Mesoporous assembled structures of Cu₂O and TiO₂ nanoparticles for highly efficient photocatalytic hydrogen generation from water

Abstract

Photocatalytic water splitting to produce hydrogen using solar energy is a particularly attractive solution to increasing energy demands. However, to be of practical use, semiconductor electrodes need to be made of inexpensive, abundant elements and have a high, yet stable, photocatalytic H₂-production activity. Here we report the first demonstration of 3D mesoporous networks of Cu₂O and TiO₂ nanoparticles as highly efficient photocatalysts for hydrogen generation from water. These assembled structures feature a highly accessible pore surface that exposes a large fraction of anatase TiO₂ and Cu₂O nanoparticles to electrolytes, and has a small grain size of the constituent nanocrystals, which lead to excellent activity for H₂ evolution via a UV-visible light-driven reduction of protons. Catalytic results associated with optical UV-vis/NIR absorption and photoluminescence (PL) data indicated that the large separation of photogenerated electrons and holes at the Cu₂O–TiO₂ p–n junctions was the main reason attributed to the improved photochemical performance. Consequently, the mesoporous Cu₂O/TiO₂ catalyst containing ∼1.5 wt% Cu reaches an average H₂ evolution rate of ∼542 μmol h⁻¹ (or ∼36 133 μmol h⁻¹ g⁻¹) with an apparent quantum efficiency (QE) of 13.5% at 365 nm and an incident photon conversion efficiency of ∼4.1% under UV-visible light illumination (360–780 nm), which is one of the best HER activities among TiO₂-based semiconductor systems reported to date.

---

Introduction

Photocatalytic water splitting promise to be an economically feasible and environmentally safe chemical process for producing hydrogen. This process uses a semiconductor-based electrode to reduce protons into hydrogen by using sunlight as the energy source. However, despite intense efforts, photocatalytic H₂ generation still faces several challenging issues, such as a low photon-to-hydrogen conversion efficiency and poor chemical stability of the photocatalysts. Current reported semiconductor photocatalysts that harvest UV to visible light and split water show low quantum efficiency (QE) with a solar-to-hydrogen (STH) efficiency less than 0.1%.^1–3 To meet these challenges, researchers strive to produce semiconductors with well-defined nanostructured morphology in an effort to increase the electron–hole separation yield and, thus, the hydrogen evolution reaction (HER) activity.^4–6 Constructing semiconductor materials with porosity at the mesoscale (with a typical pore diameter of 2–20 nm) has been regarded as an effective way to improve photocatalytic performance. In these materials, as the framework constituents can induce efficient absorption of solar light and electron delocalization properties, the mesoporous structure can provide a large number of surface active sites available for reactions. So far, there are a great number of nanostructured semiconducting materials, such as mesoporous metal oxide frameworks and nanoparticles (NPs), that have been synthesized and investigated as photocathodes for H₂ evolution.^7–9

Nowadays, TiO₂ is the most widely study material for energy conversion and environmental applications, mainly because of its appropriate energy edge potentials, chemical stability, non-toxicity and low cost.¹⁰ Nevertheless, TiO₂ photocatalyst is poorly active for producing H₂ from water, as a results of the low solar light-capturing capability (it absorbs only in UV region) and low charge carrier separation yield.¹¹ To this end, various TiO₂-based heterojunction materials have been studied in an effort to sustain high photon collection and chemical STH efficiency. Metal doping with Pt, Pd, Rh and Au co-catalysts has been a prominent quest to improve hydrogen production performance.^12,13 It is well known that these metal NPs retard the electron–hole recombination by acting as electron sinks, and facilitating the interparticle metal/TiO₂ electron-transfer process.¹⁴ However, although noble metal NPs, like Pt, allow considerable reduction in energy cost for H₂ generation, they are too scare and costly for large-scale deployment.

An alternative and cost-efficient TiO₂-based photocatalyst system is based on loading 3d-transition metal oxides, such as Fe₂O₃,¹⁵ Cu₂O,^16,17 NiO¹⁸ and Co₃O₄,¹⁸ to promote HER catalysis. Among them, Cu₂O/TiO₂ p–n heterojunction composites have attracted particular attention for their potential of achieving high H₂ evolution performance from water splitting.^19–23 In fact, owing to the intrinsic electric field across the Cu₂O–TiO₂ junctions, the photogenerated carriers can be spatially separated into the composite structure, thus suppressing the recombination of electron–hole pairs.²⁴ Thanks to the proper matching of band structure between Cu₂O and TiO₂, the photogenerated electrons can transfer from the conduction band (CB) of Cu₂O to TiO₂, while the holes at the valence band (VB) of TiO₂ can move to the VB of Cu₂O; both conduction and valence band levels of Cu₂O (−1.4 V and 0.8 V vs. NHE, pH = 7) lie above those of TiO₂ (−0.6 V and 2.6 V vs. NHE, pH = 7). Recent studies have shown that the dimensional attributes of copper and titanium oxide components in a Cu-loaded TiO₂ photocatalyst are strongly related with the photocatalytic activity. For example, one-dimensional (1D) CuO/TiO₂ nanotubes showed an efficient H₂-production activity as a result of the movement of the electrons and holes along the tube length.^25,26 Therefore, as a consequence of the different chemical reactivity of the oxide nanostructures, a beneficial improvement in charge carriers' separation and hence better photocatalytic activity can be expected.

