Photodehydrogenation of Ethanol over Cu2O/TiO2 Heterostructures.

The photodehydrogenation of ethanol is a sustainable and potentially cost-effective strategy to produce hydrogen and acetaldehyde from renewable resources. The optimization of this process requires the use of highly active, stable and selective photocatalytic materials based on abundant elements and the proper adjustment of the reaction conditions, including temperature. In this work, Cu2O-TiO2 type-II heterojunctions with different Cu2O amounts are obtained by a one-pot hydrothermal method. The structural and chemical properties of the produced materials and their activity toward ethanol photodehydrogenation under UV and visible light illumination are evaluated. The Cu2O-TiO2 photocatalysts exhibit a high selectivity toward acetaldehyde production and up to tenfold higher hydrogen evolution rates compared to bare TiO2. We further discern here the influence of temperature and visible light absorption on the photocatalytic performance. Our results point toward the combination of energy sources in thermo-photocatalytic reactors as an efficient strategy for solar energy conversion.

1. Introduction

Molecular n class="Chemical">hydrogen, a clean energy carrier and a key component in the chemical industry, is mostly produced through partial oxidation and steam reforming of natural gas and coal gasification. To move away from the exploitation of fossil fuels, cost- and energy-effective strategies for the direct production of hydrogen from renewable sources need to be defined. In this context, biomass resources are a particularly compelling alternative source of hydrogen owing to their renewable character and their near net-zero CO2 footprint [1,2,3,4,5,6]. Additional advantages of the hydrogen production from dehydrogenation of biomass-derived organics are the potential to co-produce valuable side organic chemicals for better process economics and the possibility to implement cost-effective waste abatement processes [7,8].

Among the possible dehydrogenation processes, photocatalytic routes that make use of ubiquitous, abundant and renewable solar energy are especially attractive. Photocatalytic processes also enable the dehydrogenation reaction to take place in milder conditions, which further decreases costs and can increase the side product selectivity compared with thermocatalytic analogs [2]. From another point of view, the photocatalytic production of fuels can be considered as a convenient strategy to store intermittent solar energy [9,10].

In this scenario, the photodehydrogenation of ethanol to produce molecular hydrogen and acetaldehyde using solar light as the only energy input is especially appealing [6,11]. As a liquid, ethanol can be easily stored and transported. Besides, ethanol can be easily produced from several biomass-derived feedstocks and organic residues such as sewage sludge [12,13,14]. Additionally, bioethanol aqueous solutions can be directly used, without the need for purification. Compared with water splitting, the production of hydrogen from ethanol is thermodynamically advantageous (∆G0 = +237 kJ·mol−1 for water oxidation vs. ∆G0 = +41.5 kJ·mol−1 for ethanol oxidation to acetaldehyde), which decreases the energy input required to drive hydrogen production [2,8]. Compared with water splitting, ethanol dehydrogenation also enables a much simpler product purification, preventing the H2 and O2 back reaction. Besides, compared with ethanol photoreforming, ethanol photodehydrogenation to H2 and acetaldehyde could have a threefold higher economical profitability associated with the high economic value of the side product [15].

In terms of catalysts, while photocatalyn class="Chemical">tic n class="Chemical">water splitting requires semiconductors with conduction and valence band edges sufficiently above and below the potentials for H+ reduction and water oxidation, respectively, ethanol dehydrogenation can be activated in semiconductors with significantly lower band gaps. On the other hand, the catalytic dehydrogenation of ethanol competes with the deoxygenation, reforming and decomposition reactions, which makes the selectivity of the catalytic process fundamental to ensure cost-effectiveness [1].

