Creation and Plasmon-Assisted Photosensitization of Annealed Z-Schemes for Sunlight-Only Water Splitting

Solely light-induced water splitting represents a promising avenue for a carbon-free energy future, based on reliable energy sources. Such processes can be performed using coupled semiconductor materials (the so-called direct Z-scheme design) that facilitate spatial separation of (photo)excited electrons and holes, prevent their recombination, and allow water-splitting half-reactions proceeding at each corresponding semiconductor side. In this work, we proposed and prepared a specific structure, based on WO3g–x/CdWO4/CdS coupled semiconductors, created by annealing of a common WO3/CdS direct Z-scheme. WO3–x/CdWO4/CdS flakes were further combined with a plasmon-active grating for the creation of the so-called artificial leaf design, making possible complete utilization of the sunlight spectrum. The proposed structure enables water splitting with high production of stoichiometric amounts of oxygen and hydrogen without undesirable catalyst photodegradation. Several control experiments confirm the creation of electrons and holes participating in the water splitting half-reaction in a spatially selective manner.

surface by adding CdCl 2 solution to the WO 3 aqueous suspension under vigorous stirring for 30 min, and then a Na 2 S aqueous solution was added with the suspension getting a yellowish tint [S1, S2].
The resulting water-soluble sodium chloride was removed from the product by centrifugation and washed twice with deionized water. The substance was further dried in an oven at 60 °C for 2 hours.

Preparation of WO 3 /CdWO 4 /CdS photocatalyst.
The previously obtained WO 3 -CdS system was placed in a quartz crucible and then subjected to annealing in nitrogen atmosphere at 300, 500, 600, 700, 800, or 900 °C for 4 hours. The samples were further referred as WC-X, where X is the annealing temperature.

Preparation of plasmon active Au grating.
The plasmon-active periodical Au grating was prepared by separating two transverse portions of the DVD-R disc (periodical polycarbonate template was obtained in this way) and subsequent sputtering of 30 nm thick Au film on polycarbonate surface (magnetron sputtering, 40 mA, 300 s).

Deposition of WC-X on grating surface. The uniform distribution of WC-X flakes on
Au grating surface was achieved by spin-coating method. Briefly, the optimal deposition parameters were: WO 3 /CdWO 4 /CdS concentration in methanol suspension of 1.1 mg mL -1 , the speed rotation 250 rpm and time of spin-coating 2 min. the frequency range of 1-100 000 Hz (applied potential was 1.8 V vs RHE). In impedance measurements an equivalent electrical circuit, comprising solution resistance, contact interface resistance, and constant phase element was used.
In solely light induced OWS experiments, photocatalyst was deposited on the surface of periodic gold grating (3 × 3 cm 2 ) by spin-coating method and then was immersed in self-made reaction cell filled with pure water without any sacrificial agents and illuminated with simulated sunlight (Solar Simulator SciSun-300, Class AAA). The intensity of light on sample surface was adjusted to be closer to common sunlight intensity (100 mW/cm 2 ). The evolved H 2 and O 2 were determined at 2 h intervals using an online gas chromatography system (GC-7920).

Kelvin Probe measurements and surface potential mapping.
Contact potential difference (CPD) was measured by the Scanning Kelvin Probe method (SKP) (KP Technology, U.K.).
The CPD data were recalculated to the work function (WF) values using the equation: (1) , an Au reference sample from KP Technology was used. SKP measurements were done in a glove box using a grounded steel probe tip of 2 mm diameter. The CPD was measured in the dark-light-dark-light-dark cycle switched on/off the light (simulated sunlight using AM1.5G filter, class AAA solar simulator HAL-C100, Asahi Spectra, Japan), equipped with AM 1.5G filter. In the dark the WF values reflect the intrinsic material properties and possibly retained electrostatic charge too [S3]. Under the illumination, the difference in WF corresponds to photovoltage generation on the sample surface. [S4]. Lower WF (CPD) values correspond in our setup to positive photovoltage, i.e. to generation/transfer of holes on the surface.

Photoluminiscence experiments.
The photoluminescent (PL) spectroscopy was used to confirm the heterojunction mechanism and enhanced charge separation. In particular, an aqueous coumarin 1mM solution was used as a probe. Briefly, 0.05 g of WC-700, WC-0, WO 3 , or CdS powders were dispersed in coumarin solution and illuminated with simulated sunlight (100 mW cm -2 ). The PL spectra were measured after different time of illumination, with utilization of 325 nm excitation wavelength (LED lamps, Thorlabs).

Calculation of hydrogen production quantum efficiency.
The hydrogen production quantum efficiency φ (AQE) can be calculated by following equation: where n is the number of involved electrons, N A is Avogadro constant, I = I 1 + I 2 is the number of incident photons with 780 nm (I 1 ) and 405 nm (I 2 ) wavelengths, R is produced H 2 value For the calculation of incident photons number, the following formula was used: where E is the light energy flux density, t is irradiation time, λ is wavelength, S is irradiated area, h is Plank's constant, and c is speed of light.

Description of VB and CB calculation (separately prepared WO 3 and CdS)
The levels of VB and CB were determined using the combination of Schottky-Mott and Tauc plots.
The Mott-Schottky plot of semiconductors was plotted according to the following equation (for an n-type and p-type semiconductors) [S6]: where C is the space charge capacitance, ε and ε 0 are the dielectric constant of the semiconductor and permittivity in vacuum (8.85 × 10 −14 F×cm −2 ), e is the electronic charge, N D and N A are the number of donors and acceptors for n-type and p-type semiconductors, respectively, and E, E fb , T, and k B are the applied voltage, flat-band potential, Kelvin temperature, and Boltzmann constant, respectively.
According to the slope coefficient of the Mott-Schottky plot (dependence of 1/C 2 on the potential), the type of semiconductor was determined [S7]. Further, based on the semiconductor type and applied potential, the position of closed band (VB or CB) was determined.
The band gap energy was determined from absorption spectra using the Tauc plot. It assumes that the energy-dependent absorption coefficient α can be expressed by the following equation ( ·ℎ ) 1/ = (ℎ − g), where ℎ is the Planck constant, is the photon's frequency, g is the band gap energy, and is a constant. The factor depends on the nature of the electron transition and is equal to 1/2 or 2 for the direct and indirect transition band gaps, respectively [S5]. Coefficient was determined according to the occurrence of a linear section on the Tauс plot (dependence of ( ·ℎ ) 1/ on photon energy).
Obtained results were used for creation of Fig. 6 (positions of VB and CB).

Fermi level position determination and VB, CB bands alignment.
The positions of the Fermi level (E F ) of the semiconductors were determined from the positions of the corresponding valence bands, using the standard procedure of extrapolating the edge(s) of the XPS spectra to its intersection with the background, which corresponds to the difference between the valence level (E V ) and the Fermi level (E F ).
In order to elucidate the alignment of the valence and conduction bands and in turn to confirm the mechanism of the overall water splitting, additional XPS measurements and subsequent calculations were accomplished according to previously published works [S8-S10].
The valence band offset ∆E V was calculated according to equation: Then the conductive band offset ∆E C was calculated as follows: , where and are bandgaps for WO 3 and CdS respectively.

3
Obtained results were used for creation of Fig. 6 (position of Fermi level and bands alignment).