Recently, we have reported a facile synthesis of well-controlled NP-based mesoporous titania (MTA), using a polymer-templated aggregating self-assembly of TiO₂ NPs.²⁷ Mesoporous networks of TiO₂ NPs were obtained via a sol–gel polymerization of TiCl₄ and titanium(IV)propoxide (TPP) compounds at the surface of polyoxoethylene cetyl ether (Brij-58) block copolymer micelles, followed by calcination in air (350 °C). Because of the large expose surface area and small grain size of anatase particles, MTA architecture offers a unique platform to develop new composite materials and investigate their electrochemical properties. One such interesting feature of small-sized TiO₂ nanocrystals is the efficient electron-transport to the solid/electrolyte interface.^28,29 In this work, we use a simple chemical reduction deposition method to grow Cu₂O nanocrystals on the surface of mesoporous titania assemblies. The resulting materials possess a three-dimensional (3D) network consisting of connected Cu₂O and TiO₂ NPs and exhibit high internal surface area and narrow pore size distribution. To our knowledge, this is the first report on the use of a highly porous network of Cu₂O-conjugated TiO₂ NPs for photocatalytic hydrogen production. Previous reports have demonstrated the use of Cu-loaded TiO₂ materials in nanotube,²⁶ isolated nanoparticle, nanoplate¹⁸ or thin film morphologies²⁰ as photocatalysts for photoelectrochemical water splitting. Here we show that the assembled structure of these composites is highly effective and well tolerated in producing H₂ from water under UV-visible irradiation. We found that mesoporous Cu₂O/TiO₂ composite at 1.5 wt% Cu loading leads to a considerable H₂-production enhancement, which corresponds to a ∼135-fold increase of HER activity compared to that of pristine mesoporous TiO₂ (MTA). The associated mechanism of the photocatalytic H₂ production under UV-visible light is also discussed.

Results and discussion

Morphology and structural characterization

Mesoporous Cu₂O/TiO₂ nanocomposites were synthesized by a wet chemical deposition method of copper oxide/hydroxide species on the surface of MTA, followed by reduction with glycose. A series of heterojunction composite materials with different Cu loadings, i.e. 0.5, 1, 1.5 and 2 wt% (denoted as CuMTA), was prepared using various concentrations of copper nitrate in the reaction mixture. To confirm the presence of incorporated copper oxide particles and to estimate the chemical composition, we analyzed the as-prepared CuMTA samples with energy dispersive X-ray spectroscopy (EDS). The EDS spectra indicated atomic Ti/Cu compositional data that consist of the expected Cu loadings from the stoichiometry of reactions (Table 1).

Table 1 Composition and morphological properties of CuMTA materials

Sample	Atomic ratio^a (Ti : Cu)	Cu loading (wt%)	Surface area (m^2 g^−1)	Pore volume (cm^3 g^−1)	Pore size (nm)
a Based on the EDS analysis.
CuMTA-1	99.38 : 0.62	0.49%	141	0.27	7.1
CuMTA-2	98.80 : 1.20	0.96%	130	0.24	7.0
CuMTA-3	98.17 : 1.83	1.47%	129	0.23	7.0
CuMTA-4	97.59 : 2.41	1.93%	126	0.24	6.8
CuMTA-3P	98.19 : 1.81	1.45%	121	0.23	6.7

---

The mesostructure and crystallinity of the obtained porous materials were carefully investigated by transmission electron microscopy (TEM), selected-area electron diffraction (SAED), X-ray diffraction (XRD), and nitrogen physisorption analysis. A typical TEM image of the CuMTA-3, which is the most active catalyst studied in this work, is shown in Fig. 1a. The image shows that this material consists of a porous network of connected TiO₂ NPs with uniform size (ca. 7.3 nm). Also, it is clearly observed that some discrete Cu₂O nanocrystals (appeared as dark areas) with an average diameter of 5.7 nm, as inferred by TEM analysis (Fig. 1b), are well dispersed over the TiO₂ framework (bright region). The high resolution TEM (HRTEM) image, in Fig. 1c, shows the lattice fringes of both TiO₂ and Cu₂O components, suggesting the well-defined crystal structures. The lattice fringes with d-spacing of 3.5 Å are assigned to the (101) planes of anatase TiO₂, while those with 2.4 Å lattice spacing correspond to the (111) faces of the cubic cuprite structure (Pn3̄m, a = 4.267 Å). It is worth noting that Cu₂O NPs appear to be in intimate contact with TiO₂ particles. Such a close interconnection between Cu₂O and TiO₂ particles is particularly important for photocatalytic applications where the interfacial electron transfer from Cu₂O to TiO₂ is believed to enhance the charge carrier separation and, thus, the photocatalytic efficiency. In agreement with TEM observations, scanning electron microscopy (SEM) images of CuMTA-3 show that very small particles of size less than 10 nm are agglomerated into large structures, forming a 3D network (Fig. S1†). Further structural information was obtained from selected area electron diffraction analysis. The SAED pattern of CuMTA-3 sample depicts several Debye–Scherrer diffraction rings that can be indexed to the crystal planes of the TiO₂ anatase structure (marker with red lines) and the cubic lattice of Cu₂O (marker with yellow lines), see Fig. 1d.

Fig. 1 (a) Typical TEM image, (b) size distribution plots of Cu₂O and TiO₂ NPs showing an average particle diameter of 5.7 ± 0.8 nm and 7.3 ± 1.1 nm, respectively, (c) HRTEM image and (d) SAED pattern for mesoporous CuMTA-3.