n class="Chemical">Copper oxides, Cu2−xO, have raised increasing attention as photocatalytic materials owing to their abundance, low cost, minor environmental and health impact and suitable optoelectronic properties. Cu2−xO are p-type semiconductors with a very energetic conduction band and a relatively low bandgap: 2.1 eV for Cu2O and 1.2 eV for CuO, which enables absorption of the visible range of the solar spectra. As a drawback, Cu2−xO have poor photostability, being prone to photocorrosion in reaction conditions. Besides, Cu2−xO generally presents a large defect density that results in a relatively fast recombination of photogenerated charge carriers. To solve these limitations, Cu2−xO can be combined with TiO2 within p-n heterojunctions that protect Cu2−xO against photocorrosion and reduce the charge carrier recombination. The synergism between the two materials is enabled by the appropriate conduction band edges of Cu2−xO, −1.79 V for Cu2O and −1.03 V for CuO, which allows the rapid injection of the photogenerated electrons from the Cu2−xO to the TiO2 conduction band [8,15,16,17,18,19]. Thus, the combination of Cu2−xO and TiO2 is regarded as a highly interesting photocatalyst to: (i) stabilize the Cu2−xO, (ii) boost the overall catalytic activity by extending the light absorption of TiO2 toward the visible light range and (iii) maximize external quantum yield by a rapid charge separation between the two phases enabled by their adequate band edges.

While the concept of a p-n heterojuncn class="Chemical">tionpan> betweenpan> Cu2−xO and TiO2 that promotes catalytic activity is pleasantly simple, real systems are much more complex, and Cu2−xO have been reported to promote catalytic activity through several different mechanisms: (i) Cu2−xO can absorb the visible light and transfer photogenerated electrons to TiO2, where H2 evolves, while using photogenerated holes to oxidize the organic species [20]. (ii) Cu2−xO can absorb visible light but use photogenerated electrons to evolve H2 and recombine in photogenerated holes at the Cu2−xO/TiO2 interphase within a Z-scheme mechanism [21]. (iii) Cu2−xO nanoparticles can be reduced to metallic copper during ethanol photodehydrogenation, and the resulting metal nanoparticles can act as a cocatalyst, stabilizing photogenerated electrons, promoting the water reduction reaction, simultaneously reducing the rate of charge recombination and, thus, also making more holes available for the oxidation reaction [15,22,23]. (iv) Cuδ+ and Cu0 on the surface of supported Cu clusters can also participate as catalysts in the ethanol oxidation to acetaldehyde [17]. (v) Copper ions can be partially incorporated into the TiO2 lattice by substituting for Ti4+ ions and creating oxygen vacancies that decrease the TiO2 bandgap [15,19,24,25]. All these effects strongly depend on the synthesis procedure, the TiO2 surface area and its structural and chemical properties, which affect the Cu dispersion and oxidation states [22] and the TiO2 phase that also determines the interaction with Cu and the Cu role [26].

Most previous works assign the performance promotion of Cu2−xO/TiO2 with respect to TiO2 to the extension of light absorption toward the visible range of the solar spectra. However, in most previous works, mainly UV excitation is used, and the overall and local temperature changes associated with the visible light absorption are usually neglected.

In the present work, we aim at gaining addin class="Chemical">tionpan>al understandinpan>g of the mechanism behinpan>d the synergistic promotion of the catalytic performance in Cu2O/TiO2 while simultaneously contributing to the optimization of this system. In this direction, we present a one-pot hydrothermal synthesis strategy to produce Cu2O/TiO2 nanocomposites with controlled Cu2O amounts. The photocatalytic performance of Cu2O/TiO2 toward ethanol dehydrogenation is tested using both UV and visible light irradiation. We then determined the direct contribution of visible light, beyond the increasing temperature, toward increasing catalytic activity. We tested photocatalytic activity in the gas phase as it offers additional advantages, including lower light scattering, easier scale-up, higher stability, easier product recovery and even higher selectivity [7,15]. Besides, using time-resolved photoluminescence measurements and analyzing the band alignment between the two materials, we showed the activity promotion to proceed through a conventional p-n type II heterojunction.

2. Materials and Methods

2.1. Chemicals

Titanium (IV) isopropoxide (97%, Sigma-Aldrich, St. Louis, MO, USA), copper (II) nitrate hexahydrate (98%, Fluka, Buchs, Switzerland), ethanol (96%, PanReac AppliChem GmBH, Darmstadt, Germany), polyvinylpyrrolidone (PVP, 90%, Sigma-Aldrich, St. Louis, MO, USA), and sodium sulfate (Alfa Aesar™, Ward Hill, MA, USA) were used without further purification.

