Utilizing nature's endowment: artificial leaf concept for methane activation to C–C coupled ethanol or ethylene

Abstract

Methane activation (MA) to platform chemicals under ambient conditions still remains an open challenge to be fully realised. The present work shows the fabrication of CeVO₄ quantum dots (CV-QDs) by a bottom-up approach; they are assembled from Ce³⁺ and metavanadate ions, and structurally and electronically integrated into the micro-/meso-pores of TiO₂ (CV-QD-TiO₂ (CVT)), demonstrating the conversion of MA to ethanol/ethylene by visible light-driven photocatalysis. CV-QDs in confined pores modify the quantum confinement effects and are characterized by physicochemical methods. The current synthetic strategy is potentially scalable and results in sub-quadrillion heterojunctions in a 1 mg CVT photoanode spread over 1 cm². MA with CVT under one-sun conditions demonstrates ∼100% selectivity to ethanol, yielding 4.36 μmol h⁻¹ cm⁻², with a solar-to-fuel efficiency (STFE) of 0.56. Further, by employing a co-catalyst, significant STFE (5.08) and yield (39.5 μmol h⁻¹ cm⁻²) are achieved selectively towards ethylene. A deliberate addition of methanol increases the rate of ethanol production by 17.2 times, indicating that the methyl-methoxy interaction is the origin of C–C coupling. Weight is normalized to a gram of CV-QDs in a large area CVT photoanode to yield 109 mmol h⁻¹ g_CV-QD⁻¹ of ethanol and 988 mmol h⁻¹ g_CV-QD⁻¹ of ethylene. Enhanced activity and selectivity towards the C₂-product is attributed to band-edge modulation and trillions of heterojunctions, which in turn facilitate charge separation and charge transfer for effective charge utilisation at redox sites.

---

1 Introduction

Methane stands out as one of the prime sources of voluminous energy owing to its stable C–H bonds.¹ Its tetrahedral structure renders chemical inertness, making its activation perplexing and highly challenging.² Increasing methane level, which is the second largest pollutant in the atmosphere, has accelerated over recent years, alarming the researchers. Its direct impact on global warming has significantly affected global temperature.³ Methane activation (MA) to platform chemicals, like syngas, formaldehyde, and ethanol, has been achieved by conventional heterogeneous catalysis; however, the entire process requires harsh conditions, which in turn adds to CO₂ pollution.^4,5 Currently, simple alcohol/alkene, like ethanol/ethylene, has emerged as cornerstone platform chemicals in industrial catalysis. It is imperative to develop a methodology to utilize methane for the direct production of value-added products, such as ethanol and ethylene, under ambient conditions. Though the currently practised methods of ethanol production offer significant industrial output, the yields are still far from meeting the needs, considering the input energy in terms of operational conditions and expenses.

An ethanol yield of 213 μmol g⁻¹ h⁻¹ (86% selectivity) from MA via heterogeneous catalysis using Ce_0.76Zr_0.20Co_0.02V_0.02O₂ is probably the highest reported, achieved without the use of any noble metals.⁶ Song et al. reported 25 μmol g⁻¹ h⁻¹ of ethanol yield with NiO at an applied potential of 1.40 V (vs. RHE), in which an NiO–Ni interface was created by tactfully calcining an Ni-foam.⁷ Li et al. reported 9.09 mmol g⁻¹ h⁻¹ of ethanol yield with Fe–Ni–OH catalyst at an applied potential of 1.46 V (vs. RHE), along with acetone and formate as minor products.⁸ Such catalysts either operate at low current densities or in the oxygen evolution reaction (OER) regime even at ∼20 mA cm⁻²; this leads to operational difficulties while decreasing the faradaic efficiency. Selective MA (SMA) in sunlight by photocatalysis is a potential and promising alternative method that occurs under ambient conditions.^9–11 The light-driven chemistry through charge carrier utilization has fostered the initiation and sustainability of reactions that have perplexing reaction mechanisms. After light absorption by the photocatalyst and generation of excitons, the nanosecond clock starts ticking; a maximum of a few ns is left to utilize this energy for chemical conversion. Hence, the knowledge of efficient generation and separation of charge carriers, along with their effective utilization at redox sites, holds the key to improving the efficiency of the photocatalysis of SMA and other demanding reactions.¹²

Amongst a handful of literature reports on SMA, Du et al. observed 51 μmol g⁻¹ h⁻¹ of ethanol yield under one sun illumination in the presence of water and O₂ over P-doped g-C₃N₄ (PCN).¹³ Zhou et al. valorised methane oxidation under anaerobic conditions, yielding 106 μmol g⁻¹ h⁻¹ of ethanol over Cu-modified PCN under a 500 W Xe-lamp.¹⁴ The above reports signify the wide scope of increasing efficiency through photocatalysis by introducing better-quality heterojunctions and valence band (VB) energy modulation. Photocatalysis proves worthy, even with the use of noble-metal-based catalysts, if the activity can be improved further. This is the gap area, which should be addressed aptly by following artificial leaf (AL) (or artificial photosynthesis) concept-based devices. The critical aspect of green photosynthesis is the integration of a number of components to convert water and CO₂ to glucose by absorbing visible light photons in direct sunlight. In contrast to the employment of organic and inorganic components in green photosynthesis, the integration of all inorganic components in the AL device takes the centre stage, especially to meet the sustainability and scalability aspects. Any successful AL attempt would include the major aspects of efficient light absorption from a wide wavelength regime of sunlight, efficient separation and utilization of charge carriers to the desired product. The electronic and structural integration of various material components is the key to achieving functions, like charge carrier separation and charge diffusion to the redox sites, while minimizing charge recombination. Finally, if the AL device preparation can be simplified, fast progress can be expected in this area.

The current work presents a coherent methodology for SMA to C–C coupled products (ethanol/ethylene) by methodically strategizing the engineering aspects of photocatalyst design.^13,14 Herein, an integrated photoanode system is developed, which entails the incorporation of visible-light absorbing CeVO₄ quantum dots (CV-QDs) in the pores of commercial P25-TiO₂. CV-QDs are strategically grown by the SILAR (successive ionic layer adsorption and reaction) method while restricting them in the nano-pores of TiO₂.^15,16 Hence, this strategy allows high temperature (723 K) calcination, while an indispensable surfactant in conventional QD synthesis is completely avoided for stabilization. The above strategy also helps in abundant and seamless heterojunction formation between different-size CV-QDs and TiO₂ (CeVO₄@TiO₂; CVT photoanode) along with potential gradient band-edge modulation.¹⁷ This current study demonstrates SMA to ethanol or ethylene (with a co-catalyst) under visible light irradiation under ambient conditions, with a yield of 4.36 and 39.5 μmol mg_cat⁻¹ h⁻¹, respectively, which far exceeds that of earlier reports. The integrated CVT system and synthetic strategy reported herein opens up a broader scope for the ambient activation of methane and possibly other inert molecules.