The powder XRD patterns of the mesoporous CuMTA materials clearly demonstrate the well-developed nanocrystalline structure of TiO₂, showing several broad reflections in the 2θ = 20–80° range (Fig. S2†). All reflections peaks match well with the Bragg diffractions of the standard anatase structure of TiO₂ (JCPDS card no. 21-1272) with an average grain size, as estimated from the broadening of the (101) peak and the Scherrer equation,³⁰ of about 9 nm, in good consistency with TEM results. However, Cu₂O phase is not detected by XRD analysis in CuMTA composites owing to the low amounts present and high dispersion of Cu₂O particles on the surface of TiO₂.

The mesoporosity of the title materials was determined by nitrogen physisorption at 77 K. The N₂ adsorption–desorption isotherms of mesoporous CuMTA are shown in Fig. 2a. All isotherms exhibit typical type-IV curves, according to the IUPAC classification, with a sharp capillary condensation step at a relative pressure (P/P₀) of 0.75, indicating mesoporous structure with narrow-sized pores.³¹ Of note, the H₃-type hysteresis loop between the adsorption and desorption branch of isotherms suggests the existence of an interconnected pore system between the nanoparticles. The CuMTA composites exhibited Brunauer–Emmett–Teller (BET) surface areas in the range of 126–141 m² g⁻¹ and total pore volumes in the range of 0.23–0.27 cm³ g⁻¹, which are slightly lower than that of the pristine MTA sample (153 m² g⁻¹, 0.29 cm³ g⁻¹, see ESI Fig. S3†), plausibly due to the deposition of Cu₂O NPs on the titania surface. The pore width in CuMTA materials was determined by using the non-local density functional theory (NLDFT) model on the adsorption data, and was found to be in the range 6.8–7.1 nm with narrow size distribution (Fig. 2b). In comparison with the pore size of the pristine MTA (ca. 7.2 nm), the pore diameter in CuMTA is slightly lower and progressively decreased with increasing Cu content. This suggests that Cu₂O particles are embedded in the porous structure of MTA. All the textural properties of the mesoporous CuMTA materials are listed in Table 1.

Fig. 2 (a) Nitrogen adsorption–desorption isotherms at 77 K and (b) the corresponding NLDFT pore-size distribution plots calculated from the absorption branch of isotherms for mesoporous CuMTA materials. For clarity, the isotherms of CuMTA-2, CuMTA-3 and CuMTA-4 are offset by 70, 120 and 140 cm³ g⁻¹, respectively.

The presence of copper oxide NPs in CuMTA mesoporous was also verified by ultraviolet-visible/near-IR (UV-vis/NIR) light diffuse reflectance spectroscopy. The sharp optical absorption at wavelengths shorter than 390 nm is assigned to the intrinsic interband (O_2p → Ti_3d) electron transitions in TiO₂ (Fig. 3). The band gap of the CuMTA semiconductors can be estimated from the Tauc plot [(αhν)² versus hν, in which α, h and ν are the absorption coefficient, Planck's constant and light frequency, respectively], and thus was found to be 3.2–3.3 eV (Fig. 3, inset). Furthermore, an apparent tail in the visible light region (at wavelength over 400 nm) of the UV-vis/NIR spectra can be clearly observed with the increase of Cu₂O loading. Given that the absorption feature at ∼460 nm is very close to the interfacial transition energy from the VB of TiO₂ (2.6 V vs. NHE, pH = 7) to Cu₂O/Cu (0.05 V vs. NHE, pH = 7),³² it is reasonable to suggest that this absorption may be due to the interface charge transfer (IFCT) from the VB of TiO₂ to Cu₂O.³³ Such photoinduced electron transfer can cause a partial reduction of Cu₂O to Cu⁰ in CuMTA composites, the presence of which has been identified by HRTEM, as we shall discuss below. The absorption at wavelengths longer than 500 nm is attributed to the band edge transition in Cu₂O.³⁴ Notably, we did not observe any appreciable red shift in the energy gap of TiO₂ with increasing the Cu loading. This suggests that Cu elements do not incorporated into the titania lattice but rather form Cu₂O deposits on the surface.

Fig. 3 UV-vis/NIR absorption spectra, obtained from the corresponding diffuse reflection data according to the Kubelka–Munk function, for mesoporous CuMTA catalysts. Inset: the corresponding (αhν)² versus energy plots.

Photocatalytic study

We examined the HER activity of the mesoporous CuMTA materials towards the UV-visible light irradiated reduction of water, using methanol (20% by volume) as a hole scavenger. For comparison, similar HER measurements for the mesoporous TiO₂ (MTA) catalyst were also performed. First we optimize the amount of catalyst added to water by measuring the H₂ production for different concentrations of CuMTA-3. We found that the H₂ evolution rate increases with increasing the catalyst addition until reaching a maximum, and then slightly decreases (Fig. S4†). We interpret the improvement of HER rate with the concentration of catalyst to the increased photon absorption by the catalyst's NPs. The highest H₂ evolution rate by the CuMTA-3 was obtained within the range of ∼0.75 g L⁻¹ of the catalyst loading. At an even higher concentration of catalyst, however, the H₂ evolution rate slightly decreased, which is probably due to the scattering of light by the particles. Next, we proceeded to optimize the catalyst composition in a fixed mass of CuMTA composites. As shown in Fig. 4, the photocatalytic activity of the samples increases with increasing Cu content and reaches a maximum H₂ evolution rate at 1.5 wt% of Cu loading (CuMTA-3). As for the lower HER activity of CuMTA-4 sample, which contains 2 wt% Cu, the excessive Cu₂O NPs presumably shield the surface active sites of TiO₂ or act as charge-carrier recombination centers, deteriorating the photocatalytic performance.³⁵ Interestingly, we found that the CuMTA-3 achieves an average H₂ evolution rate as high as ∼542 μmol h⁻¹ (or ∼36 133 μmol h⁻¹ g⁻¹ mass activity) with a 13.5% QE at λ = 365 ± 10 nm and an incident photon conversion efficiency of around 4.1% under UV-visible light illumination (λ = 360–780 nm). This activity represents a significant HER improvement compared to that of mesoporous TiO₂ (MTA) (∼4 μmol h⁻¹), which clearly indicates the presence of Cu₂O makes a significant contribution to enhance the photocatalytic efficiency. Therefore, the enhanced HER activity of the Cu₂O/TiO₂ composite structures can be inferred as the enhancement of the visible light absorption and/or suppression of the electron–hole recombination process within the anatase structure. Control experiments showed that no H₂ evolution can be detected (over a 3 h reaction period) when the reaction performed in the dark or in absence of catalyst or methanol. This suggests that hydrogen was produced by photocatalytic reactions. Table 2 summarizes the HER activity data of the photocatalysts studied.