2.2. Synthesis of Photocatalysts

n class="Chemical">PVP (0.45 g) was dissolved inpan> Milli-Q n class="Chemical">water:ethanol (1:2) (40 mL) under stirring at room temperature. To this solution, a proper amount of Cu (NO3)2·6H2O was added (0, 11.3, 22.5, 45 and 112.5 mg to reach 0%, 0.5%, 1%, 2% and 5%, respectively) by stirring for 5 min. Then, 2.3 mL of titanium (IV) isopropoxide was added dropwise, followed by stirring for 10 h at room temperature. Finally, the suspension was transferred to a 50-mL Teflon-lined autoclave and maintained at 170 °C for 14 h.

2.3. Structural and Chemical Characterization

The morphology and size of the parn class="Chemical">ticles were obtainpan>ed by transmissionpan> electronpan> microscopy (TEM) usinpan>g a ZEISS LIBRA 120 (Carl Zeiss, Jenpan>a, Germany) inpan>strumenpan>t. Elemenpan>tal analysis was carried out usinpan>g an Oxford enpan>ergy dispersive X-ray spectrometer (EDX) combinpan>ed with the Zeiss n class="Chemical">Auriga SEM (Carl Zeiss, Jena, Germany) working at 20.0 kV. The crystal structure of the samples was determined by X-ray diffraction (XRD) using a D8 Advance (Bruker, Billerica, MA, USA) equipment with Ni-filtered Cu-Kα radiation (λ = 0.15406 Å) operating at 40 mA and 40 kV. UV-Vis absorption spectra were recorded on a UV-Vis spectrophotometer (Shimadzu, UV-3600i Plus, Tokyo, Japan), and BaSO4 was used as a reference standard. The spectra were recorded at room temperature in the air within the range of 300–800 nm. High-resolution transmission electron microscopy (HRTEM) images and scanning transmission electron microscopy (STEM) studies were conducted on an FEI Tecnai F20 field emission gun microscope operated at 200 kV with a point-to-point resolution of 0.19 nm, which was equipped with high angle annular dark-field (HAADF) and a Gatan Quantum electron energy loss spectroscopy (EELS) detectors. X-ray photoelectron spectroscopy (XPS) was done on a SPECS system (SPECS GmbH, Berlin, Germany) equipped with an Al anode XR50 source operating at 150 mW and a Phoibos 150 MCD-9 detector (SPECS GmbH, Berlin, Germany). Data processing was performed with the CasaXPS program (Casa Software Ltd., Teignmouth, UK). Steady-state photoluminescence (PL) spectra were conducted by a high-resolution photoluminescence spectrofluorometer (Horiba Jobin Yvon Fluorolog-3, Palaiseau, France). For the time-resolved photoluminescence spectroscopy (TRPL) measurements, a nanosecond LED with a 350-nm peak wavelength (Horiba NanoLED N390, Palaiseau, France, pulse width < 1.3 ns) was applied to excite the samples. The TRPL decay was resolved at 400 nm. Average lifetimes were obtained by fitting the TPPL spectra with DAS6 software (Horiba, Palaiseau, France).

2.4. Photoelectrochemical Measurements

Photoelectrochemical (PEC) propern class="Chemical">ties were measured usinpan>g CHI760e (CHI 760E, n class="Disease">CH Instrument, Austin TX, USA) in a three-electrode cell with a platinum mesh as the counter electrode, and an Ag/AgCl reference electrode. Na2SO4 (0.5 M) was used as the electrolyte solution. The working electrode was prepared by depositing Cu2O/TiO2 on an indium tin oxide (ITO) glass electrode (1 cm × 1 cm) and heating at 200 °C for 1 h. Potentials vs. Ag/AgCl were converted into potentials vs. reversible hydrogen electrodes (RHE), according to the Nernst equation (ERHE = EAg/AgCl + 0.059 pH + 0.196). Electrochemical impedance spectroscopy (EIS) measurements were carried out with a sinusoidal ac perturbation of 5 mV applied over the frequency range of 0.01–100,000 Hz. The transient photocurrent (TPC) of the as-prepared photocatalysts was measured with an AM1.5G solar power system used as the light irradiation source (100 mW·cm−2) at an ambient temperature and without any light irradiation source. Mott–Schottky (M–S) measurements were carried out in the dark with a scanning speed of bias potential ranging from −1.4 to 0.2 V at a scan rate of 0.01 V∙s−1. The linear sweep voltammetry was carried out with a scanning speed of bias potential ranging from −1.2 to 0.6 V at a scan rate of 0.01 V∙s−1.