2 Results and discussion

2.1 Chemicals required

Cerium(III) nitrate hexahydrate, ammonium metavanadate, titanium dioxide (Degussa P25), glacial acetic acid, ethanol, terpineol, ethylcellulose, titanium chloride, and FTO-coated glass plates were purchased from Sigma.

2.2 Doctor-blade method

The following synthetic procedure was utilised to assemble and integrate CV-QDs into the micro and mesopores present in Degussa TiO₂.¹⁸ First, the TiO₂ paste was made, wherein 1 g of TiO₂ powder was placed in a round bottom flask, to which 33 mL of ethanol and 0.33 mL of glacial acetic acid were added. After thorough dispersion of TiO₂ in the solvent mixture by 10 minutes of continuous stirring followed by 10 minutes of sonication, 0.5 g of ethyl cellulose was added as a binder, followed by stirring and sonication for 10 minutes each. To provide viscosity to the solution mixture, 3 mL of terpineol was added, and the solution mixture was stirred for 30, followed by 30 minutes of sonication. The solution was evaporated in a rota-vapour at 323 K to remove the solvents and obtain the desired viscous thick paste. The paste obtained was coated over carefully cleaned (with isopropyl alcohol, followed by TiCl₃ treatment on the conductive side) FTO plates of 1 × 1 cm² area. The TiO₂ film was uniformly spread with a thickness between 9 and 10.5 μm and measured by a surface profilometer for confirmation. The films were finally dried at 333 K for 2 h, followed by calcination at 723 K for 30 minutes.

2.3 Fabricating the binary CeVO₄@TiO₂ (CVT) semiconductor photocatalyst

The SILAR method¹⁹ was employed to assemble the Ce³⁺ and metavanadate (VO₃⁻) ions into the pores of TiO₂ for the synthesis of the integrated CVT system. Ce(NO₃)₃·6H₂O and NH₄VO₃ were used as the cationic and anionic precursors, respectively. 40 mL of 25 mM solutions of each were placed in separate beakers and kept at 348 K. The as-prepared TiO₂ films were dipped into the solutions one after the other for 30 seconds each, followed by washing in deionized water. This is considered one SILAR cycle. This was repeated for ten SILAR cycles for the best SMA performance. The films were then dried in an oven at 333 K for 1 h. It is reiterated that precursor ions are simply assembled by applying the SILAR process, and the CV structure is yet to form. After drying in an oven, the above film was subjected to static air-calcination at 723 K for 4 h at a ramping rate of 5 deg. min⁻¹ in a muffle furnace. This is the critical step leading to the formation of CV-QDs in the TiO₂ pores, as well as their integration with TiO₂ into the final photoanode device. A synthesis scheme is shown in Fig. 1a. After optimizing the number of SILAR cycles to 10, it was observed that 1 cm² of the area holds 1 ± 0.1 mg of the integrated photocatalyst, i.e., CeVO₄@TiO₂ (CVT). For all calculations, it is assumed that the weight of the CVT is 1 mg cm⁻². The photocatalytic SMA was evaluated under one-sun conditions in a quartz reactor, and the set-up is displayed in Fig. S1.

Fig. 1 (a) Schematic of the CVT photoanode fabrication by the SILAR method. n = number of SILAR cycles. (b) X-ray diffraction (XRD) stack plot for CVT (pink) along with bulk CV (dark yellow) and TiO₂ (blue). Reference XRD pattern for CeVO₄ JCPDS No. 79-1065 (green) is displayed at the bottom. Green and black asterisks denote CeVO₄ and anatase TiO₂ features, respectively. The FTO features are denoted with #. (c) Enlarged area depicting the shift in peak positions induced by CV-QD integration with TiO₂. Also note the absence of few diffraction features of bulk CV in CVT. The observed shift is denoted by solid and dash-dot black arrows.

2.4 Material characterisation

The integrated binary CVT photocatalyst was thoroughly characterized to study its structural and morphological properties. Diffuse reflectance UV-Vis measurement was carried out using a Shimadzu spectrophotometer (model UV2550) and converted into absorbance. Powder XRD data were collected using a Pan Analytic X'Pert Pro dual goniometer diffractometer with Cu K_α (1.5418 Å) radiation and Ni filter. Microstructural analysis was carried out using a JEOL JEM F-200 HRTEM operating at 200 kV. Nitrogen adsorption–desorption isotherm measurements at 77 K were carried out using a Quantachrome Quadrasorb automated gas sorption system to obtain textural properties. Electronic structure aspects were explored using X-ray photoelectron spectroscopy (XPS) (Thermo Fisher Scientific Instruments), equipped with an Al K_α (monochromatic) X-ray radiation source and operated at a 6 mA beam current and 12 kV. Energy dispersive X-ray spectral (EDS) analysis was carried out to explore the weight percentage of CV-QDs incorporated into the CVT. Further, the weight difference was carefully measured from the initial TiO₂ thin film to the CVT photoanode, and it is in good correlation with the EDS findings. Electron paramagnetic resonance (EPR) spectra were obtained using a Bruker A300 X-band spectrometer. The data were collected at a scanning power of 0.998 mW at 298 K. A TechnoS IndiRAM CTR 500C Micro Raman spectrometer was used to derive the Raman features with an excitation laser wavelength of 532 nm. The electrochemical measurements herein were carried out using a Gamry 3000 potentiostat. The IPCE measurements were carried out using a Newport IPCE system. The physical and chemical properties measured were reproduced from at least three different batches of CVT photoanodes.

2.5 Selective photo-oxidation of methane to ethanol and ethylene

The CVT photoanode device was investigated for benevolent utilization of the photon energy to initiate the activation of the stable methane molecule and steer the C–C coupling under ambient conditions. A photocatalytic methane oxidation reaction was carried out using the CVT device with a photocatalyst coated on a 1 × 1 cm² area. The photocatalytic device was placed inside the carefully cleaned custom-built photoreactor containing 10 mL of deionized water at pH = 7 ± 0.2. It was sealed to ensure no leakage. The said photoreactor was made of quartz with a flat surface that can expose the photocatalytic device to solar irradiation, as shown in Fig. S1. Initially, the photoreactor was placed in an ice bath (∼2–4 °C) and purged with methane (99.999% purity) for 30 min to saturation. The low temperature facilitates a higher solubility of methane in water (23 mg L⁻¹). The reactor was closed airtight using a rubber septum and exposed to one sun solar illumination for 3 h under ambient conditions. Photocatalytic oxidation was evaluated by analyzing the liquid and gaseous samples collected at steady intervals using a leak-tight syringe. The formation of liquid products was identified and quantified using ¹H-NMR spectroscopy (Bruker BioSpin AG; 400 and/or 500 MHz; AV-Neo 400 and 500). Similar steps were followed to evaluate the photo-oxidation study under direct sunlight. CVT-Pt was also employed in a similar manner, except for gas-phase product detection by GC. Gaseous products were analysed by an Agilent GC system equipped with a TCD detector and a Carbosieve S II column, with He as the carrier gas for CH₄, CO, and CO₂; FID detector with an HP-PLOT/Q column was employed for ethane and ethylene identification. All photocatalysis experiments were repeated at least three times, with photoanodes prepared from different batches, and the error values were observed to be within ±10%.