Fig. 4 Photocatalytic H₂ evolution rate under UV-visible light irradiation (using a 360–780 nm band-pass filter) for mesoporous MTA and CuMTA composite materials. The error bars indicate standard deviation.

Table 2 Photocatalytic H₂ production activity of the mesoporous MTA and CuMTA composite materials under UV-visible light

Catalyst	Evolved H2, 3 h (μmol)	Rate of H2 evolution^a (μmol h^−1)
a Average H2 evolution rate over 3 h irradiation period.b Total amount and average H2 evolution rate under visible light irradiation (λ > 420 nm).
MTA	15	∼4
CuMTA-1	241	82
CuMTA-2	988	341
CuMTA-3	1630 (∼11)^b	542 (∼3.6)^b
CuMTA-4	1224	409
CuMTA-3P	374	120

---

Note here that the HER activity of CuMTA-3 is higher to that obtained for Cu₂O/rGO (rGO, reduced graphene oxide) (∼264 μmol g⁻¹ h⁻¹)³⁶ and Cu₂O–TiO₂/rGO (110 968 μmol g⁻¹ h⁻¹)³⁷ composites, Cu₂O/TiO₂ nanoplates (37.2 μmol h⁻¹, 372 μmol g⁻¹ h⁻¹),¹⁸ CuO/TiO₂ nanotubes (499 μmol h⁻¹, 99 823 μmol g⁻¹ h⁻¹),²⁵ and Cu₂O/TiO₂ (P25) (∼134 μmol h⁻¹, 668 μmol g⁻¹ h⁻¹),¹⁹ Cu–Cu₂O/TiO₂ (P25) (511 μmol h⁻¹, 12 779 μmol g⁻¹ h⁻¹),³⁸ Cu₂O/TiO₂ (P25) (290 μmol h⁻¹, 2900 μmol g⁻¹ h⁻¹)³⁹ and Au/Cu₂O–TiO₂ (1140 μmol h⁻¹, 11 400 μmol g⁻¹ h⁻¹, QE ∼ 7–8%)⁴⁰ NPs, and it is comparable to the activity of mesoporous Cu–CuO/TiO₂ ternary structures (7150 μmol h⁻¹, 286 000 μmol g⁻¹ h⁻¹, QE ∼ 10.5%).⁴¹ To our knowledge, the prepared CuMTA composite system is one of the most efficient photocatalysts reported until now for TiO₂-based photocatalytic hydrogen generation.

The CuMTA-3 catalyst also demonstrated very good stability under the examined conditions. The reusability study was carried out by purging the reaction cell with argon at certain intervals (5 h), until no H₂ and O₂ were detected (by GC analysis). As shown in Fig. 5a, the CuMTA-3 sustains its activity after five catalytic cycles. After 25 hours of illumination, a total amount of H₂ of ∼11 978 μmol (∼268 mL) was obtained, which corresponds to an average HER rate of ∼479 μmol h⁻¹. Assuming that this H₂ evolution rate corresponds to a 114 J energy generated by the reduction of water and the catalyst particles absorb all incident UV and visible light (0.51 W cm⁻²), the photon-to-hydrogen energy conversion efficiency was determined to be ∼2% (in the 360–780 nm interval), which is among the highest values reported in the literature.^42,43 Moreover, elemental X-ray microanalysis and N₂ physisorption studies showed no detachment of Cu₂O NPs from the TiO₂ surface and dissolution of the TiO₂ assembled structure after 25 h of reaction, confirming the high durability of the CuMTA-3 composite under the examined conditions. These results indicated that the regenerated sample have a Cu loading of ∼1.42 wt%, a surface area of 121 m² g⁻¹ and a total pore volume of 0.22 cm³ g⁻¹ (Fig. S5 and S6†). Furthermore, the CuMTA-3 catalyst retains, to a great extent, the high distribution of Cu₂O nanocrystals on the titania matrix after catalysis, as derived from the contrast different and the phase assignment from the TEM images in Fig. 5b and c. The HRTEM also revealed that some Cu nanocrystals with a diameter of about 4–5 nm were formed between the Cu₂O and TiO₂ particles; the image in Fig. 5c shows the lattice fringes with d-spacings of 1.8 Å and 2.1 Å corresponding to the (200) and (111) facets of Cu, along with the lattice fringes of the Cu₂O (111) and anatase TiO₂ (101) structures. The color change of the obtained catalyst to darker gray is also a reasonable indication of the presence of metallic Cu NPs. However, the regenerated catalyst fully regained its original color upon exposure to air for a few hours. It is noted that, for TEM analysis, significant effort was taken to minimize the air exposure time of the reused sample (it was handled under N₂ atmosphere) in order to avoid spontaneous oxidation of Cu metals to Cu₂O.⁴⁴ In agreement with TEM results, the optical absorption spectrum of the regenerated catalyst suggested that metallic Cu is also present in the composite structure, displaying a broad feature at ∼600 nm due to the localized surface plasmon resonance (LSPR) of Cu NPs (Fig. S7†).^45,46 Taken together, it is obvious that a partial transformation of Cu₂O into metallic Cu occurs during the photocatalytic process. A similar structural transformation of Cu₂O into Cu⁰ upon UV irradiation of Cu₂O/TiO₂ composites have been observed by other researchers.^19,38,40