2.5. Photocatalytic Test

In a typical experiment, a cellulose paper impregnated with 2.0 mg of the photocatalyst was n class="Chemical">placed inpan>side a photocatalytic reactor that was equipped with UV LEDs (365 ± 5 nm, from SACOPA S.A.U, Gerona, Spain) (Figure S1). A light irradiation of 79.1 ± 0.5 mW·cm−2 was measured for UV light at the sample position. A saturated Ar gas stream was prepared by bubbling dry Ar gas through a Dreschel bottle with a water:ethanol vapor mixture (9:1, molar ratio, 20 mL∙min−1). The photoreactor effluent was monitored online every 4 min using gas chromatography (GC) (Agilent 3000A MicroGC, Santa Clara, CA, USA) with three columns: MS 5 Å, Plot U and Stabilwax. The system was purged with the saturated Ar stream (20 mL∙min−1, 30 min) to remove oxygen before performing the experiments. The UV-visible light source contained two LEDs emitting at 372 ± 5 nm and two LEDs emitting visible light (correlated color temperature (CCT) 6099 K and color rendering index (CRI) 74) in Figure S1. In this system, UV light irradiation was 11.2 ± 0.5 mW·cm−2 at the sample position.

2.6. Apparent Quantum Yield (AQY) Calculation

The AQY was esn class="Chemical">timated usinpan>g the followinpan>g equation: where is the number of evolved hydrogen molecules, and np is the number of incident photons reaching the catalyst. The number of incident photons can be calculated by , where ET is the total energy reaching the catalyst, and Ep is the energy of a photon. ET can be calculated by , where P (W·m−2) is the power density of the incident monochromatic light, S (m2) is the irradiation area and t (s) is the duration of the incident light exposure. Ep can be calculated by , where h is the Planck’s constant, c the speed of light and λ (m) is the wavelength of the incident monochromatic light. The number of hydrogen molecules can be calculated as , where n is the H2 moles evolved during the time of light exposure (t), and NA is the Avogadro constant. In our experimental conditions with UV light, the wavelength of the incident light was λ = 365 nm, the power density of the incident light at the paper surface was P = 79.1 mW·cm−2 and the irradiation area was S = πR2 = 3.14 × 0.752 = 1.77 cm2.

3. Results and Discussion

3.1. Structural, Chemical and Optical Properties

n class="Chemical">Cu2O/n class="Chemical">TiO2 nanocomposites with different Cu2O loading, between 0.5% and 5%, were synthesized by the hydrothermal reaction of copper (II) nitrate hexahydrate and titanium (IV) isopropoxide at 170 °C for 12 h. Figure 1a shows the XRD patterns of the TiO2 and Cu2O/TiO2 nanopowders. The main XRD peaks of all patterns could be indexed with the tetragonal anatase TiO2 phase (JCPDS No. 01-071-1167). Additional XRD peaks at 2θ = 36.4° and 42.3° were identified in the Cu2O/TiO2 samples containing 1% and higher Cu2O amounts and were associated with the (111) and (200) family planes of the cubic Cu2O cuprite phase. From the XRD patterns, using the Scherrer equation, the size of the TiO2 and Cu2O crystal domains was calculated to be ca. 7 nm and 50 nm, respectively, which pointed at the presence of some large Cu2O crystals.

Figure 1

(a) Powder XRD pattern of TiO2 and 0.5%, 1%, 2% and 5% Cu2O/TiO2 nanocomposites. (b) TEM micrograph of 1% Cu2O/TiO2, with a scale bar of 200nm. (c) HRTEM analysis of the 1% Cu2O/TiO2 sample. The upper image shows a crystal with a tetragonal anatase phase of TiO2 visualized along the [010] zone axis. The lower image shows a cubic Cu2O crystallite visualized along the [111] zone axis. (d) STEM-ADF and STEM-EELS analysis of the 1% Cu2O/TiO2 sample. Cu L-edges at 931 eV (red), O K-edge at 532 eV (green) and Ti L-edge at 456 eV (blue). (e) High resolution XPS spectra for the Ti 2p core level of TiO2 and 1%, 2% Cu2O/TiO2 nanocomposites.

