High Performance of Manganese Porphyrin Sensitized p-Type CuFe 2 O 4 Photocathode for Solar Water Splitting to Produce Hydrogen in a Tandem Photoelectrochemical Cell

: A novel composite composed of (5, 10, 15, 20-tetraphenyl) porphinato manganese sensitized p-type CuFe 2 O 4 was developed for constructing the photocathode of a tandem photoelectrochemical (PEC) cell. The prepared material was characterized by X-ray diffraction (XRD), transmission electron microscopy (TEM), X-ray photoelectron spectroscopy (XPS) and UV-vis diffuse reﬂectance spectroscopy (DRS). Light-driven water splitting to produce hydrogen can be achieved through the PEC cell, and the results show that H 2 and O 2 can be collected separately at low applied bias. This work demonstrates that manganese porphyrin sensitized CuFe 2 O 4 is an effective hybrid material for building the photocathode of a PEC cell for solar water splitting to produce H 2 .


Introduction
In recent years, hydrogen, acting as an environmentally friendly energy carrier with high capacity, has attracted great attention owing to increasing environmental concerns and energy demands. Photocatalytic water splitting to produce hydrogen can be regarded as one of the most promising strategies for converting solar energy into chemical fuels [1,2]. Over the past decades, much effort has been devoted to exploring new materials for fabricating electrodes for photoelectrochemical (PEC) cells [3,4]. In general, a PEC cell is constructed by an anode and a cathode, from which oxygen and hydrogen can be produced and collected. The photocathode is usually composed of a p-type semiconductor with suitable bandgap. Under sunlight illumination, the electrons in the valence band of the p-type photocatalyst are excited to its conductor band. If the potential of the photoexcited electrons accumulated on the conduction band is high enough to be able to overcome the surface potential and migrate to the interface of the electrode and electrolyte, they can act as a reducing agent to reduce water into hydrogen. So far, a large number of p-type semiconductors, including Cu 2 O or CuO [5][6][7], NiO [8,9], InP [10,11], etc., have been reported as photocathode materials for light-driven water splitting [12][13][14][15][16]. Among them, Cu 2 O has received the most attention, with favorable CB edge position and direct band gap; however, it is unstable in aqueous solutions under illumination [17]. CuFe 2 O 4 as a spinel ferrite has been regarded as a promising catalyst for photocatalytic H 2 production owning to the relatively narrow band gap, low cost and photochemical stability [18,19]. Although aqueous solutions under illumination [17]. CuFe2O4 as a spinel ferrite has been regarded as a promising catalyst for photocatalytic H2 production owning to the relatively narrow band gap, low cost and photochemical stability [18,19]. Although the photocurrent of CuFe2O4 is relatively modest compared to that of Cu2O, its onset potential is at a lower level, which is beneficial for photocatalysis [20].
Porphyrin-based dyes have been known to be used as potential sensitizers, owing to their high extinction coefficients in the visible light region, excellent chemical and thermal stability, efficient photochemical electron transfer ability, and low cost [21][22][23]. Porphyrin dye-modified semiconductor electrodes can improve the photophysical and electronic properties of PEC cells greatly [24].