2.6 Liquid product quantification from ¹H-NMR

To quantify the liquid products, 0.5 mL of the post-reaction sample is taken and mixed well in an NMR tube containing 0.020 mL of KHP as an internal standard (corresponding to 1 mM in an NMR solution) and 0.080 mL of D₂O. The concentration of the products was determined using the following formula:
where n represents the molar concentration, I represents the integral area in the ¹H-NMR spectra, and N represents the number of protons/nuclei associated with the peak signal. x and y represent the internal standard and the liquid product, respectively.

2.7 Numeric depiction of photocatalytic efficiency

2.7.1 Solar-to-fuel efficiency. The solar to fuel (STF) conversion efficiency for the photocatalytic system is calculated under one-sun conditions over a 1 cm² area. It is noteworthy that a 1 cm² thin film contains around 40 μg of CeVO₄. The ethanol yield is calculated to be 4.36 μmol h⁻¹. The STF is calculated as follows:

where of ethanol is 234 kJ mol⁻¹ and P is the average power density of one-sun conditions (100 mW cm⁻²) over a 1 cm² area. The STF values were calculated to be 0.28% and 0.1% in one-sun conditions and direct sunlight, respectively. As ethylene formation occurs via ethanol generation and further dehydration, the STF is multiplied 9.07 times as per the activity observed (Table 2). It is also to be indicated that there are two methane molecules to be activated for the generation of one ethanol/ethylene molecule. Hence, STF in terms of methane activation (to fuel) is multiplied by a factor of 2; for ethanol/ethylene, it is 0.56%/5.08%.

2.7.2 Apparent quantum efficiency. The AQE for the reaction was calculated using a Newport light source (300 W Xe lamp) equipped with a Newport 74125 monochromator and a Newport-1918-R power meter to measure monochromatic light and incident light intensity, respectively.

The number of incident photons is calculated as . Here, the power of the lamp is 400 W cm⁻², and the wavelength used is 550 nm; h is Planck's constant, and c is the speed of light. The yield of ethanol observed is 1 μmol h⁻¹ cm⁻².

2.7.3 Turn over frequency. ToF is an important factor that defines the sustainability of the photo-catalyst. This is determined by the ratio of moles of the product formed to the moles of the active catalyst. For comparative purposes, we consider CV-QDs to be active sites. 40 μg of CeVO₄ corresponds to 0.196 × 10⁻⁶ moles. Considering a 4.36 μmol h⁻¹ ethanol yield, the TOF is as follows:

3 Results and discussion

3.1 Textural properties and microscopy analysis

Fig. 1b and c displays the regular (enlarged) XRD patterns of CVT along with pristine tetragonal CeVO₄ and TiO₂. All the features are indexed, and profound changes are observed in the CVT pattern. The narrow (200) facet of CV (24.26°) broadens and shifts to 23.86° in CVT. Similarly, the predominant TiO₂ (101) feature at 25.16° shifts to 25.05° in CVT. It is also to be noted that three intense diffraction features of CV (37.9, 51.7 and 65.6°) show negligible intensity after integration in CVT. The ratio of (200) (CV in CVT) to (101) (TiO₂ in CVT) is 0.984, while that of (112) (CV in CVT) to (101) (TiO₂ in CVT) is 0.835. This suggests that (200) is the predominant facet of CV-QDs that interacts strongly with TiO₂ in CVT and thus could serve as the active facet, followed by (112). The shift to a lower angle suggests a significant increase in interplanar distances and an elongation of the associated bond lengths. A shift in diffraction features underscores the structural integration (SI) of TiO₂ and CeVO₄ components into one system. Although TiO₂ and CV amounts are 96 and 4 weight percent, respectively, the parent features from TiO₂ (101) and CeVO₄ (200) show comparable intensities in the integrated CVT, supporting the SI through selected facets of CV with TiO₂.

Fig. 2a–f shows TEM and HRTEM images and provides a nanoscale view of the CVT. Fig. S2 shows more such images. These results demonstrate the incorporation of CV-QDs (indicated with yellow arrows and purple circles) of ≤5 nm in the nanopores of TiO₂. Fig. 2c shows the zoomed-in view of a single CV-QD in TiO₂ particle, while Fig. 2a and S2a–c shows many such particles. The interface of CV and TiO₂ forms a heterojunction, and they are known to promote charge separation efficiency. A careful look at the images shows the presence of two types of QDs. Although most of the QDs are integrated into the pores, a few are semi-incorporated and can be observed embedded on the surface (Fig. 2c (green circle) and Fig. 2d (right-most arrow)). The d-spacing of CV-QDs conveys this vividly. The d-spacing of QDs found on the surface was measured to be 0.314 nm (Fig. 2e and f), corresponding to the (112) facet of CeVO₄, while TiO₂ shows d = 0.345 nm, corresponding to the anatase (101) facet. Although the areas where the CV-QDs are predominantly found in deep TiO₂ pores, no distinct d-spacing could be calculated. The CV-QDs found just beneath the surface exhibit an intersection of the lattice fringes of TiO₂ and CV (Fig. 2c–e, purple encircled areas). Fig. S2d and e illustrate the intersection of such lattice fringes in CVT, fully supporting the presence of nano-heterojunctions. It is reiterated that any CV-QD not integrated with TiO₂ would easily agglomerate under calcination (723 K), which is not observed in the present case, underscoring the integrated aspect of the CVT photoanode.

Fig. 2 (a, b) TEM and (c–e) HRTEM images of the CVT photocatalyst at different magnifications, showing the formation of CV-QDs in the pores of TiO₂. One CV-QD inside the pore and another one partly embedded are indicated by purple and green circles, respectively, in (c). Lacey grid support is on the right-most side of panel (a and b). (f) d-spacing of the lattice fringes calculated for CV. (g–k) STEM images for the line profile analysis, and the elemental distribution of CV-QDs in the TiO₂ pores. Distribution of individual elements of CVT, namely, (h) Ti, (i) O, (j) Ce, and (k) V. Ce and V, concentrated in the mesopores, are encircled. (l) Contribution of different elements to the line profile.