Fig. 5 (a) Recycling study of CuMTA-3 catalyst and typical (b) TEM and (c) HRTEM images of the regenerated CuMTA-3 sample. In inset of panel (c), the corresponding FFT pattern, taken from the area marked by the dashed circle, was indexed as the [011] zone axis of face-centered cubic (Fm3m) Cu.

To better understand the effect of Cu₂O NPs on H₂ production activity of CuMTA, we ran visible light (λ > 420 nm) irradiated HER using the CuMTA-3 as catalyst. In this experiment, CuMTA-3 exhibited little H₂ evolution activity (∼3.6 μmol h⁻¹), which is more than an order of magnitude lower than that under UV-light illumination (Table 2). The results from this study clearly indicate that even through incorporation of Cu₂O NPs on the titania surface can induces visible light absorption (Cu₂O has an energy gap of ∼2.1–2.2 eV),³² as inferred by UV-vis spectroscopy, it does not seem to be a reliable explanation for the increased photoactivity. Therefore, we can infer that the greatly improved HER performance of CuMTA mesoporous mainly arising from the efficient charge carrier dissociation at the Cu₂O/TiO₂ interface that retards the recombination of photogenerated electrons and holes. To investigate whether electron–hole pair separation occurs in the CuMTA heterostructures, we performed photoluminescence (PL) measurements. As seen in ESI Fig. S8,† the mesoporous TiO₂ (MTA) exhibits two intense light emissions at around 414 and 470 nm due to the radiative recombination of self-trapped excitons (between the sub-bandgap states of TiO₂).⁴⁷ The weak signal at 500–550 nm range could be related to surface oxygen vacancies and defects in TiO₂ lattice.⁴⁴ Comparatively, the PL emission of CuMTA-3 heterostructure is almost vanished. This suggests a limited interband recombination of photoexcited electrons and holes in CuMTA-3 due to the resonance charge transport within the Cu₂O/TiO₂ heterojunctions.

It should be stressed that the chemical nature and phase morphology of Cu species attached on the titania surface are important for promoting photocatalytic H₂ production. As such, we also prepared a Cu-loaded TiO₂ catalyst by photo-deposition of ∼1.5 wt% Cu on the surface of MTA (CuMTA-3P) and we compared its HER activity with that of CuMTA-3. These two materials exhibit very similar textural properties and chemical composition (see Table 1 and Fig. S9†), but feature copper species with different valence state on the titania surface. The HRTEM images in ESI Fig. S10† show that the surface of CuMTA-3P is dominated by metallic Cu deposits with an average size of about 4–5 nm; these NPs display lattice fringe spacings of 2.1 and 1.8 Å nm, which are characteristic of the (111) and (200) planes of cubic Cu. Consistent with TEM observation, UV-vis/NIR absorption spectrum of CuMTA-3P showed a broad feature in the 700–800 nm range due to the plasmonic excitation of metallic Cu (Fig. S11†),⁴⁶ further confirming the Cu chemical nature. Catalytic results indicated that CuMTA-3P affords a ∼120 μmol h⁻¹ H₂ evolution rate under identical illumination conditions, which is remarkably low when compared to the activity of mesoporous CuMTA-3. This observation is consistent with previous studies,⁴⁸ and suggests that Cu/TiO₂ composite is not a suitable photocatalyst for H₂ evolution.

On the basis of the above results, a schematic representation of photocatalytic H₂ production from the mesoporous CuMTA is shown in Scheme 1. In short, upon UV-visible light illumination, both Cu₂O and TiO₂ semiconductor components are photoexcited to produce electrons (e⁻) and holes (h⁺) in the conduction and valence band, respectively. Next, because of the potential gradient at the Cu₂O/TiO₂ interface, the photogenerated electrons from Cu₂O can transfer to the CB of TiO₂ and promote the hydrogen production. In addition, during the course of irradiation a fraction of photogenerated electrons in CB of Cu₂O may be also consumed by the reduction of Cu⁺ ions to Cu NPs. This could be due to the lower reduction potential of Cu₂O to Cu⁰ (E⁰ = 0.05 V vs. NHE) than the CB edges of Cu₂O and TiO₂. Also, the direct IFCT from the VB of TiO₂ to Cu₂O, which results in the partial reduction of Cu₂O to Cu, is a reasonable option. The presence of Cu metal clusters between the TiO₂ and Cu₂O particles in the CuMTA-3 was evidenced by HRTEM analysis and UV-vis/NIR spectroscopy. It was recognized that such Cu/Cu₂O/TiO₂ ternary heterojunctions can enhance tunneling of electrons across the junction, dissociating more efficiently the photoexcited electrons and holes that are available for reactions.^37,49 Meanwhile, the photogenerated holes in the VB of TiO₂ can be transferred to the Cu₂O VB (or to the metallic Cu NPs oxidizing Cu⁰ to Cu⁺) and consumed by methanol. As noted above, the Cu₂O–TiO₂ junctions can facilitate the spatial separation of photoexcited excitons (electron–hole pairs) and promote HER activity. The improved separation of photogenerated electrons and holes within the structure of CuMTA was verified by PL spectroscopy.