Morphology and Structure
XRD patterns of the samples are shown in Figure 1A. Six diffraction peaks at 30.1°, 35.4°, 43.2°, 50.4°, 62.5°, and 74.1° correspond to (220), (311), (400), (107), (440), and (533) planes of the cubic spinel structure of CuFe2O4 (JCPDS card no. 25−0283). In addition, the relatively regular diffraction peaks without the peaks of other impurities show the high purity of CuFe2O4. The diffraction peaks of MnTPP/CuFe2O4 match well to those of pure CuFe2O4, indicating that the presence of MnTPP in the hybrid does not change the crystallinity of CuFe2O4. The TEM image of the CuFe2O4 sample ( Figure 1B) shows that the as-prepared CuFe2O4 sample is composed of nanospheres with a rough surface. The average diameter of the nanospheres is 50-60 nm.  The Fe 2p XPS spectrum of CuFe 2 O 4 ( Figure 1C) shows two dominant peaks with binding energies of Fe 2p 3/2 and Fe 2p 1/2 at 711.3 and 725.1 eV along with a typical satellite peak at 718.6 eV, suggesting the presence of Fe 3+ in the sample [25]. Each Fe 2p lines can be deconvolved into two peaks, as shown in Figure S1A. The doublets of Fe 2p 3/2 binding energy at 710.4 eV and Fe 2p 1/2 binding energy at 723.7 eV are assigned to the Fe 3+ ions in octahedral sites, while the doublets of Fe 2p 3/2 binding energy at 712.8 eV and Fe 2p 1/2 binding energy at 725.7 eV are assigned to Fe 3+ ions in tetrahedral sites [26]. Moreover, the relative contributions to the overall intensity of Fe 3+ ions at the octahedral and tetrahedral sites are 40% and 60%, calculated from the ratio of peak area. The Cu 2p spectrum of CuFe 2 O 4 ( Figure 1D) shows two main peaks located at 933.2 and 952.8 eV, corresponding to the Cu 2p 3/2 and Cu 2p 1/2 of Cu 2+ , respectively [20]. Along with a typical satellite peak at 942.8 eV, the above result indicates the presence of Cu 2+ in synthesized CuFe 2 O 4 . Moreover, each Cu 2p peak can also be fitted to two peaks, as shown in Figure S1B. The doublets of Cu 2p 3/2 binding energy at 933.2 eV and Cu 2p 1/2 binding energy at 952.8 eV are assigned to the Cu 2+ ions in octahedral sites, while the doublets of Cu 2p 3/2 binding energy at 935.1 eV and Cu 2p 1/2 binding energy at 955.0 eV are assigned to the Cu 2+ ions in tetrahedral sites. The relative contributions to the overall intensity of Cu 2+ ions at the octahedral and tetrahedral sites are 51% and 49%, respectively [27]. These results are consistent with that of the CuFe 2 O 4 nanospheres prepared by the reported hydrothermal method [28]. The Cu 2p peaks' intensity and the width of the MnTPP/CuFe 2 O 4 showed a decrease compared to CuFe 2 O 4 , which may attributed to a decrease of the content and the number of inequivalent sites of Cu on the composite's surface when increasing the MnTPP ratio in the composite [29,30]. In addition, a negative shift or a smaller peak binding energy is observed for both Fe 2p and Cu 2p peaks in the XPS spectra of the MnTPP/CuFe 2 O 4 sample, this result shows the formation of new chemical bonds after combining MnTPP with CuFe 2 O 4 , demonstrating that there is an interaction between MnTPP and CuFe 2 O 4 in the nanohybrid [31]. The XPS spectra of Mn 2p in the MnTPP/CuFe 2 O 4 material ( Figure S1C) exhibits two peaks centered on 642.4 and 654.5 eV, corresponding to Mn 2p 3/2 and Mn 2p 1/2 , respectively. The above results provide an evidence that the MnTPP-modified CuFe 2 O 4 hybrid had been successfully synthesized. Figure 2 displays the UV-vis diffuse reflectance spectra of the prepared samples. From the figure, we can see that CuFe 2 O 4 demonstrates a wide light absorption throughout the whole UV-vis region, which is in agreement with previous reports [32,33], and the MnTPP dye exhibits a strong Soret band at 455 nm and two weak Q bands at 570 and 610 nm, respectively [34]. The spectrum of the MnTPP/CuFe 2 O 4 composite exhibits broad absorption ability in visible light region, with the characteristic absorption of MnTPP bands centered at 464 and 614 nm. These facts demonstrate that the MnTPP/CuFe 2 O 4 composite has an enhanced ability to harvest visible light, as compared to bare CuFe 2 O 4 . The Tauc plot (inset of Figure 2) shows that the band gap is 1.42 eV for CuFe 2 O 4 in the hybrid structure, which is in agreement with the value given in previous reports [35,36]. The linear scan voltammograms of CuFe2O4 are shown in Figure 3A. It is obvious that the cathodic current of CuFe2O4 under illumination is larger than the dark current, which is characteristic of a p-type semiconductor [37,38]. The fact that dark current of the fabricated photocathode is a bit high may be due to the existing surface states on the photocathode materials, resulting in the leakage current [39,40]. The negative slope of the Mott-Schottky curve of CuFe2O4 ( Figure 3B) further proves that p-type CuFe2O4 semiconductor has been successfully fabricated, and the calculated flat-band potential of CuFe2O4 is 0.14 V vs. RHE, which can be approximately considered to be the valence band value for p-type semiconductors. Combined with the value of the band gap (1.42 eV), we can conclude that the conduction band edge of CuFe2O4 is −1.28 V vs. RHE. The photoelectrochemical results of the prepared samples are shown in Figure 3C. The CuFe2O4/ITO electrode shows a weak photocurrent density (−0.04 μA cm −2 ) under UV-vis light irradiation; whereas the observed photocurrent density of the MnTPP/ITO electrode is −0.11 μA cm −2 in the same conditions, which can be caused by the strong optical absorption of MnTPP and effective photoexcited electron transfer from the sensitizer to FTO. Compared with the photocurrent of CuFe2O4 and MnTPP, the MnTPP/CuFe2O4/FTO electrode demonstrates a remarkably enhanced photocurrent density (−0.34 μA cm −2 ), owing to broad and strong absorption of the MnTPP moiety and efficient electron transfer from the excited dye to CuFe2O4, resulting in effective photoexcited charge separation [41]. The electrochemical impedance spectra (EIS) of CuFe2O4 electrode and MnTPP/CuFe2O4 electrode are shown in Figure 3D. The radius of the MnTPP/CuFe2O4 plot is smaller than that of CuFe2O4, indicating a smaller resistance and an improvement of charge transfer on the interface between MnTPP/CuFe2O4 and the electrolyte [22,42]. The linear scan voltammograms of CuFe 2 O 4 are shown in Figure 3A. It is obvious that the cathodic current of CuFe 2 O 4 under illumination is larger than the dark current, which is characteristic of a p-type semiconductor [37,38]. The fact that dark current of the fabricated photocathode is a bit high may be due to the existing surface states on the photocathode materials, resulting in the leakage current [39,40]. The negative slope of the Mott-Schottky curve of CuFe 2 O 4 ( Figure 3B) further proves that p-type CuFe 2 O 4 semiconductor has been successfully fabricated, and the calculated flat-band potential of CuFe 2 O 4 is 0.14 V vs. RHE, which can be approximately considered to be the valence band value for p-type semiconductors.

