Hybrid TaON/LaTiO₂N photoelectrode for water oxidation

Abstract

Efficient and stable photoelectrodes for water oxidation are highly desirable in the field of photoelectrochemical (PEC) water splitting. However, photoelectrodes with low externally applied bias usually exhibit weak photocurrent and vice versa. Herein, novel and efficient CoO_x-TaON/LTON composite photoanodes have been successfully prepared by a microwave assisted method followed with a particle transfer procedure. The obtained photoanode generated an anodic photocurrent of ~7.2 mA cm⁻² at 1.2 V_RHE and initiated the anodic photourrent at ~0.5 V_RHE. The HC-STH of the composite photoelectrode reached 1.0% at 1.2 V_RHE. Further, stoichiometric oxygen and hydrogen are stably produced on the photoanode and the counter electrode with a Faraday efficiency of unity for 2 h.

1. Introduction

Considering the exhausting of fossil resources at the global level, photoelectrochemical (PEC) and photocatalytic water splitting using semiconductors have attracted much attention owing to their great potential for efficiently generating renewable hydrogen from solar energy [1–3]. The development of a stable semiconductor that can perform efficiently under visible light is essential to practically harness solar energy [4–6]. Much effort has been devoted to find band gaps and band positions consistent with high efficiency and water splitting. To utilize visible light, the semiconductor band gap should be sufficiently narrow (i.e., <3.0 eV) for visible light absorption. In addition, the band energy positions should be suitable for water splitting; the conduction band (CB) should be sufficiently negative for H₂ production and the valence band (VB) should be sufficiently positive for water oxidation [7]. Although many photocatalysts have been developed in the past decades, efficient and stable photocatalysts responsive to visible light in aqueous solution are still few [4–6].

Oxynitride semiconductors, such as, ZnO:GaN [8,9], TaON [7, 10], and LaTiO₂N (LTON) [11, 12], have received considerable attention due to their suitable properties as photocatalysts and photoelectrodes for water splitting. Among them, TaON and LTON are the two most widely studied oxynitrides, owing to their appropriate band levels for both water oxidation and reduction, and a narrow band gap that allows visible light absorption [10–12]. Their ideal band structures originate from the shallow potential of VB maximum, which consists of N 2p orbitals hybridized with O 2p orbitals [10]. On the other hand, the bottom of the CB in TaON is predominantly composed of empty tantalum orbitals, which are sufficiently negative for water reduction [7]. Thus, TaON, as a promising photoelectrode, has the potential to produce H₂ and O₂ from water under visible light irradiation with a relatively low externally applied bias. However, its relatively large band gap, 2.5 eV [10], limits its visible light absorption and restricts its photocurrent. In contrast, LTON has a narrower band gap, 2.1 eV, and can absorb visible light up to 600 nm [11, 12]. However, the onset potential of its anodic photocurrent is more positive than that of TaON, when they prepared into photoelectrodes [7, 13]. In other words, it need relatively large applied bias in LTON photoanode for water splitting. Thus, such systems still exhibit low PEC water splitting performances, which implies that it is difficult to improve photocurrent with low externally applied bias.

Herein, we developed an efficient hybrid TaON/LTON photoelectrode for PEC water splitting. The obtained photoanode exhibits a low onset potential at ~0.5 V_RHE, closing to that of TaON, and a high photocurrent, ~7.2 mA cm⁻², at 1.2 V_RHE, which is over 2 times of that for LTON photoanode. And the photoanode showed stable oxygen evolution from water for 2 h with a faradaic efficiency of unity under simulated AM 1.5G light.

2. Experimental

2.1 Raw materials and reagents

La₂O₃ (Kanto Chemical co., Inc., 99.99%) had been freshly calcined at 1273 K for 10 h before use. Rutile TiO₂ (Rear Metallic, 99.99%) and Ta₂O₅ (Rear Metallic Co., LTD., 99.99%), NaCl (Wako Pure Chemical Industries, Ltd, 99.5%), poly(4-styrene sulfonic acid) solution (Sigma-Aldrich Japan, MW~75,000, 18 wt% in H₂O), Ethanol (Wako Pure Chemcail Industries, Ltd) Etylene glycol (Wako Pure Chemical Industries, Ltd) and Hexammine cobalt(III) chloride (Sigma-Aldrich Japan, 99%) were used without further purification.