Scheme 1 Photocatalytic H₂ production mechanism on the Cu/Cu₂O/TiO₂ interface under UV-visible light irradiation (VB: valence band, CB: conduction band, OPs: oxidation products).

Conclusions

In summary, we have reported the synthesis of high-surface-area mesoporous networks of TiO₂ and Cu₂O NPs and demonstrated their high photocatalytic activity and stability for H₂ production under UV-visible light irradiation. Among the composites with different Cu loadings, the mesoporous CuMTA-3, which contains ∼1.5 wt% Cu, achieves an average hydrogen evolution rate of 542 μmol h⁻¹ with a 13.5% QE at λ = 365 ± 10 nm, while displaying good stability for at least 25 h. This particular catalyst outperforms mesoporous TiO₂ assemblies (MTA) by a factor of 135 under identical reaction conditions. Using UV-vis/NIR diffuse reflectance and photoluminescence (PL) spectroscopy, we showed that the superior HER activity of the CuMTA catalysts arises from the large separation of photogenerated electrons and holes in ternary Cu/Cu₂O/TiO₂ semiconductor. In addition to the electronic effects, the high porosity and small grain size of TiO₂ and Cu₂O crystals may also contribute to the higher photocatalytic activity of CuMTA materials. Moreover, by comparing the activity of two different ∼1.5 wt% Cu-loaded TiO₂ catalysts that were synthesized using distinct methods, i.e. chemical reduction deposition and photo-reduction, we also showed that the HER is controlled by the chemical nature of copper oxide species deposited on the surface of TiO₂; the result is that the Cu/TiO₂ catalyst is less active than the Cu₂O/TiO₂ catalyst for HER. Taken together, these findings suggest the great possibility for the implementation of this new photocatalyst system as a sustainable and low-cost alternative to hydrogen production from water.

Experimental

Materials

Brij-58 surfactant (HO(CH₂CH₂O)₂₀C₁₆H₃₃, M_n ∼ 1124), titanium tetrachloride (99.9%), titanium(IV)propoxide (≥98%) and absolute ethanol (99.8%) were purchased from Sigma-Aldrich. Cu(NO₃)₂·2H₂O (98%) was purchased from Alfa Aesar.

Synthesis of CuMTA catalysts

Mesoporous assemblies of TiO₂ NPs (MTA) were prepared through sol–gel polymerization between TiCl₄ and titanium(IV)propoxide (TPP) compounds in the presence of polyoxoethylene cetyl ether (Brij-58) block copolymer. Subsequent removal of the polymer template by calcination at 350 °C led to mesoporous structures composed of TiO₂ NPs.²⁷ Cu-loaded TiO₂ (CuMTA) catalysts with different content of Cu₂O were obtained by chemical reduction deposition method. Typically, 0.1 g of MTA was dispersed in 2 mL of copper(II) nitrate ethanol solution with a Cu content ranging from 0.5 to 2 wt% (on the basis of TiO₂). The resulting suspension was then slowly stirring at 40 °C until completely evaporation of solvent. Reduction of copper ions on the surface of MTA was carried out by suspending the Cu/TiO₂ materials into an aqueous solution containing 2 eq. of glucose and 3 eq. of NaOH, on the basis of Cu loading, and the resulting mixture was kept under stirring for 1 hour at 60 °C. The slight pink colored solid was then collected by filtration, washed with water and ethanol, and dried 60 °C for 12 h. A series of mesoporous CuMTA catalysts with different Cu content, i.e. 0.5, 1, 1.5 and 2 wt%, was prepared by varying the copper amount in the reaction mixture.

Photocatalytic study

The photocatalytic hydrogen evolution experiments were carried out in a gas-tight reaction cell using a 300 W Xe lamp (Variac Cermax) with an optical long-pass filter allowing λ between 360 and 780 nm. For visible-light photocatalytic reactions, an optical UV cut-off filter (λ > 420 nm) (Ophir Optronics Ltd) was used. In a typical reaction, 15 mg of the catalyst was suspended with vigorous stirring in 20 mL of aqueous solution containing 20% (v/v) methanol. The reaction mixture in the cell was cooled (20 ± 2 °C) by a continuous flow of water. Before irradiation, the suspension was purged with argon for at least 45 min to drive off the air inside. The concentration of the evolved H₂ as a function of irradiation time was measured by a gas chromatograph (Shimadzu GC-2014) equipped with a thermal conductivity detector and using Ar as the carrier gas.

The apparent quantum efficiency (QE = 2N_H2/N_ip, where N_H2 and N_ip are the number of evolved hydrogen molecules and number of incident photons) and the UV-visible photon conversion efficiency of the CuMTA-3 sample were determined by using a StarLite power meter equipped with a FL400A-BB-50 fan cooled thermal sensor (Ophir Optronics Ltd). The average intensity of incident light was measured to be 40 mW cm⁻² using a band-pass cutoff filter of λ = 365 ± 10 nm (Asahi Spectra, Japan) and the total incident power of irradiation was 0.51 W cm⁻² in the 360–780 nm interval.