Photocatalytic Activity
For all photocatalytic experiments carried out in the tandem PEC cell, the molar amount of O2 detected from the counter electrode is about 1/2 of that of H2. Figure S2 shows the amount of hydrogen collected from photocathode under −0.2 V applied bias vs. RHE. In 2 h irradiation of a solar simulator (Ultra-vitalux, 300 W, Osram, Augsburg, Germany), the amount of hydrogen collected from the CuFe2O4 chamber is 11.1 μmol g −1 . Under the same conditions, the amount of hydrogen collected from the MnTPP/CuFe2O4 chamber reaches 23.0 μmol g −1 , indicating a positive photocatalytic influence of the combination of MnTPP and CuFe2O4. Platinizing the electrode can further strengthen the activity of the dye-sensitized electrode. The amount of hydrogen evolved from the Pt/MnTPP/CuFe2O4 electrode after 2 h of reaction increases to 760.0 μmol g −1 under −0.2 V vs. RHE bias. For comparison, we used a Pt plate to replace the prepared photocathode. Neither hydrogen nor oxygen can be detected from the tandem PEC cell under the same reaction conditions. Figure 4 and the inset of Figure 4 show that hydrogen can also be collected from CuFe2O4 (9.3 μmol g −1 ), MnTPP/CuFe2O4 (12.0 μmol g −1 ), and Pt/MnTPP/CuFe2O4 (748.0 μmol g −1 ) electrode under −0.1 V applied bias vs. RHE, respectively, indicating that as-prepared photocathodes still have catalytic activity at very low applied bias. The Faraday efficiency for H2 production and solar-to-hydrogen conversion efficiency ηSTH from Pt/MnTPP/CuFe2O4 PEC cell at −0.1 V vs. RHE are 53% and 0.14%, respectively.