2.2 Preparation of LaTiO₂N

LaTiO₂N particles were synthesized from La₂Ti₂O₇ according to the previously reported method [13]. La₂Ti₂O₇ precursor was prepared by solid-state reaction with NaCl flux. A mixture of La₂O₃ and TiO₂ and NaCl with molecular ratio of La: Ti: Na = 1:1:10 was annealed in air at 1423 K for 5 h, followed by cooling to 1023 K at a rate of 1 K/min. After cooling, the obtained precipitate was washed with water to give pure La₂Ti₂O₇ oxide. Then, LaTiO₂N particles were obtained under nitridation of the oxide under NH₃ flow (250 mL/min) at 1223 K for 5 h.

2.3 Preparation of TaON

TaON powder was obtained according to the previously reported method [10, 14]. Briefly, Ta₂O₅ was annealed under NH₃ flow (20 mL/min) at 1123 K for 15 h to obtain crude product. Then, red particles were removed from surface of this powder to give greenish- yellow TaON powder.

2.4 Poly(4-styrene sulfonic acid) surface treatment

Previously, it was reported by Domen group that surface treatment of LaTiO₂N using aqua regia is an effective tool to improve the photocatalytic activity [15]. But, in this case, we need to treat them quickly in order to suppress the degradation of photocatalyst due to the strong acidity of aqua regia. Recently, we found that the poly(4-styrene sulfonic acid) (PSS) solution is more effective against the surface treatment. The procedure is as below. First, LaTiO₂N powder was added to PSS solution (MW~75,000, 18% in H₂O) and stirred at room temperature for 30 min. After filtration, the powder was washed with H₂O and EtOH sufficiently to give less-defective LaTiO₂N. TaON was also treated with PSS in the same way described above.

2.5 Preparation of composites with microwave (MW)

The composites were prepared by using a microwave (Anton Paar microwave reaction system: Monowave 300 (800 W, 2.54 GHz)). Co(NH₃)₆Cl₃ was used as a CoO_x precursor because of its high solubility and reactivity. Typically, total 150 mg of LaTiO₂N and TaON (1:1, 3:1, and 1:3) was mixed in ethanol and milled for 5 minutes in a mortar. The pre-composite was added to ethylene glycol and stirred vigorously. After adding an aqueous Co(NH₃)₆Cl₃ (0.05 M) solution (the amount of Co was weighed to give a weight fraction to the composites of 2 wt%), the dispersion was heated at 150°C for 2 h with microwave apparatus. The as-synthesized composite was filtered and washed with H₂O and EtOH, followed by drying at 60°C in vacuum for several hours. The same method was applied to prepare CoO_x-TaON and CoO_x-LTON, excepting the MW treating time was shorten to 1 h.

2.6 Electrode preparation process

The electrodes were prepared by a particle transfer method [13]. Typically, the composites were casted onto a glass substrate using a suspension of composite powders in 2-propanol several times, followed by drying. The deposition of the contact layer (Ta) and conducting layer (Ti) were performed by radio-frequency (RF) magnetron sputtering. The layers were bounded to another glass with conductive carbon tape and then lifted off the primary glass. Finally, the electrode was ultrasonicated in distilled water to remove the extra particles stacked on the particles layer. The same process was applied to prepare CoO_x-TaON and CoO_x-LTON electrodes.