Physical characterization

Scanning electron microscopy (SEM) imaging and elemental microprobe analysis were carried out using a JSM-6390LV (JEOL) SEM equipped with an Oxford INCA PantaFET-x3 energy dispersive X-ray spectroscopy (EDS) detector. Data acquisition was performed using an accelerating voltage of 20 kV and 60 s accumulation time. Transmission electron microscope (TEM) images were recorded on a JEOL JEM-2100 microscope with an acceleration voltage of 200 kV. Samples were prepared by dropping the ethanol suspension containing dispersed samples onto the carbon-coated copper grids. X-ray diffraction (XRD) patterns were collected on a Panalytical X'Pert Pro MPD powder diffractometer (45 kV and 40 mA) with Cu Kα radiation (λ = 1.5418 Å). Nitrogen adsorption and desorption isotherms were measured at 77 K on a NOVA 3200e volumetric analyzer (Quantachrome, USA). Before analysis, the samples were degassed at 100 °C under vacuum (<10⁻⁵ Torr) for 12 h to remove moisture. The specific surface areas were evaluated using the Brunauer–Emmett–Teller (BET) method⁵⁰ in the 0.05–0.24 relative pressure (P/P₀) range. The total pore volumes were derived from the adsorbed volume at the P/P₀ = 0.98, and the pore size distributions were calculated using the non-local density functional theory (NLDFT) method⁵¹ on the adsorption data. UV-vis/NIR diffuse reflectance spectra were obtained on a Perkin Elmer Lambda 950 optical spectrophotometer using BaSO₄ as reflectance standard. Reflectance data were converted to absorption (α/S) using the Kubelka–Munk equation:⁵² α/S = (1 − R)²/(2R), where R is the reflectance and α, S are the absorption and scattering coefficients, respectively. Photoluminescence (PL) spectra were obtained at room temperature on a Lumina Fluorescence Spectrometer (Thermo scientific) equipped with a 150 W xenon lamp and operated from 200 to 1000 nm.

Acknowledgements

This work was financially supported by the Greek Ministry of Education and the European Union under the ERC grant schemes (ERC-09) and the University of Crete – Special Account for Research (KA 3475).