Photocatalytic Activity
For all photocatalytic experiments carried out in the tandem PEC cell, the molar amount of O 2 detected from the counter electrode is about 1/2 of that of H 2 . Figure S2 Figure 4 and the inset of Figure 4 show that hydrogen can also be collected from CuFe 2 O 4 (9.3 µmol g −1 ), MnTPP/CuFe 2 O 4 (12.0 µmol g −1 ), and Pt/MnTPP/CuFe 2 O 4 (748.0 µmol g −1 ) electrode under −0.1 V applied bias vs. RHE, respectively, indicating that as-prepared photocathodes still have catalytic activity at very low applied bias. The Faraday efficiency for H 2 production and solar-to-hydrogen conversion efficiency η STH from Pt/MnTPP/CuFe 2 O 4 PEC cell at −0.1 V vs. RHE are 53% and 0.14%, respectively.  The incident photon-to-electron conversion efficiency (IPCE) spectrum of MnTPP/CuFe2O4 electrode ( Figure S3) shows a maximum IPCE of 22.8% at 475 nm and 22.7% at 585 nm. This result fits well the UV-visible diffuse reflectance spectra of MnTPP/CuFe2O4. Figure S4A-C shows the results of the photocurrent experiments of the as-prepared photocathodes. The steady-state photocurrent changes in the order: Pt/MnTPP/CuFe2O4 > MnTPP/CuFe2O4 > CuFe2O4, which is consistent with the amount of hydrogen evolved from the PEC cell. The photocurrent of the Pt/MnTPP/CuFe2O4 electrode remains constant from the very beginning to 6 h of continuous operation at the −0.1 V and −0.2 V bias, indicating sufficient stability for the photocatalysis.

Photoelectrode Preparation
CuFe2O4 nanospheres were prepared by a hydrothermal method [28]. In a typical synthesis, 2 mmol of CuCl2•2H2O, 4 mmol of FeCl3•6H2O, 6 mmol of NaAc and 0.5 g of polyvinylpyrrolidone (PVP) were dissolved into 30 mL of ethylene glycol, and stirred at room temperature for 30 min. Then the mixture was transferred into an 80 mL Teflon-lined autoclave and maintained at 160 °C for 24 h. The solid was achieved from the mixture through centrifugation, and washed with deionized water three times. Finally, the solid was dried at 60 °C under vacuum. For fabricating the CuFe2O4 photoelectrode, a CuFe2O4 suspension was prepared by adding and grinding the CuFe2O4 powder (20 mg) with a solution composed by 3 mL of trichloromethane, ethylene glycol, ethanol, and PVP (10 mL: 15 mL: 75 mL: 30 mg). CuFe2O4 was deposited by doctor blading the suspension on a clean

Photoelectrode Preparation
CuFe 2 O 4 nanospheres were prepared by a hydrothermal method [28]. In a typical synthesis, 2 mmol of CuCl 2 ·2H 2 O, 4 mmol of FeCl 3 ·6H 2 O, 6 mmol of NaAc and 0.5 g of polyvinylpyrrolidone (PVP) were dissolved into 30 mL of ethylene glycol, and stirred at room temperature for 30 min. Then the mixture was transferred into an 80 mL Teflon-lined autoclave and maintained at 160 • C for 24 h. The solid was achieved from the mixture through centrifugation, and washed with deionized water three times. Finally, the solid was dried at 60 • C under vacuum. For fabricating the CuFe 2 O 4 photoelectrode, a CuFe 2 O 4 suspension was prepared by adding and grinding the CuFe 2 O 4 powder (20

