrapure water including 0.15 M FeCl3·6H2O (Shanghai Hushi, 99.99%), 90 μl TiCl3 (3AChem, 15.0–20.0% TiCl3 basis in 30% HCl), and 0.3 M CO(NH2)2 (Shanghai Hushi, 99%). Four pieces of cleaned FTO were placed against the wall of the autoclave. The reaction is 6 h for 100°C, then cooled to room temperature naturally. After being annealed at 550°C and 650°C in air, α-Fe2O3 photoanodes were prepared.
FeOOH/Fe2O3 photoanodes were formed by the room temperature impregnation method. The Fe2O3 photoanode was immersed in 25 ml solution of 0.075 M FeCl3·6H2O (Shanghai Hushi, 99.99%) and 0.15 M CO(NH2)2 (Shanghai Hushi, 99%) for varying durations (1 h, 2 h, and 3 h) at room temperature (25°C). The sample is rinsed with ultrapure water and then blown dry.
2.2. Characterization of Materials
The morphologies of the samples were observed by a scanning electron microscope (SEM) Nova Nano SEM 230 FEI Co, and a JEM-200CX TEM transmission electron microscope is used for observing transmission electron microscopy (TEM). The UV-vis spectra were observed with a UV-2550 (Shimadzu) UV-vis spectrophotometer. The crystal structures of the samples were examined by Ultima III Rigaku (Cu Kα radiation) X-ray diffraction (XRD). The binding energy was measured by a photoelectron spectrometer (XPS) of Thermo ESCALAB 250 X-ray and was calibrated by C1s (284.5 eV).
2.3. Photoelectrochemical Performance
The PEC test was performed under a three-electrode system using a CHI-760E electrochemical workstation manufactured by Shanghai Chenhua. The voltage scanning speed should not be too fast to prevent charging and discharging. The voltage scanning rate selected in this paper is 20 mV/s, and the range is 0.4 V to 1.8 V vs. RHE. A 500 W xenon lamp is used for the light source, and the light intensity used in this experiment is 52.7 mW/cm2. When the photocurrent is tested, the sample is irradiated by light from the back (α-Fe2O3/electrolyte interface) and the exposed area of the photoanode under illumination was 0.28 cm2. In the PEC test, the working electrode is Fe2O3 or FeOOH/Fe2O3 samples, the counter electrode is Pt electrode, and the reference electrode is Ag/AgCl electrode. A 1 M NaOH aqueous solution (pH 13.6) was used as electrolyte. The calculated formula is , where represents the potential versus reversible hydrogen electrode and .
3. Results and Discussion
As shown in Figure 1, FeOOH nanorods were grown on FTO by the hydrothermal method to prepare the FeOOH/Fe2O3 photoanode. And then, a red-orange Fe2O3 film is obtained by high-temperature annealing (650°C) as described previously [3]. Subsequently, the α-Fe2O3 photoanodes are immersed in precursor aqueous solutions to allow ultrathin FeOOH to grow on the Fe2O3 nanorod surface. The FeOOH cocatalyst was loaded by immersing Fe2O3/FTO into a precursor aqueous solution of FeCl3·6H2O and urea at room temperature. The urea slowly hydrolysed and form NH3 to enable the formation of FeOOH. The presence of Fe3+ apparently promotes the hydrolysis reaction at room temperature.
In the SEM images (Figures 2(a) and 2(b)), both bare Fe2O3 and FeOOH/Fe2O3 exhibited a similar morphology. The measurement diameter is about 50 nm and the length is about 400 nm. No morphological differences between the Fe2O3 and FeOOH/Fe2O3 nanorod arrays were observed, which indicates that the immersion treatment of Fe2O3 at room temperature has no effect on the Fe2O3 nanorods. The low-resolution TEM images clearly show the morphologies of bare Fe2O3 and FeOOH/Fe2O3 nanorods (Figures 2(c) and 2(d)). As shown in Figure 2(e), high-resolution transmission electronic microscopy (HR-TEM) images show that the lattice distance was 0.24 nm, corresponding to the interplanar spacing of 110 planes of hematite (α-Fe2O3). Figure 2(f) shows a uniform FeOOH coating on the surface of Fe2O3 with a thickness of 2.23 nm. The XRD patterns of FeOOH/Fe2O3 and Fe2O3 exhibit diffraction peaks which are consistent with hematite (α-Fe2O3, JCPDS 33-0664) and F-doped tin oxide (FTO, JCPDS 41-1445) (Figure 3(a)). In addition to this, it can be found that the photoanode of FeOOH/Fe2O3 is matched with FeOOH (JCPDS No. 34-1266) by two relatively weak peaks of 11.8° and 26.7° in 2θ, respectively, detected in XRD. Figure 3(b) shows that the wavelength-dependent light absorption of FeOOH/Fe2O3 is almost the same as that of the bare Fe2O3 and it is only slightly higher (~3%) in the wavelength region of 450-550 nm. The nearly identical light absorption between Fe2O3 and FeOOH/Fe2O3 is attributed to the FeOOH ultrathin thickness. XPS spectra of the FeOOH/Fe2O3 and Fe2O3 photoanodes show that the binding energies of Fe2p were basically the same in the two samples before and after treatment (Figure 3(c)), while the binding energies of O1s were changed (Figure 3(d)). Three binding energies (O1, O2, and O3) of O1s (529.7 eV, 530.4 eV, and 531.3 eV) are observed in Figure 4(b), belonging to the O2- species, OH- group, and adsorbed oxygen species, respectively. The XPS spectra of FeOOH/Fe2O3 show an OH- peak with higher intensity than that of Fe2O3, which proved the existence of FeOOH. The above results lead to the conclusion that the ultrathin FeOOH nanolayer can be uniformly loaded on the Fe2O3 photoanodes successfully via a simple method of room temperature immersion method.