CV-QDs confined in the micro/mesopores of TiO₂ are evident from the scanning transmission electron microscopy (STEM), and line profile analysis (LPA) results are shown in Fig. 2g–l. A careful analysis of the encircled areas (Fig. 2j and k) indicates the presence of a group of particles, but they are neither connected nor agglomerated, even after calcination at 723 K. This is mainly due to the containment of CV-QDs in TiO₂ pores, and TiO₂ walls prevent them from agglomerating. LPA reveals a similar intensity pattern variation for both Ce and V, while it is high for Ti. Even within LPA, the presence of CV-QDs of different sizes is evident in the count pattern. Apart from the Ce and V concentrated (big circle) areas, a scarce distribution of Ce and V is also observed. This suggests the filling of smaller micropores, too, with CV-QDs. HRTEM shows evidence of this. The uniform distribution of the CV-QDs throughout the film thickness is further confirmed by FE-SEM cross-sectional analysis (Fig. S3). EDS analysis was carried out for CVT prepared in different batches. A representative result is displayed for the elemental composition in Fig. S4. EDS results demonstrate that the atomic ratio of Ce : V is close to 1 : 1, and the weight of CV-QDs is 4.03%, which is integrated in the pores of TiO₂ (∼96%).

N₂ adsorption–desorption isotherms recorded for TiO₂ and CVT appear to be similar (Fig. S5). Type-IV adsorption isotherm with H1 hysteresis-loop indicates a narrow and close-size range of mesopores. A decrease in surface area (and pore volume) from 57.0 (0.22) for TiO₂ to 41.9 m² g⁻¹ (0.14 cm³ g⁻¹) for CVT was observed, which is attributed to the restraint of CV-QDs in the pores. Pore volume reduction occurs predominantly due to the filling of mesopores. Assuming a 1 : 4 weight ratio of CV-QDs occupying 2.5 and 4 nm pores, respectively, a calculation is made (SI S1 and note below Fig. S5) to determine the number of heterojunctions.^19,20 An estimated total number of CV-QD-TiO₂ heterojunctions formed is 293 trillion in 1 mg CVT material. When pore size is varied between 1 and 5 nm, it results in a similar number from ∼150 trillion to 1 quadrillion.

3.2 Optical, Raman and XPS spectral studies

Fig. 3a shows the UV-Vis absorption spectra recorded for TiO₂, CeVO₄, and CVT. CeVO₄ shows a weak absorption onset at ∼680 nm.^21–23 Nonetheless, CVT exhibits a distinct absorption onset at ∼720 nm with a significantly increased extinction coefficient and main absorption feature around 470 nm. The inset in Fig. 3a demonstrates a gradually increasing extinction coefficient and a simultaneous red shift in absorption onset due to increasing CV-QD content in TiO₂. Interestingly and concurrently, a blue-shift observed with the main absorption feature ≤460 nm, compared to CV, underscores an increase in band-gap (E_g) of CV-QDs (from Tauc plot; Fig. 3b) due to quantum confinement. However, 5 nm CV nanoparticles are also observed in TiO₂ mesopores. Hence, CVT absorption is extended and is similar to bulk CV (∼500 nm). The changes observed above reflect the colour of CV (pale yellow-brown) to CVT (dark brown) (Fig. 3b). Due to the polydispersity of CV-QDs, light absorption is efficient for the entire range of visible light. Due to CV structure formation on calcination and hence volume expansion, electronic and structural integration (E&SI) occurs necessarily at the interface of CV-QDs and TiO₂, with a potential gradient across the interface, which depends on the size and crystallographic facet of CV-QDs. Besides, an indirect E_g displayed in Fig. S6 appears in CVT that evolved with a higher number of SILAR cycles, showing a value of 1.7 eV. The combined understanding derived from the optical absorption studies of the integrated CVT photocatalyst thus underlines the quantum confinement of CV-QDs in the TiO₂ nanopores that encapsulate the properties intermediate between the bulk semiconductor and a single unit cell. Additionally, a careful analysis of CV and CVT reveals an extended absorption by the latter (∼720 nm) in spite of the QD size of CV in the pores of TiO₂. Extended absorption is attributed to the possible inter-band transition between CV and TiO₂ in CVT. This is discernible in the Tauc plot. This further enhances charge transfer ability by lowering the charge recombination rate, as observed in the PL spectra (Fig. S7). The CVT photocatalyst hereby shows a notable decrease in the charge recombination rate upon the introduction of CV-QDs into the pores of TiO₂. The critical issue in photocatalysis is the lack of timely utilization of the photogenerated charge carriers. The creation of abundant heterojunctions provides the right handle to efficiently separate electrons and holes; hence, their utilization increases. Although electrons are transferred to TiO₂, holes are retained in the CV-QDs, facilitating oxidation reactions.

Fig. 3 a) UV-Vis spectra of various photocatalysts; inset spectra show the evolution of absorption pattern with increasing number of SILAR cycles from 3 (CVT_3CYC) to 10 (CVT_10CYC). A parallel increase in the absorption onset to higher λ and the interface interaction are observed up to 10 SILAR cycles. (b) Tauc plot displaying an increase in the band-gap from CV to CVT and indicating the containment of CV-QDs in TiO₂ pores. Inset shows the change in colour associated with CV (pale brown) to CVT (dark brown), although the latter contains a mere 4 wt% CV-QDs in TiO₂. (c) Comparison of Raman spectra of TiO₂ and CVT; in spite of 4 wt% of CV-QDs, the significant increase in the intensity of the Raman features of CV underscores its high crystallinity.

The Raman spectra recorded for TiO₂ and CVT are shown in Fig. 3c. The black arrows display peaks typical (144, 399, 515, and 640 cm⁻¹) of anatase TiO₂, and CV features are indicated by small pink arrows. The strongest vibrational feature of CeVO₄ at 865 cm⁻¹ depicts the intra-tetrahedral V–O bonds of the symmetric stretching mode of A_1g along with asymmetric stretching and bending modes at 795 and 466 cm⁻¹, respectively. The bending mode of the VO₄ tetrahedron is observed at 265 cm⁻¹.²³ Raman features support the tetrahedral zircon-type CV-QDs integrated with the TiO₂ framework. Although CV-QD content is only 4 wt% in CVT, the observed narrow and strong Raman features underscore its high crystallinity in correspondence with the XRD results. It is noteworthy that the E_g feature of CV at 380 nm overlaps with the B_1g feature of TiO₂ at 399 cm⁻¹; hence, a broadening is observed with CVT.²⁴ TiO₂ features are the same in CVT and TiO₂.

XPS was employed to understand the changes in the oxidation state and E&SI aspects of CVT; deconvoluted XPS results are shown in Fig. 4a–c. Ti 2p_3/2 shows a major amount of Ti⁴⁺ (458.5 eV) with a small amount of Ti³⁺ (457.4 eV) (Fig. 4a). However, Ti³⁺ content increases from CVT to the post-reaction catalyst (CVT-P). The oxygen vacancy (O_v) feature indicates the oxygen species around O_v, which is observed with significantly different electron densities. This feature also increases from CVT to CVT-P (Fig. S8). However, Ce³⁺ was exclusively observed in CV, CVT and CVT-P catalysts (Fig. 4b and SI S2). Predominant V⁵⁺ and minor V⁴⁺ states are observed in CVT; in fact, an increase in V⁴⁺ content is observed in CVT-P (Fig. 4c). Concurrent increase in O_v, V⁴⁺ and Ti³⁺ contents corroborates well from CVT to CVT-P.