Characterization
Transmission electron microscopy (TEM) studies were carried out on a FEI TECNAI-G2 electron F20 200 KV microscope (JEOL JEM-2100, JEOL Ltd., Tokyo, Japan). The X-ray diffraction (XRD) patterns were measured on X'Pert-Pro MRD (Amsterdam, The Netherlands) equipped with Ni-filtered Cu Kα radiation. X-ray photoelectron spectroscopy (XPS) measurement was carried out on ESCALAB 250 XI with an AXIS Ultra DLD system (Kratos Analytical Ltd., Manchester, UK). UV-vis diffuse reflectance absorption spectra (DRS) were carried out on a UV3600 spectrophotometer (Shimadzu, Kyoto, Japan). Photoelectrochemical measurements were performed on a CHI 660D potentiostat/galvanostat electrochemical analyzer (CHI Instrucment, Inc., Shanghai, China) in a three-electrode system consisting of a working electrode, a saturated calomel electrode (SCE) and a platinum plate. The working electrode was prepared by dipping ca. 0.5 mL of the sample dispersion (0.4 mg mL −1 ) on a clean fluorine-doped tin oxide (FTO) glass and drying under ambient conditions. The efficient area of the electrode was about 1.0 cm 2 . Three electrodes were immersed in 0.2 M Na 2 SO 4 (pH = 6), which acted as the electrolyte solution. The working electrode was irradiated by a GY-10 xenon lamp (150 W) as light source.

Photocatalytic Reaction
The photocatalytic reaction was performed in a three-electrode PEC device, as depicted in Scheme S1. The CuFe 2 O 4 or MnTPP/CuFe 2 O 4 photoelectrode acted as photocathode, a platinum plate as counter electrode and an SCE as the reference electrode. Three electrodes are connected through a potentiostat, and the system was deaerated by pumping argon into the system for 30 min before the reaction. The working electrode was illuminated by a solar simulator (Ultra-vitalux, 300 W) at 298 K. The distance between the lamp and the quartz window of the PEC cell was 20 cm. The gases produced from the reaction were analyzed by an online gas chromatograph (GC1690, Ke Xiao Instruments Co., Ltd., Hangzhou, China), which has a thermal conductivity detector and 5 Å molecular sieve columns; argon was used as carrier gas. GC signal calibration was performed by the standard H 2 /Ar gas mixtures with known concentrations.
The apparent quantum yield (AQY) was studied by the solar simulator with a 420 nm filter acting as light source. The focused light intensity was measured on a radiometer. AQY is calculated by the Equation (1): where n H 2 is the measured molar number of H 2 at the reaction t, and I 0 is the number of photons per unit time. I 0 there was ca. 4.5 × 10 −8 mol s −1 . Solar-to-hydrogen conversion efficiency η STH from the PEC cell at −0.1 V vs. RHE is calculated by the following equation [43][44][45]: where J op is the effective operating current density measured during the reaction (35.7 µA cm −2 ), V is the water splitting potential required (1.23 V), V bias is the bias voltage between two electrodes, P light is the incident light power (16 mW cm −2 ), and η F is the Faraday efficiency (55%).

Conclusions
In summary, we demonstrated a novel material composed of porphinato manganese sensitized p-type CuFe 2 O 4 , which has been successfully applied to fabricate a photocathode for water splitting to produce hydrogen. The gases produced from the PEC cell can be collected separately at very low bias. This study provides a new approach to obtaining both practical and reliable materials for photocatalytic applications. Scheme S1: Scheme of three-electrode PEC device. Table S1: comparison of the performance of photocathodes.