2.7 Characterizations of powder and electrode

The powder-samples and electrode were characterized by UV-visible diffuse reflectance spectroscopy (DRS; V-670, JASCO), X-ray powder diffraction (XRD; RINT-Ultima III, Rigaku; Cu Kα), scanning electron microscopy (SEM; SU-8020, Hitachi), high-resolution transmission electron microscopy (HRTEM; JEM-2010F, JEOL), X-ray photoelectron spectroscopy (XPS, XPS analyzer Omicron-EA 125, X-ray sources XR3E2, VG Microtech) using Mg Kα X-ray as the excitation source. The binding energies determined by XPS were corrected by reference to the Fermi-edge for each separate sample. The quantitative chemical composition of the photoanode surface was also measured using an energy dispersive X-ray spectroscopy (EDX; EMAX-7000, Horiba, microscope working at 10 kV) attached to the cold field SEM.

2.8 Photoelectrochemical measurement

The photoelectrochemical performance was measured in the typical three-electrode configuration cell, consisted of a CoO_x-TaON/LTON electrode, a counter electrode (Pt coil), and an Ag/AgCl/sat. KCl as the reference electrode. An electrolyte (aqueous 0.1 M NaOH, pH = 13). The potential of the working electrode was controlled using a potentiostat (Hokuto Denko, HSV-100) swept at 10 mV s⁻¹. The solution was purged with Ar for over 30 min prior to the measurement with vigorous stirring. The electrode current-potential was measured under intermittent illumination from a solar simulator (SAN-EI electronic, XES40S1, AM 1.5G, 100 mW cm⁻²). All measured potential versus Ag/AgCl were converted to the reversible hydrogen electrode potential (V_RHE) scale as calculated using the Nernst equation V_RHE = V_Ag/AgCl + 0.199 + 0.0591 × pH (V°_Ag/AgCl = 0.199 V at 25°C).

2.9 IPCE measurement

2.10 Photoelectrochemical water splitting

An air-tight three-electrode system PEC cell with an Ag/AgCl reference electrode, a Pt coil counter electrode and a micro gas chromatography (Micro GC) was used for a gas evolution measurement of electrode. Counter electrode of Pt-coil was coated with a Cr₂O₃ to prevent the back reaction. The PEC cell connected to a vacuum pump and a gas chromatograph with the circulation of cooling water (20°C). Before the measurement, the PEC cell was pumped to low vacuum and then purged with Ar flow with sufficient times until no nitrogen and oxygen gases can be detected in GC. A solar simulator (AM 1.5G, 100 mW cm⁻²) was used for a light sources and the amount of oxygen and hydrogen evolved from the photoelectrode and the Pt counter electrode were measured with the gas chromatograph (Inficon, GC-3000). Simultaneously current was also measured by a potentiostat (Versa STA3, Princeton Applied Research). The optimal photoelectrode was biased at 1.0 V versus RHE in a stirred aqueous solution of 0.1 M NaOH (pH = 13) under AM 1.5G simulated sunlight.

2.11 Theoretical simulation

We analyze band structure of La₂TiON and TaON for understanding the possible mechanism for charge transfer in the composites. The La₂TiON and TaON are assumed to be electronically contacted, and are formed hetero-junction. We regard the electrolyte as a hypothetical metal with a work function of 4.4(eV) [17], and employ 4.33(eV) for metal work function of Ti back contact [18]. The band structure are computed by using a semiconductor device simulator (AFORS-HET [19]) with parameters listed in Table S1. In the calculation, we adopt typical values of semiconductors for unknown parameters: dielectric constant, conduction and valence band density, and donor concentration. The electron and hole mobility of these materials are assumed to be identical, and the values are represented by La₂TiON mobility [20].

3. Results and discussion

The TaON and LTON were prepared by heating the oxide precursors under an NH₃ flow, respectively. After the preparation of TaON and LTON, the obtained products were treated in poly(4-styrene sulfonic acid) solution for 30 minutes for improving their surface conditions [21]. Then, the hybrid TaON/LTON composites were procedured through a micro-wave assisted method. Simultaneously, CoO_x nanoclusters, as efficient oxygen evolution cocatalysts [12], were deposited on the surface of the composites in-situ during the microwave process. Then, photoelectrodes were fabricated from the hybrid TaON/LTON powders using a particle transfer method developed by our group previously [13].