Fig. 4 XPS core level spectra of (a) Ti 2p, (b) Ce 3d, and (c) V 2p collected from TiO₂, CVT, CVT-P and CV. CVT-P denotes the post-reaction CVT catalyst. (d) X-ray valence band spectra of the photocatalysts, showing the valence band (VB) features. Note that the Ce 4f¹ feature from CVT shifts toward E_F, indicating the highly reactive nature of VB.

The X-ray VB (XVB) results are depicted in Fig. 4d. It gives the VB-Maximum (VB_Max) energy that helps in the direct elucidation of CB-Minimum (CB_Min) and hence the electron transfer mechanism for ensuing photocatalysis. Interestingly, VB_Max-CV and VB_Max-CVT are observed at the same energy, 0.75 eV, demonstrating a vivid shift (by 2.21 eV) from TiO₂, which appears at 2.96 eV. The CB_Min positions are calculated using E_g from the Tauc and XVB plots, as listed in Table 1. The CB_Min-CVT calculated to be at −1.90 eV, at a high potential, facilitates the injection of photoexcited electrons from CV-QDs into TiO₂ in the CVT, thereby demonstrating the type-II heterojunctions. Remarkably, the Ce-4f band shifts from 2.45 eV (CV)²⁵ to 2.0 eV with CVT, underscoring the highly reactive nature of Ce in CVT. The above findings demonstrate the rich E&SI of CV-QDs with TiO₂.

Table 1 Measured VB and CB edge positions from the XVB and E_g values

Photocatalyst	Band gap (Eg)/eV	VB maximum (VBMax)/eV	CB minimum (CBMin)/eV
TiO2	3.13	2.96	−0.17
CV	2.5	0.75	−1.75
CVT	2.65	0.75	−1.90

---

3.3 Photoelectric response

The chronoamperometry results are shown in Fig. 5a, with an enhanced current density (38 μA cm⁻²) for CVT, which is much higher than that of (TiO₂) CV. Enhanced photocurrent generation demonstrates improved charge separation at heterojunctions in CVT and their diffusion to the electrodes despite only 4 wt% of CV-QDs in CVT.²⁶ This observation underscores the efficient light absorption and pathways available for efficient electron transfer from the source point to the electrode or redox sites on the photoanode. Critically, the potential gradient between CV-QDs and TiO₂ at the interface facilitates charge separation.

Fig. 5 (a) Chronoamperometry measurement carried out at 0.4 V under one-sun conditions to show transient photocurrent response for light on and off conditions, periodically, for CV, CVT and TiO₂. (b) Wavelength-dependent IPCE performance measured at zero applied potential, and the inset shows the integrated current density over the wavelengths of visible light range.

Wavelength-dependent IPCE (incident photon to current efficiency) measurements were carried out at 298 K (Fig. 5b). Expectedly, TiO₂ shows poor IPCE response in visible light. An enhanced IPCE over the entire visible light range is observed with CVT compared to CV. Bulk CV shows 24% IPCE at λ = 420 nm, which increases to 58% with CVT, despite a mere 4 wt% CV in it. Inset in Fig. 5b shows the integrated current density as a function of wavelength; a 3-times higher current density observed with CVT than CV underscores an efficient light absorption and charge separation. An observation worth underscoring is that the significant photocurrent observed is ≥600 nm. This depicts the efficient absorption of longer wavelength photons (red) for CVT and CV. Note that an exceptional increase in IPCE performance was observed between 400 and 460 nm, and a 27% value even at ≥600 nm. It is noteworthy that CV contains 100% CeVO₄, while CVT contains only 4 wt% CV-QDs in TiO₂. An unusual increase in current density is observed ≤460 nm with CVT. It is known from earlier results that CV-QDs are integrated in the TiO₂ pores, which increases the E_g of the former. It is known that the E_g of CV-QDs increases with decreasing size, and an enhancement in IPCE between 400 and 460 nm underscores the possibility of the integration of smaller CV-QDs (≤2 nm) in the micropores. The high current generation between 400 and 460 nm demonstrates an efficient electron injection at a significantly higher photovoltage from CV-QDs into TiO₂. These facts suggest that an increase in the extinction coefficient between 460 and 400 nm increases IPCE performance effectively. An electrochemical impedance spectral (EIS) study was carried out for further support. The Nyquist plot obtained for CVT and bulk CV depicts a smaller arc radius than TiO₂ (Fig. S9). However, between CV and CVT, the latter shows a smaller arc, underscoring efficient charge separation and supporting low charge transfer resistance.^22,27

A few points are worth highlighting. First, a higher IPCE with a smaller CV-QD size accentuates the filling of CV-QDs in the micropores and small mesopores of TiO₂. This is in agreement with the textural properties alongside improved visible light absorption (Fig. 2 and 3a). Second, the cumulative IPCE current density over 410–430 nm is significantly higher than that of 470–510 nm. Thus, a shift in the CB potential of smaller CV-QDs (1–3 nm) toward more negative values is indicative, which is expected to generate higher current and potential due to the quantum confinement effect of CV-QDs. This is observed in the XVB results. Third, the marked increase in current density with the CVT photoanode reiterates the influence of E&SI in CVT on efficient charge separation at the heterojunctions and charge migration to the redox sites.

3.4 Photocatalytic activity evaluation

CVT was evaluated for photocatalytic methane oxidation, initially without employing any co-catalyst, under one-sun conditions. ¹H NMR spectrum observed, after 3 h illumination of CVT in the CH₄-saturated water, is shown in Fig. 6a, demonstrating selective ethanol formation. The results are reproduced at least six times with CVT prepared in different batches, and the NMR spectra observed for three CVTs are compared, as shown in Fig. S10. Device stability is verified across three consecutive 3 h cycles, yielding consistent results, as shown in the inset in Fig. 6a. It is noteworthy that the same device was employed for all three cycles. A water suppression sequence was applied to eliminate the dominance of water; however, it suppresses the –OH feature of alcohols. A small amount of methanol was also formed, indicating the oxidation of methane to ethanol through methanol as an intermediate. 1 cm² area CVT photoanode gives 13.6 μmol/3 h ethanol yield. Ethanol formation is verified by spiking experiments in NMR measurements with a known amount of ethanol. Reference experiments using TiO₂ and CeVO₄ were also carried out under conditions identical to those of CVT. However, no significant value-added ethanol or other products were observed, confirming that E&SI and consequential heterojunctions between CV-QDs and TiO₂ are crucial for SMA.

Fig. 6 (a) ¹H-NMR spectra of liquid products of the representative reaction analyzed after 3 h illumination under one-sun conditions. Inset depicts the stability of the photocatalytic device over three cycles of 3 h each. (b) Schematic of the plausible mechanistic pathway for photocatalytic methane oxidation to ethanol over the CVT photocatalytic device and methane to ethylene over Pt-CVT.