(a)
(b)
(c)
(d)
(e)
(f)
(a)
(b)
(c)
(d)
(a)
(b)
(c)
(d)
Figure 4(a) shows the photocurrent comparison of Fe2O3 photoanodes after different times of solution impregnation method treatment. It can be seen that the photocurrent of the two-hour treatment FeOOH/Fe2O3 photoanode is the highest with the lowest of starting voltage. Figure 4(b) shows the photocurrent density-potential curves of bare Fe2O3 and FeOOH/Fe2O3 photoanodes. It can be seen that compared with the bare Fe2O3, the FeOOH/Fe2O3 photoanode has a relatively enhanced photocurrent density, with an increase from 1.55 to 2.40 mA/cm2 at 1.23 V vs. RHE. Simultaneously, the cathodic shift of onset potential with about 80 mV from 0.66 to 0.58 V is also observed. The results show that the deposition of an FeOOH ultrathin overlayer can improve the photocurrent density availably, and the onset potential of Fe2O3 photoanodes can be effectively decreased. The transient photocurrent response curves of photoelectrodes were acquired at 1.23 V vs. RHE under chopped light illumination (Figure 4(c)). The photocurrent intensity of FeOOH/Fe2O3 exhibits obviously higher than that of bare Fe2O3 when irradiated by the light source. The photocurrent densities obtained are basically consistent with the values from the density-potential curves at 1.23 V vs. RHE. Under on/off cyclic light illumination, the electrodes have fast light response and excellent photosensitivity. At the moment the light is turned on, the slight photocurrent transient spikes were observed, which were caused by a slight recombination of holes and electrons. The photocurrent density in a hole scavenger (0.1 M Na2SO3) containing electrolyte solution was conducted to clarify the charge injection efficiency of bare Fe2O3 and FeOOH/Fe2O3 photoanodes. The charge injection efficiency of FeOOH/Fe2O3 reached 82% at 1.23 V vs. RHE, which was obviously better than the 51% of Fe2O3 (Figure 4(d)). The increased charge injection efficiency contributes to the enhanced photocurrent density of FeOOH/Fe2O3 photoanodes.
In order to explain the different charge transfer of the bare Fe2O3 and FeOOH/Fe2O3 photoanodes, the EIS tests are examined under the xenon lamp illumination at 1.23 vs. RHE (Figure 5(a)). Equivalent resistance and capacitance circuit model is fit to these data (Figures 5(c) and 5(d)). Here, Rs represents series resistance, which includes the resistance of interface between FTO and Fe2O3 films and contact resistance of the test system [3]. The charge transfer resistance (Rct) of the FeOOH/Fe2O3 photoanode was smaller than that of the Fe2O3 photoanode, which proves the faster charge transfer from the FeOOH/Fe2O3 photoanode to the electrode/electrolyte interface [21]. The reduction of resistance indicated that the FeOOH nanolayer allows that more holes are efficiently captured and promotes the hole migration on the interface of electrode and electrolyte, thus achieving an effective water oxidation reaction. As far as we know, photostability is a considerable characteristic of photoanodes, thus the photocurrent density versus time of the FeOOH/Fe2O3 photoanode was further tested. As shown in Figure 5(b), it reveals no significant attenuation under xenon lamps for 5 h in 1 M aqueous NaOH and it exhibits excellent photostability. It can also be seen that the enhancement of FeOOH/Fe2O3 photoanode photocurrent density is not caused by the photo corrosion of the photoanode and the enhanced performance of PEC is not at the expense of the stability of light.
(a)
(b)
(c)
(d)
4. Conclusions
A simple room temperature immersion method is used to form an FeOOH nanolayer on an Fe2O3 photoanode. The FeOOH nanolayer, as an efficient OEC, can increase the hole injection efficiency and effectively decrease the interfacial resistance between electrode and electrolyte, so the photocurrent density greatly improved, appreciably lowering the onset potential for the Fe2O3 photoanode. Excellent photostability is also observed. The desirable enhanced photocurrent density is 2.40 mA/cm2 at 1.23 V vs. RHE, and the evidently lower onset potential is 0.58 V.
Data Availability
The data used to support the findings of this study are available from the corresponding authors upon request.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
The study was supported by the NSF of China (No. 21773114), the NSF of Jiangsu Province (No. BK20171246), and the Fundamental Research Funds for the Central University (020414380135).