Figure 1(a) shows the X-ray diffraction (XRD) patterns of the prepared CoO_x-TaON, CoO_x-LTON and CoO_x-TaON/LTON composites, respectively. It can be seen that the prepared CoO_x-TaON and CoO_x-LTON have pure phases each other. No impurities can be detected in TaON and LTON samples because the amount of CoOx is relative small, 2 wt%. The XRD pattern of composites clear shows that it is composed of TaON and LTON. All the peaks can be indexed into TaON or LTON. No impurities can be found in the composites. The prepared CoO_x-TaON and CoO_x-LTON are yellowish-green and red powders, respectively. While, the composites transferred into a brownish-red powder (Figure S1). The UV-Vis light absorption spectra present that the light absorption of the composites are located between those of TaON and LTON but more close to LTON (Figure 1(b)), indicating its efficient visible light absorption. Comparing with that of LTON, the composites weakened the visible light absorption above 600 nm, which caused by the reduced Ti³⁺ species in LTON [21, 22]. Scanning electron microscopy (SEM) images and the corresponding energy dispersive X-ray spectroscopy (EDX) mapping images (Figure 1(c), 1(d) and Figure S2) show the composite photoelectrodes were well assembled by small TaON and large LTON particles. The large LTON particles were surrounded by small TaON particles. The CoO_x nanoclusters were highly dispersed on the surfaces of the composites. The cross-sectional observation (Figure S3 and S4) revealed that the average thickness of the composite particle layer was ~2 μm, which is comparable to the secondary particle size of the LTON and TaON [13]. The well mixture of TaON and LTON and the good dispersion of CoO_x in the composites were further confirmed by transition electron microscopy (TEM) images and the corresponding EDX images (Figure S5). The good attachment of CoO_x nanoclusters to the composites surfaces are also observed from the high resolution TEM(HRTEM) images (Figure S6).

Notably, the cross-link of TaON and LTON particles through the CoO_x nanoclusters could also be observed in in the composites (Figure 2). The surface X-ray photoelectron spectroscopy (XPS) spectra were recorded to determine the surface composition and chemical states of the surface elements (Figure S7 and S8). It can be seen that no obvious differences could be seen in the chemical states of element La, Ti, and Ta, demonstrating that neither the mixing of TaON and LTON nor the surface grafting of CoO_x nanoclsuters affected the bonding structure among elements of metal, oxygen and nitrogen [10–12].

Current potential (I-E) curves of CoO_x-TaON, CoO_x-LTON and CoO_x-TaON/LTON composite photoelectrodes under simulated AM 1.5 G light irradiation are shown in Figure 3. It can be seen that the anodic photocurrent for CoO_x-TaON/LTON photoelectrode increased at ~0.5 V_RHE, which is close to that of CoO_x-TaON photoelectrode, and reached ~7.2 mA cm⁻² at 1.2 V_RHE. Comparatively, the anodic photocurrent increased at ~0.8 V_RHE for CoO_x-LTON photoelectrode, and the photocurrent for CoO_x-TaON and CoO_x-LTON photoelectrodes are only ~1.8 and ~3.3 mA cm⁻², respectively, which are much lower than that of CoO_x-TaON/LTON photoelectrode. The half-cell solar-to-hydrogen conversion efficiencies (HC-STH) reached ~1.0% at ~1.0 V_RHE (Figure S9). While, the HC-STHs for CoO_x-TaON and CoO_x-LTON are only ~0.5% at ~0.7 V_RHE and ~0.4% at ~1.0 V_RHE. The good stability and high incident photon-to-charge carrier efficiency (IPCE) of the CoO_x-TaON/LTON photoelectrode (Figure S10-S12) indicate the potential efficient photoelectrode for water splitting.