To further understand the pathway for C–C coupling, a known amount (395 μmol) of methanol was deliberately added, and MA was carried out, as described above. The resultant ¹H-NMR spectrum (Fig. S11) demonstrated a 17.2 times larger (75.1 μmol h⁻¹ cm⁻²) amount of ethanol than without methanol addition. Additionally, no other products were detected. This result reiterates three points: (a) photogenerated methyl species finding a methanol/methoxy species was highly enhanced; (b) the diffusion length of the short-lived methyl species decreased significantly, and (c) it suggests that C–C coupling occurs after MA to methyl/methoxy species. As no ethane/ether was observed, no C–C or C–O–C coupling occurred among the reactive species. The above-enhanced ethanol yield and three points indicate that half of the 75.1 μmol carbon comes from methane, and the other half comes predominantly from added methanol. Hence, the effective enhancement of methane activation to ethanol may be considered 8.6 times or 37.5 μmol h⁻¹ cm⁻². These observations underscore the bifunctional nature of CVT. As CV-QDs absorb visible light and electrons are injected into the TiO₂ matrix, water oxidation and methane activation occur with holes concentrated in it. However, C–C coupling to ethanol is expected to occur at the interface of CV-QD and TiO₂, as this minimises the dispersion of electrons to longer distances and possible recombination of charge carriers. Further, this helps to enhance the interaction between short-lived intermediates, like the methyl radical. This aspect is further confirmed by applying a Pt co-catalyst to increase the MA activity, which will be discussed later.

MA was also carried out in direct sunlight. Compared to one-sun conditions, the ethanol yield obtained is lower (1.49 μmol h⁻¹ cm⁻²). This is attributed to the high temperatures (45–50 °C) observed in sunlight with the catalyst and solution, while 22 ± 2 °C is maintained under one-sun conditions. Methane solubility decreases at 45–50 °C; hence, its availability becomes a rate-limiting step. Nonetheless, the ethanol yield could be enhanced with reactor engineering, especially with methane recycling. It is worth evaluating MA in colder and high-altitude places on bright sunny days, such as Iceland and Ladakh. A major limitation of MA is the possibility of over-oxidation to undesired CO₂ by prolonging the reaction. However, the product analysis, after a 7 h reaction, showed no runaway reaction to CO₂ (Fig. S12), suggesting that the products did not undergo further oxidation.

Commercially available Pt/C (PT-1 platinum paste) was employed as a co-catalyst, coated on the edge of the CVT photoanode (Pt-CVT) and then evaluated for MA under the same conditions described earlier. A very interesting and critical change in the nature of the product was observed. Although a small amount of ethanol was observed, ethylene was selectively (≥95%) produced, and the same was confirmed by GC analysis (Fig. S13). Table 2 illustrates the ethylene and ethanol productivity with (and without) Pt co-catalyst. The ethylene yield observed (Pt-CVT) is nearly an order of magnitude higher than ethanol. This single observation underscores that effective charge separation at heterojunctions alone would not suffice, but electrons need to be either utilized immediately or stored in the co-catalyst. With 39.54 μmol h⁻¹ cm⁻² ethylene yield, the strategically designed artificial leaf device^18,19 can selectively demonstrate MA to ethylene. These results suggest that the use of the right co-catalyst promotes the efficient utilization of electrons towards improving selectivity and provides active sites for ethylene formation from ethanol through dehydration. A 4 cm² CVT photoanode with a Pt co-catalyst was also evaluated for MA. Although an increase in ethylene yield (97 μmol cm⁻² h⁻¹) was observed, the device configuration and reaction conditions need to be optimized for further improvement. STFE observed shows a trend similar to that of yield for ethanol/ethylene. Although the STFE for ethylene increases to 2.54, the same for ethanol increases to 4.82 with deliberately added methanol. To the best of our knowledge, this is the first time that a STFE of more than 1 has been reported for methane to ethanol/ethylene. Apparent quantum efficiency (AQE) was also measured at 550 nm under reaction conditions, and the details are presented in the methods section. AQE is observed to be 3.02%. Turn over frequency (ToF) was also calculated for ethanol/ethylene, and details are presented in the Methods section. 0.36 and 3.3 min⁻¹ were observed to be the TOF values for ethanol and ethylene, respectively. Nonetheless, ethanol ToF jumps to 6.2 with deliberately added methanol.

Table 2 C2 product yield and STFE observed with or without Pt co-catalyst and CVT under one-sun conditions

Conditions^a	Ethylene yield μmol^−1 cm^−2 h^−1 [STFE]	Ethanol yield μmol^−1 cm^−2 h^−1 [STFE]
a Methane saturated in 10 mL water at ice-cold conditions with a 1 cm^2 device under one-sun conditions for all reactions. b Under direct sunlight conditions. c With deliberately added methanol (395 μmol).
1 cm^2 CVT photoanode	Negligible amount	4.36/(1.49)^b [0.28]/[0.1]^b
1 cm^2 CVT photoanode	Negligible amount	75.0^c [4.82]
1 cm^2 Pt-CVT photoanode	39.54 [2.54]	Negligible
4 cm^2 Pt-CVT photoanode	96.95 μmol h^−1	Negligible
1 cm^2 CV photoanode	None	Negligible

---

Table 3 compares the ethanol/ethylene yield and reaction conditions for photo/electrocatalysts reported in the literature.^6,7,10,11 1 mg (gram) of the photoanode material spread over 1 cm² (1000 cm²) leads to 4.36 μmol mg_CVT⁻¹ h⁻¹ (4.36 mmol g_CVT⁻¹ h⁻¹) ethanol. Reference experiments demonstrated that without CV-QD integration, there is no meaningful activity. Hence, the activity was also normalized to 1 g of CV-QDs. Critically, the 1 cm² area of CVT contains 40 μg of CV-QDs in 960 μg TiO₂, or 1 : 24 = CV-QDs : TiO₂. Normalising the CV-QDs amounts to 1 g in 24 g of TiO₂ spread over 2.5 m² scales up the ethanol yield to 109 mmol g_CV-QD⁻¹ h⁻¹, which is far higher than that observed with noble-metal-based catalysts. Ethanol yield could be increased by ∼18 times to 1875 mmol g_CV-QDs⁻¹ h⁻¹ with deliberately added methanol in the reaction mixture. Similarly, ethylene formation reaches a value of 987.5 mmol g_CV-QD⁻¹ h⁻¹. Compared to any of the literature-reported values, the present study reports a new benchmark for SMA under ambient photocatalysis conditions.^28,29 Even if the large-sized device works at 50% efficiency, it is still lucrative. Since methods like spray-coating are available to make large-sized devices, device scale-up is a definite possibility; however, process conditions need to be optimized for the best performance.