TEM micrographs showed n class="Chemical">TiO2 and n class="Chemical">Cu2O/TiO2 nanopowders that consisted of small nanoparticles with irregular shapes and an average size of ca. 10 nm (Figure 1b and Supplementary Figure S2). HRTEM characterization of the 1% composite further confirmed the presence of both the tetragonal anatase TiO2 and cubic Cu2O phases (Figure 1c). STEM-EELS compositional map displayed the elemental distribution (Figure 1d). By performing the quantitative relative compositional analysis, we could extrapolate that the Ti and O compositions oscillate between 30–35% and 65–70%, respectively. Only traces of Cu could be detected from the 1% composite (TS1). This limitation and the small size of the Cu2O domains observed by HRTEM resulted in a STEM-EELS compositional map showing homogeneous-like copper distributions (Figure 1d). SEM-EDX analysis showed the Cu concentration to match the nominal amount in low Cu-loaded samples but to be lower than expected in 2% and 5% Cu2O/TiO2 nanocomposites (Table S1).

XPS spectra showed the incorporan class="Chemical">tionpan> of n class="Chemical">Cu not to influence the Ti chemical state (Figure 1e and Supplementary Figure S3), which displayed the Ti 2p3/2 and Ti 2p1/2-binding energies at 458.5 eV and 464.2 eV, respectively, consistent with Ti4+ within a TiO2 chemical environment [27,28,29]. Besides, the Cu 2p3/2 and Cu 2p1/2-binding energies were 931.9 eV and 951.9 eV, pointing at a Cu+ chemical state [30,31]. The surface composition of Cu matched the nominal amount of Cu in the 1% Cu2O/TiO2 nanocomposite, but it was lower for the 2% Cu2O/TiO2 nanocomposite, which is, in part, consistent with SEM-EDX analysis and, in part, associated to the formation of relatively large Cu2O particles when increasing the Cu loading, as observed by XRD.

Figure 2 shows the UV-vis spectra of n class="Chemical">TiO2 and n class="Chemical">Cu2O/TiO2 nanopowders and the corresponding Tauc plot calculated as (αhν)1/2 vs. hν to determine the direct bandgap of TiO2 (Figure 2b) and, as (αhν)2 vs. hν, to determine the indirect bandgap of Cu2O (Figure 2c). UV-vis absorption data showed a clear absorption edge at around 3.2 eV consistent with the TiO2 bandgap. No clear shift of the absorption edge was observed with the introduction of Cu, which ruled out a possible bandgap change related to the incorporation of Cu ions within the TiO2 lattice. Besides, when incorporating Cu2O, additional light absorption in the visible region and with an absorption edge of ca. 2.0 eV was clearly observed, consistent with the presence of the Cu2O phase [32].

Figure 2

UV-vis absorption spectra (a) and Tauc plot calculated as (αhν)1/2 vs. hν (b) and as (αhν)2 vs. hν t (c) for TiO2 and 0.5%, 1%, 2% and 5% Cu2O/TiO2 nanocomposites.

3.2. Photocatalytic Activity

Figure 3a,b disn class="Chemical">plays the UV (365 ± 5 nm) photocatalytic activity of TiO2, Cu2O/TiO2 and Cu2O nanopowders toward hydrogen production from a gas phase 10% ethanol solution in water. The composition of the effluent gas was monitored using gas chromatography, which showed acetaldehyde (2) and hydrogen (3) in a 1:1 molar ratio to be the two unique products of the reaction. These results proved both that the hydrogen was generated from the dehydrogenation of ethanol and not from water splitting and that the reaction proceeded with very high selectivity toward acetaldehyde production, following the scheme [33]:

Figure 3

(a) Photocatalytic H2 evolution on TiO2, Cu2O, 0.5%, 1%, 2% and 5% Cu2O/TiO2 nanocomposites under UV light irradiation (365 ± 5 nm and 79.1 ± 0.5 mW·cm−2). (b) HER from data displayed in panel (a). (c) HER measured on TiO2, 0.5%, 1%, 2% and 5% Cu2O/TiO2 nanocomposites under different conditions: (1) UV light irradiation (372 ± 5 nm and 11.2 ± 0.5 mW·cm−2), (2) UV (372 ± 5 nm and 11.2 ± 0.5 mW·cm−2) plus visible light irradiation (0.017 ± 0.005 mW·cm−2), (3) UV light irradiation and (4) UV light irradiation and heating to compensate for the temperature (~36−37 °C). (d) HER obtained from the data displayed in panel (c).