Notes and references

1. T. Hisatomi, J. Kubota and K. Domen, Chem. Soc. Rev., 2014, 43, 7520.
2. T. Bak, J. Nowotny, M. Rekas and C. C. Sorrell, Int. J. Hydrogen Energy, 2002, 27, 991.
3. D. M. Fabian, S. Hu, N. Singh, F. A. Houle, T. Hisatomi, K. Domen, F. E. Osterlohf and S. Ardo, Energy Environ. Sci., 2015, 8, 2825.
4. A. Kudo and Y. Miseki, Chem. Soc. Rev., 2009, 38, 253.
5. Y. Quab and X. Duan, Chem. Soc. Rev., 2013, 42, 2568.
6. W. Fan, Q. Zhang and Y. Wang, Phys. Chem. Chem. Phys., 2013, 15, 2632.
7. L. M. Peter and K. G. U. Wijayantha, ChemPhysChem, 2014, 15, 1983.
8. F. E. Osterloh, Chem. Soc. Rev., 2013, 42, 2294.
9. M. R. Gholipour, C.-T. Dinh, F. Bélandb and T.-O. Do, Nanoscale, 2015, 7, 8187.
10. W. Chen, Q. Kuang, Q. Wang and Z. Xie, RSC Adv., 2015, 5, 20396.
11. D. Barreca, G. Carraro, V. Gombac, A. Gasparotto, C. Maccato, P. Fornasiero and E. Tondello, Adv. Funct. Mater., 2011, 21, 2611.
12. M. Murdoch, G. I. N. Waterhouse, M. A. Nadeem, J. B. Metson, M. A. Keane, R. F. Howe, J. Llorca and H. Idriss, Nat. Chem., 2011, 3, 489.
13. P. Bera and M. S. Hegde, RSC Adv., 2015, 5, 94949.
14. P. Zhang, T. Wang and J. Gong, Adv. Mater., 2015, 27, 5328.
15. L. Deng, S. Wang, D. Liu, B. Zhu, W. Huang, S. Wu and S. Zhang, Catal. Lett., 2009, 129, 513.
16. T. Montini, V. Gombac, L. Sordelli, J. J. Delgado, X. Chen, G. Adami and P. Fornasiero, ChemCatChem, 2011, 3, 574.
17. J. Ran, J. Zhang, J. Yu, M. Jaroniec and S. Z. Qiao, Chem. Soc. Rev., 2014, 43, 7787.
18. K. Zhao, S. Zhao, J. Qi, H. Yin, C. Gao, A. M. Khattak, Y. Wu, A. Iqbal, L. Wu, Y. Gao, R. Yu and Z. Tang, Inorg. Chem. Front., 2016, 3, 488.
19. Z. Xi, C. Li, L. Zhang, M. Xing and J. Zhang, Int. J. Hydrogen Energy, 2014, 39, 6345.
20. A. Devadoss, P. Sudhagar, C. Ravidhas, R. Hishinuma, C. Terashima, K. Nakata, T. Kondo, I. Shitanda, M. Yuasa and A. Fujishima, Phys. Chem. Chem. Phys., 2014, 16, 2137.
21. W. Yuan, J. Yuan, J. Xie and C. M. Li, ACS Appl. Mater. Interfaces, 2016, 8, 6082.
22. W. Fan, X. Yu, H.-C. Lu, H. Bai, C. Zhang and W. Shi, Appl. Catal., B, 2016, 181, 7.
23. D. P. Kumar, N. L. Reddy, B. Srinivas, V. Durgakumari, V. Roddatis, O. Bondarchuk, M. Karthik, Y. Ikuma and M. V. Shankar, Sol. Energy Mater. Sol. Cells, 2016, 146, 63.
24. D. Barreca, G. Carraro, E. Comini, A. Gasparotto, C. Maccato, C. Sada, G. Sberveglieri and E. Tondello, J. Phys. Chem. C, 2011, 115, 10510.
25. D. P. Kumar, M. V. Shankar, M. M. Kumari, G. Sadanandam, B. Srinivas and V. Durgakumari, Chem. Commun., 2013, 49, 9443.
26. M. Wang, L. Sun, Z. Lin, J. Cai, K. Xie and C. Lin, Energy Environ. Sci., 2013, 6, 1211.
27. I. Tamiolakis, I. N. Lykakis, A. P. Katsoulidis and G. S. Armatas, Chem. Commun., 2012, 48, 6687.
28. J. B. Joo, M. Dahl, N. Li, F. Zaera and Y. Yin, Energy Environ. Sci., 2013, 6, 2082.
29. Y. Q. Cao, Y. Georgievskii and R. A. Marcus, J. Chem. Phys., 2000, 112, 3358.
30. P. Scherrer, Gottinger Nachr. Math.-Phys., 1918, 2, 98.
31. F. Rouquerol, J. Rouquerol and K. S. W. Sing, Adsorption by Powders and Porous Solids. Principles Methodology and Applications, Academic Press, London, 1999.
32. P. E. de Jongh, D. Vanmaekelbergh and J. J. Kelly, Chem. Commun., 1999, 1069.
33. H. Tada, Q. Jin, H. Nishijima, H. Yamamoto, M. Fujishima, S.-I. Okuoka, T. Hattori, Y. Sumida and H. Kobayashi, Angew. Chem., Int. Ed., 2011, 50, 3501.
34. S. Zhang, S. Zhang, F. Peng, H. Zhang, H. Liu and H. Zhao, Electrochem. Commun., 2011, 13, 861.
35. Y. Lai, L. Sun, Y. Chen, H. Zhuang, C. Lin and J. W. Chin, J. Electrochem. Soc., 2006, 153, 123.
36. P. D. Tran, S. K. Batabyal, S. S. Pramana, J. Barber, L. H. Wong and S. C. J. Loo, Nanoscale, 2012, 4, 3875.
37. S. G. Babu, R. Vinoth, D. P. Kumar, M. V. Shankar, H.-L. Chou, K. Vinodgopal and B. Neppolian, Nanoscale, 2015, 7, 7849.
38. W. Jian Foo, C. Zhangb and G. W. Ho, Nanoscale, 2013, 5, 759.
39. K. Lalitha, G. Sadanandam, V. D. Kumari, M. Subrahmanyam, B. Sreedhar and N. Hebalkar, J. Phys. Chem. C, 2010, 114, 22181.
40. L. Sinatra, A. P. LaGrow, W. Peng, A. R. Kirmani, A. Amassian, H. Idriss and O. M. Bakr, J. Catal., 2015, 322, 109.
41. I. Mondal and U. Pal, Phys. Chem. Chem. Phys., 2016, 18, 4780.
42. X. Li, J. Yu, J. Low, Y. Fang, J. Xiao and X. Chen, J. Mater. Chem. A, 2015, 3, 2485.
43. J. W. Ager, M. R. Shaner, K. A. Walczak, I. D. Sharpae and S. Ardo, Energy Environ. Sci., 2015, 8, 2811.
44. G. H. Chan, J. Zhao, E. M. Hicks, G. C. Schatz and R. P. Van Duyne, Nano Lett., 2007, 7, 1947.
45. H. Guo, Y. Chen, M. B. Cortie, X. Liu, Q. Xie, X. Wang and D.-L. Peng, J. Phys. Chem. C, 2014, 118, 9801.
46. T. Nakamura, Y. Tsukahara, T. Sakata, H. Mori, Y. Kanbe, H. Bessho and Y. Wada, Bull. Chem. Soc. Jpn., 2007, 80, 224.
47. F. J. Knorr, C. C. Mercado and J. L. McHale, J. Phys. Chem. C, 2008, 112, 12786.
48. M. Junga, J. N. Hartb, J. Scotta, Y. H. Nga, Y. Jianga and R. Amal, Appl. Catal., A, 2016 DOI:10.1016/j.apcata.2015.11.013.
49. J. Xing, Z. P. Chen, F. Y. Xiao, X. Y. Ma, C. Z. Wen, Z. Li and H. G. Yang, Chem.–Asian J., 2013, 8, 1265.
50. S. Brunauer, L. S. Deming, W. S. Deming and E. Teller, J. Am. Chem. Soc., 1940, 62, 1723.
51. P. I. Ravikovitch, D. Wei, W. T. Chueh, G. L. Haller and A. V. Neimark, J. Phys. Chem. B, 1997, 101, 3671.
52. P. Kubelka and F. Munk, Z. Tech. Phys., 1931, 12, 593.

---

Footnote

† Electronic supplementary information (ESI) available: SEM images, XRD patterns and catalytic data of CuMTA samples, N₂ physisorption isotherms of MTA, EDS, N₂ physisorption and UV-vis/NIR data of the regenerated CuMTA-3 catalyst, PL spectra of MTA and CuMTA-3, N₂ physisorption isotherms, TEM images and UV-vis/NIR spectrum of CuMTA-3P. See DOI: 10.1039/c6ra08546f

---