Table 3 Ethanol yield by photocatalytic/electrochemical methane oxidation under ambient conditions by some of the top catalytic systems reported in the literature

Photocatalyst/photoanode	Ethanol yield	Reaction conditions	Ref.
a Activity normalized to per gram of CV-QDs. b Ethylene yield. c Total C2 (ethane and ethylene) yield. d Activity normalized to per gram of CV-QDs with deliberately added methanol.
NiO/Ni	25 μmol gNiO^−1 h^−1	0.1 M NaOH@1.4 V	7
Fe3Ni7(OH)x	9.09 mmol gcat^−1 h^−1	0.1 M NaOH@1.46 V vs. RHE	8
P-doped g-C3N4	51 μmol gcat^−1 h^−1	300 W Xenon lamp	13
Cu-modified g-C3N4	106 μmol gcat^−1 h^−1	500 W Xenon lamp	14
Pd-modified ZnO-Au	12.75 μmol gcat^−1 h^−1^b	300 W Xenon lamp	28
PdCu nano alloy over TiO2	1240 μmol gcat^−1 h^−1^c	365 nm LED (40 W)	29
CeVO4@TiO2	4.36 mmol gCVT^−1 h^−1	300 W Xenon lamp with one-sun condition	Present work
CeVO4@TiO2	109 mmol gCV-QDs^−1 h^−1^a
CeVO4@TiO2 with Pt	987.5 mmol C2H4 gCV-QDs^−1 h^−1^a^,^b
CeVO4@TiO2	1875 mmol gCV-QDs^−1 h^−1^d

---

3.5 Mechanism

SMA activity observed in the current study is attributed to the forced heterojunctions between CV-QDs and TiO₂ that help in the band-edge reconstruction of CVT, compared to its bulk TiO₂/CV counterparts. The potential gradient across the CV-QD-TiO₂ interface depends on the size of the CV-QD. The smallest CV-QDs boost the photocurrent generation at 400–460 nm, thereby efficiently generating the charge carriers and their dispersion. This, in turn, renders a better flow of the photogenerated charge carriers for SMA to one of the C–C coupled products. The IPCE result shows a gradual decrease in the IPCE response with increasing λ between 400 and 600 nm; thereafter, a constant current was observed. In fact, a near-plateau region was observed between 470 and 520 nm, instead of a steep drop in current, indicating that CVT is fully supportive of effective charge separation and is adequate to accomplish the redox reactions. As displayed in Fig. 6b, the CVT forms a type-II heterojunction; the excited electrons from CB_CV-QD are transferred to the CB_TiO2 and subsequently consumed to produce superoxide radicals, which can generate hydroxyl species further.^30,31 This pathway for electron transfer is substantiated by a reference experiment involving the deliberate addition of an electron scavenger (SI S3) to illustrate its effect on photocatalytic activity. In the presence of the Pt co-catalyst, electrons injected from the CV-QDs into TiO₂ are partially stored in Pt through the FTO bottom plate. This particular configuration helps for efficient electron storage and utilization of ethanol for ethylene reduction. However, the holes present in the VB_CV-QD are utilized to oxidize water, generating hydroxyl radicals.^14,31 Hydroxyl radicals are likely to activate the first C–H bond of methane to generate methyl radicals, which is the rate-determining step.^2,15 VB_Max at 0.75 eV with a highly active Ce³⁺ centre present at the heterojunctions of CVT is expected to activate methane and allow hydroxylation with OH radicals available nearby. The availability of more than one active site in close proximity enables the oxidation of methyl radicals with hydroxyl species to form methanol first, followed by coupling with another methyl radical to form ethanol.¹⁴

Experiments to trap active hydroxyl radicals were conducted using diammonium terephthalate and coumarin, which are widely used as hydroxyl radical trapping agents. A linear increase in the photoluminescence intensity with illumination time is observed, reiterating the continuous generation of the hydroxyl radical (Fig. S14a and b); without illumination, no intensity was observed, underscoring the light-induced reactions.³² Further, EPR studies were carried out with a methane-saturated solution containing the photocatalytic device in the custom-made quartz reactor along with DMPO. First, the spectrum of the aliquot was recorded before illumination, and no signal was detected. Then, it was irradiated under an Xe lamp for 15 min., and the aliquot collected was measured for trapped radicals by EPR. The appearance of the trapped radicals, i.e., methyl and hydroxyl, could be clearly observed, as shown in Fig. S14c, substantiating the proposed mechanistic pathway. Our results are also in good agreement with earlier reports on similar trapping EPR measurements.³³

The addition of Pt co-catalyst enhances nearly an order of magnitude increase in C₂-product (ethylene) yield (39.54 μmol h⁻¹ cm⁻²), compared to that of ethanol (4.36 μmol h⁻¹ cm⁻²). Photoexcited electrons from CV-QDs are not only injected into TiO₂ effectively, but they are also stored efficiently in Pt. This, in turn, decreases charge carrier recombination and increases the generation of methoxy/methyl species due to the efficient utilization of holes. This leads to enhanced C–C coupling to ethanol, which in turn undergoes dehydration to ethylene. Pt addition promotes charge separation/storage and consumption at a faster rate; hence, an overall rate of ethylene generation via ethanol is observed. An order of magnitude increase in CH₄ conversion also indicates that, as long as CH₄ is available, the rate of reaction remains high. Apart from charge carrier separation, their utilization is equally important for increasing the rate of overall reaction.

4 Conclusions

Assembling and integrating CeVO₄ QDs into the micro/mesopores of TiO₂ act as further stimuli for quantum confinement and band structure alignment with concomitant generation of ∼293 trillion heterojunctions. These QDs can be stabilized without any surfactant while exploiting the best properties, like quantum confinement, using the present method; E&SI leads to further fine-tuning of the electronic structure. CV-QD-TiO₂ heterojunctions facilitate the pathway of photo-generated charge carriers that plays a pivotal role in utilizing charge carriers for SMA to ethanol. With the addition of the Pt co-catalyst, the extent of charge separation increases, which leads to an enhancement of MA and, hence, selective ethylene formation.

The reported CVT system is better than other benchmark catalysts and hence opens a wide horizon of photocatalytic reactions to be innovated, incorporated and investigated.^3,34,35 It is reiterated that light absorption is uniform due to thin film geometry with a fixed thickness of the photocatalyst layers; hence, it can be scaled up to larger areas. The present findings not only provide a greener way of synthesising platform chemicals but also scale down the intensive energy input dramatically, hence increasing economic and environmental viability. Understanding the excited electronic states of nano-meter-sized pore-confined semiconductor crystallite³⁶ and its exploitation for charge carrier generation to drive appropriate redox reactions holds the key to the evolution of photocatalytic activation of inert molecules, like CH₄, and can be extended to N₂ and CO₂.