The n class="Chemical">hydrogen evolution rate (HER) measured under UV light for the reference TiO2 was 2.4 mmol h−1·g−1 (Figure 3a,b). HER strongly increased with the introduction of Cu2O (Table S2). Among the series of Cu2O/TiO2 samples tested, the highest HRE were obtained for the 0.5% and 1% Cu2O/TiO2 samples that displayed a HER of 20.5 mol∙g−1∙h−1 and 24.5 mmol∙h−1∙g−1, a factor of 10 above bare TiO2. Higher Cu2O loadings that resulted in lower HER, 13.6 and 10.7 mmol∙g−1∙h−1 for the 2% and 5% samples, respectively. We hypothesize the lower HER obtained when increasing the Cu loading above 1% to be related with an increase of the recombination rate associated with a faster recombination of the charge carriers photogenerated in the Cu2O phase than in the TiO2 phase. Besides, the formation of larger Cu2O domains when increasing the Cu loading could also play an important role. The AQY of the 1% Cu2O/TiO2 was 6.4%, whereas the AQY for TiO2, 0.5%, 2% and 5% Cu2O/TiO2 were 0.6%, 5.3%, 3.5% and 2.8%, respectively (Figure S4). Table S2 displays a comparison of the AQY obtained here with those obtained in previous works. On the other hand, the HER of bare Cu2O was very moderate, just 0.8 mmol∙h−1∙g−1, demonstrating both the important role played by TiO2 in the separation of charge carriers and the synergism between the two materials to optimize photocatalytic activity. Figure S5 displays the HER of the 1% Cu2O/TiO2 sample, measured three consecutive times during 1 h, showing the notable HER stability of the system. Besides, in contrast to some previous works, we observed no color change of our samples during the photocatalytic reaction in the presence of ethanol [34,35]. It should be noted, that beyond the convenient use of aqueous ethanol solutions, as produced from biomass processing, the presence of water is beneficial to increase of the catalyst activity and stability by preventing active sites to be blocked by acetaldehyde, which exhibits a strong affinity towards inorganic oxide surfaces [36].

Figure 3c shows the photocatalyn class="Chemical">tic HER acn class="Chemical">tivities of Cu2O/TiO2 under UV light (372 ± 5 nm) and when combining UV light with visible light or heat (see the experimental section for details). It should be noted that, under visible light, there is an increase in the temperature of the photocatalyst; thus, it is necessary to separate the effect on HER of the temperature increase and the photogenerated charge carriers obtained with the visible light absorption. Thus, the photocatalytic test was divided into four consecutive steps: (i) After turning on the UV light, HER began to rise until it stabilized. At this stage, the sample temperature was ca. 25 °C. (ii) Keeping the UV light on, the visible light was turned on, which increased the HER of all samples. The introduction of visible light also increased the sample temperature, up to ca. 36–37 °C (Table S3). (iii) With the UV light on, the visible light was turned off, which resulted in a relatively slow decline of the HER and a temperature decrease down to 25 °C. The slow HER decrease already denoted a significant effect of temperature on the increase of HER observed with the visible light. (iv) Finally, still maintaining the UV light on, the reactor was heated to 36–37 °C (Table S3), which also resulted in an increase of the HER for all catalysts. By comparing stages 2 and 4, the effect of temperature and photogenerated electron–hole pairs can be differentiated.

Non class="Chemical">tice that the addition of visible light increased the HER of TiO2 by a factor of two, which was associated with a 10 °C increase in temperature (Figure 3c,d). This twofold HER increase points toward the combination of energy sources in thermo-photocatalytic reactors as an efficient strategy of solar energy conversion. Such a strong influence of temperature on HER is likely related to the high adsorption energy of acetaldehyde on the oxide surface, blocking the catalyst active sites and, thus, slowing down the reaction. A moderate increase in temperature can significantly reduce the acetaldehyde adsorption strength, thus unblocking active sites and increasing the activity [37].

n class="Chemical">Cu2O/n class="Chemical">TiO2 catalysts displayed a much higher increase of activity with the addition of visible light (Figure S6) by close to a factor of three in 1% Cu2O/TiO2. Only a small fraction of this increase in activity can be associated with the increase of temperature, as observed in Figure 3c,d. The much larger increase of HER obtained with visible light irradiation compared to the sample heating to the same temperature suggests a significant contribution of photogenerated charge carriers in Cu2O.