Figure 4 shows the gas evolution performance on the CoO_x-TaON/LTON photoelectrode at 1.0 V_RHE under simulated AM 1.5G light irradiation. The corresponding I-t curve is shown in Figure S12. The estimated amounts of the gases based on the passed current and a faradaic efficiency of unity are also shown, with solid curves labeled e⁻/2 and e⁻/4 for hydrogen and oxygen, respectively. The stable gas evolution was achieved for 2 h over CoO_x-TaON/LTON photoelectrode, indicating its stability for PEC water oxidation [23]. The comparative studies of the oxygen evolution performances over CoO_x-TaON and CoO_x-LTON under simulated AM 1.5G light irradiation were also recorded in Figure S13. It can be seen that the evolved oxygen over CoO_x-TaON/LTON is much higher than those of CoO_x-TaON and CoO_x-LTON, revealing the composite photoelectrode are sufficiently efficient for water splitting.

To investigate the possible mechanism for the performance enhancement in the TaON/LTON composite, time-resolved diffuse reflectance (TRDR) spectroscopy and theoretical calculation were procedure [24,25]. Figure 5 shows the comparative TRDR spectroscopy analyses over CoO_x-TaON, CoO_x-LTON and CoO_x-TaON/LTON, respectively. It can be seen that the composites obviously enhance the lifetime of excited electrons with respect to the sole TaON and LTON, indicating the composites facilitate the survive of the charge carriers [24]. Moreover, theoretical simulation of TaON/LTON composites in the real electrolyte/LTON/TaON/Ti electrode system (Figure 6 and S14) reveal that holes are facile to transfer from TaON to LTON. The SEM and TEM studies showed that the TaON and LTON were well mixed and the large LTON were surrounded with small TaON. The LTON/TaON/Ti structures are also possible existed in the composite electrode. In contrast, the TaON/LTON/Ti structures are also possible existed. This opposite structure will reduce the PEC performance of the electrode (Figure S15). Fortunately, this opposite structure is rare observed in the composite electrode (Figure 1, S2-S4). Further, owing to the virtue of the in-situ CoO_x deposition during the microwave process, TaON and LTON were partially connected with each other through CoO_x nanoclusters. These features are beneficial for the charge transfer and separation. Thus, the composites exhibited enhanced PEC performances.

We further optimized the ratio between the amount of TaON and LTON as shown in Figure S16. The results in Figure 3 demonstrate our best performance for visible light PEC water splitting. It may be due to the ratio of TaON to LTON will seriously influences the light absorption and the surround degree of TaON to LTON, which are important for efficient visible light activity.

4. Conclusions

In summary, a CoO_x-TaON/LTON composite photoanode has been successfully prepared by a microwave assisted method followed a particle transfer process. The obtained photoelectrode generated an anodic photocurrent of ~7.2 mA cm⁻² at 1.2 V_RHE, which is much larger than that of CoO_x-LTON, and increased the anodic photourrent at ~0.5 V_RHE, which is close to that of CoO_x-TaON. The HC-STH of the composite photoelectrode reached 1.0% at 1.2 V_RHE, which is the highest HC-STH for reported oxynitrides. The obtained composite photoelectrode exhibited a stable water splitting for 2 h with a unity faradaic efficiency under simulated AM 1.5 G light irradiation. The TRDR spectroscopy and theoretical simulation studies revealed that the hybrid composite facilitate the charge transfer and separation. Thus, the reasonable hybridism semiconductors may provide an efficient route to enlarge their advantages and suppress their disadvantages. Moreover, the current work also provide an efficient method to prepare efficient functional materials with the assistance of microwave process.

Acknowledgement

We thank the Natural Science Foundation of China (Grant No. 21872174 and U1932148), Project of Innovation-Driven Plan in Central South University (Grant No. 20180018050001), State Key Laboratory of Powder Metallurgy, International Science and Technology Cooperation Program (Grant No. 2017YFE0127800), Hunan Provincial Science and Technology Program (2017XK2026), Shenzhen Science and Technology Innovation Project (Grant No. JCYJ20180307151313532), Thousand Youth Talents Plan of China and Hundred Youth Talents Program of Hunan.

Conflict of interest statement. None declared.

References