3.3. Photoluminescence and Photoconductivity

The photocatalyn class="Chemical">tic performance of the semiconpan>ductor photocatalyst is n class="Chemical">tightly related to their charge transport, separation and transfer processes, which closely rely on their relative electronic energy level positions. To understand the photocatalytic process and to gain insights from the enhanced performances of the Cu2O/TiO2 nanocomposites, a series of spectroscopic analyses was performed. Figure S7 displays the steady-state PL spectra of TiO2 and 1% Cu2O/TiO2. In both spectra, a peak at around 400 nm, associated with the band-to-band radiative recombination in TiO2, was observed [38,39,40,41,42]. The presence of Cu2O resulted in a decrease of the peak intensity, which denoted an influence of Cu2O on the recombination of charge carriers photogenerated in TiO2. Figure 4a displays the TRPL spectra of TiO2 and 1% Cu2O/TiO2 at 400 nm. The PL intensity of both samples was observed to decay at a similar rate, with an average photocarrier lifetime of 32.3 ns for 1% Cu2O/TiO2 and 34.0 ns for TiO2. This result demonstrated a minor influence of Cu2O on the band-to-band recombination within TiO2, thus pointing again toward a minor or null influence of Cu within the TiO2 lattice.

Figure 4

(a) TRPL decay of the TiO2 and 1% Cu2O/TiO2 composites. (b) Transient photocurrent response for TiO2 and 0.5%, 1%, 2% and 5% Cu2O/TiO2 composites. (c) Current density vs. potential (RHE) and (d) Nyquist plot with the EIS data obtained from TiO2 and the 1% Cu2O/TiO2 composite in the dark (off) and under illumination (on) at the AM1.5G solar power system 100 mW·cm−2 light irradiation.

The photoelectrochemical behavior of n class="Chemical">TiO2 and 1% n class="Chemical">Cu2O/TiO2 samples supported on an ITO-covered glass substrate were measured under dark and 100-mW·cm−2 AM 1.5G irradiation. As shown in Figure 4c, the photocurrent density measured for 1% Cu2O/TiO2 was higher than that obtained for TiO2. Figure 4b displays the TPC data obtained from the TiO2 and Cu2O/TiO2 composites with different Cu2O loadings. The 1% Cu2O/TiO2 electrode showed the highest photocurrent densities, well above those obtained for bare TiO2. The stable photocurrent of all Cu2O/TiO2 samples pointed at a good stability of the composites under illumination in the solution. The 5% Cu2O/TiO2 sample showed the largest TPC transient spikes, indicating the highest degree of surface charge recombination, which is consistent with its lower HER catalytic performance (Figure 3b) and suggests that an excessive amount of Cu2O hampers the photocatalytic activity due to excessive charge carrier recombination [43].

Figure 4d disn class="Chemical">plays the Nyquist n class="Chemical">plot with the EIS data obtained from TiO2 and 1% Cu2O/TiO2 in the dark and under illumination. EIS analysis showed the 1% Cu2O/TiO2 sample to be much less resistive than TiO2 [44], suggesting that the formation of the heterojunction facilitates the charge transport and injection. Data were fitted with a Randles equivalent circuit consisting of a series resistor , a bulk resistor for charge transport resistance and a bulk capacitor for space charge region capacitance (Table S4) [45]. With the incorporation of only 1% of Cu2O, the value of was reduced from 4.18 Ω to 15.5 Ω in the dark and to even lower values under AM1.5G irradiation.

3.4. Determination of Heterojunction Band Position

To determine the band alignment of the n class="Chemical">Cu2O/n class="Chemical">TiO2 heterojunction; the M–S analysis was performed on pristine TiO2, Cu2O and 1% Cu2O/TiO2, considering: where is the space charge capacitance in the semiconductor, is the electron carrier density, is the elementary charge (1.60 × 10−19 C) and is the vacuum permittivity (8.85 × 10−12 F∙m−1). The considered relative permittivity was ε = 55 for TiO2 and ε = 6.3 for Cu2O [46].

Figure 5a shows the M–S plots of TiO2, Cu2O and 1% Cu2O/TiO2. is determined as: where is the best fit of their linear range of vs. V (12 × 109 cm4 ∙F−2 for TiO2 and 9 × 1010 cm4∙ F−2 for Cu2O). As expected, TiO2 shows a positive value in the linear region in accordance with its n-type character, while Cu2O shows a negative value consistent with its p-type behavior [30]. The M–S analysis resulted in = for TiO2 and = for Cu2O.

Figure 5

(a) M-S analysis of TiO2 and 1% Cu2O/TiO2. (b) M-S analysis of a Cu2O. (c) Energy band diagrams for Cu2O and TiO2 before contact. (d) Scheme of the Energy band structure of a Cu2O/TiO2 heterojunction and the ethanol dehydrogenation reaction.

The effecn class="Chemical">tive denpan>sity of states inpan> the conpan>duction band () is given by: where mde is the density-of-state effective mass for electrons of nano-crystalline anatase TiO2, is Planck’s constant (6.62607004 × 10−34 m2∙kg∙s−1), is Boltzmann’s constant (1.38064852 × 10−23 m2∙kg∙s−2∙K−1) and is the absolute temperature (298 K). For TiO2 a =10 was used for Nc calculations, where (9.109 × 10−31) is the mass of a free electron. For Cu2O = 0.58 m0 is taken as the effective hole mass.

Boltzmann stan class="Chemical">tistics was applied to determine the position of the bottom of the conduction band for TiO2 and the maximum of the valence band for Cu2O (3): where is the Fermi level position ( = ). was found to be 0.033 eV below the for TiO2 and 0.081 eV above the for Cu2O [47]. Based on the M–S analysis, the electronic band structure of Cu2O and TiO2 is displayed in Figure 5c. The values of TiO2 and Cu2O are −0.44 eV and 0.72 eV vs. RHE, respectively. The for TiO2 was −0.41 eV and the for Cu2O is 0.64 eV. As the two materials are brought into contact, there is a net transfer of electrons from n-type TiO2 to p-type Cu2O that results in a bending of the band structure at the interface. Due to the small size of the crystal domains, this bending extends through all the whole TiO2 and Cu2O crystals that are in contact with each other. (Figure 5d). In the resulting heterostructure, photogenerated electrons in the conduction band of Cu2O tend to move toward TiO2, where hydrogen generation takes place, and photogenerated holes in TiO2 tend to move toward the Cu2O, where ethanol is oxidized to acetaldehyde (Figure 5d) [48,49].

4. Conclusions

A simn class="Chemical">ple onpan>e-pot method for the synthesis of n class="Chemical">p-n Cu2O/TiO2 heterostructures was presented. Using UV-vis spectroscopy, and M–S analyses, we showed the formation of a p-n heterojunction between Cu2O and TiO2, which favors the separation of electron–hole pairs. They obtained nanocomposites at 0.5%, 1%, 2% and 5% Cu2O loading were tested for the photocatalytic dehydrogenation of ethanol in water:ethanol vapor mixture. We demonstrated the composites to be photostable catalysts capable of working in a light absorption towards the visible range, with an outstanding selectivity to the production of acetaldehyde and hydrogen from ethanol. The optimum composition contained 1% of Cu2O and showed a yield for HER of 24.5 mmol∙g−1∙h−1 and an AQY = 6.4%. The EIS analysis showed the 1% Cu2O/TiO2 sample to be less resistive than TiO2 sample and suggested that the heterojunction facilitated the charge transport and injection. The addition of visible light increased the HER of the samples by a factor of two, which was partially associated with an increment in the reaction temperature of around 10 °C. We further discerned the influence of temperature and photogenerated electron–hole pairs in the HER increase upon visible light irradiation, demonstrating the important role of photogenerated charge carriers in the presence of Cu2O. Besides, our results open new opportunities for efficient solar energy conversion by the combination of energy sources in thermo-photocatalytic reactors.