A new and green way of making quantum dots while integrating with another matrix, reported in the present study, is expected to widen the research field of quantum dots, especially towards many applications, such as optoelectronics, catalysis, and electrochemistry. Designing an AL device that can mimic green photosynthesis or similar chemical transformation with an STFE of more than 10% is a major challenge. Trillions of mandatory heterojunctions were created by assembling and integrating visible light absorbing semiconductor QDs in the pores of a wide bandgap semiconductor, which can be considered an innovative and skilful way of generating and utilizing charge carriers in the present AL device. With a minimum amount of material, enhanced activity in a variety of applications is possible. The ability to assemble quantum dots of metals and other compounds in another matrix, which can also withstand harsh reaction conditions, would further widen the spectrum of applications. For example, the assembly of Pt clusters in a porous carbon matrix could revolutionise fuel cells and electrolysers.

Conflicts of interest

There is no conflict of interest.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information: supplementary text for (S1) number of heterojunctions calculation, (S2) XPS details of Ce 3d core level, and (S3) scavenging experimental conditions. Supplementary figures are: (Fig. S1) photograph of experimental setup, (Fig. S2) TEM, (Fig. S3) FESEM, (Fig. S4) EDS, (Fig. S5) adsorption–desorption isotherm, (Fig. S6) indirect bandgap calculation, (Fig. S7) photoluminiscence, (Fig. S8) O 1s core level XPS, (Fig. S9) impedance analysis, (Fig. S10 and S11) NMR results, (Fig. S12 and S13) GC results, and (Fig. S14) spin-trapping results by PL and EPR. See DOI: https://doi.org/10.1039/d5se01178g.

Acknowledgements

SSK acknowledges the research fellowship received from INSPIRE, DST, New Delhi. HB, HJB and KNS acknowledge UGC and SARTHI CSMNRF-2019, respectively, for the research fellowship. We thank CSIR, New Delhi, for the financial support through an NCP project under HCP-044.

References

1. M. Ravi, M. Ranocchiari and J. A. van Bokhoven, Angew. Chem., Int. Ed., 2017, 56, 16464.
2. N. Feng, H. Lin, H. Song, L. Yang, D. Tang, F. Deng and J. Ye, Nat. Commun., 2021, 12, 4652.
3. (a) X. Chen, Y. Li, X. Pan, D. Cortie, X. Huang and Z. Yi, Nat. Commun., 2016, 7, 12273; (b) K. K. Patra and C. S. Gopinath, Chem. Commun., 2023, 59, 6774.
4. A. Dubey, S. K. Kolekar and C. S. Gopinath, ChemCatChem, 2016, 8, 2296.
5. A. Dubey, S. K. Kolekar and C. S. Gopinath, ChemCatChem, 2016, 8, 3650.
6. S. S. Kanungo, A. K. Mishra, N. B. Mhamane, U. Kiran Marelli, D. Kumar and C. S. Gopinath, Inorg. Chem., 2022, 61, 19577.
7. Y. Song, Y. Zhao, G. Nan, W. Chen, Z. Guo, S. Li, Z. Tang, W. Wei and Y. Sun, Appl. Catal., B, 2020, 270, 118888.
8. J. Li, L. Yao, D. Wu, J. King, S. S. C. Chuang, B. Liu and Z. Peng, Appl. Catal., B, 2022, 316, 121657.
9. (a) H. Song, X. Meng, Z. Jun Wang, H. Liu and J. Ye, Joule, 2019, 3, 1606; (b) H. Sato, A. Ishikawa, H. Saito, T. Higashi, K. Takeyasu and T. Sugimoto, Commun. Chem., 2023, 6, 1.
10. (a) S. S. Mani, S. Rajendran, T. Mathew and C. S. Gopinath, Energy Adv., 2024, 3, 1472; (b) S. Rajendran, S. Saju, T. Mathew and C. S. Gopinath, Chem. Commun., 2025, 61, 16738.
11. Y. Cao, W. Yu, C. Han, Y. Yang, Z. Rao, R. Guo, F. Dong, R. Zhang and Y. Zhou, Angew. Chem., Int. Ed., 2023, 62, 1.
12. (a) C. S. Gopinath and N. Nalajala, J. Mater. Chem. A, 2021, 9, 1353; (b) P. Vijyarengan, A. R. Maria, K. S. Ashadevi, N. Nalajala and C. S. Gopinath, Mater. Today Catal., 2025, 10, 100115; (c) N. Nalajala, K. K. Patra, P. A. Bharad and C. S. Gopinath, RSC Adv., 2019, 9, 6094; (d) B. Tudu, N. Nalajala, K. P. Reddy, P. Saikia and C. S. Gopinath, ACS Appl. Mater. Interfaces, 2019, 11, 32869; (e) S. Rajaambal, K. Sivaranjani and C. S. Gopinath, J. Chem. Sci., 2015, 127, 33; (f) P. Devaraji, M. Mapa, H. M. Abdul Hakkeem, V. Sudhakar, K. Krishnamoorthy and C. S. Gopinath, ACS Omega, 2017, 2, 6768.
13. X. Du, Z. Yang, X. Yang, Q. Zhang, L. Liu and J. Ye, Energy Fuels, 2022, 36, 3929.
14. Y. Zhou, L. Zhang and W. Wang, Nat. Commun., 2019, 10, 506.
15. (a) Y. Fan, W. Zhou, X. Qiu, H. Li, Y. Jiang, Z. Sun, D. Han, L. Niu and Z. Tang, Nat. Sustainable, 2021, 4, 509; (b) M. K. Sambhare, K. N. Salgaonkar, A. Saha and C. S. Gopinath, ChemCatChem, 2025, 17, e202500641.
16. S. Murcia-López, K. Villa, T. Andreu and J. R. Morante, ACS Catal., 2014, 4, 3013.
17. M. A. Becker, J. G. Radich, B. A. Bunker and P. V. Kamat, J. Phys. Chem. Lett., 2014, 5, 1575.
18. (a) K. K. Patra, B. D. Bhuskute and C. S. Gopinath, Sci. Rep., 2017, 7, 6515; (b) K. K. Patra, P. A. Bharad, V. Jain and C. S. Gopinath, J. Mater. Chem. A, 2019, 7, 3179.
19. K. N. Salgaonkar, H. Bajpai, N. B. Mhamane, N. Nalajala, I. Chauhan, K. Thakkar, K. Joshi and C. S. Gopinath, J. Mater. Chem. A, 2023, 11, 15168.
20. H. Bajpai, T. R. Nivedhitha, E. Dais, S. S. Kanungo and C. S. Gopinath, J. Catal., 2014, 437, 115644.
21. (a) H. Farhadi, M. M. Kamazani, N. Keramati and S. Alamdari, Sci. Rep., 2014, 1, 14824; (b) V. K. Jha, K. N. Salgaonkar, A. Saha, C. S. Gopinath and E. S. S. Iyer, Nanoscale, 2025, 17, 12738.
22. (a) L. Yao, X. Li, H. Liu, Z. Li and Q. Lu, J. Nanopart. Res., 2019, 21, 140; (b) K. Sivaranjani and C. S. Gopinath, J. Mater. Chem., 2011, 21, 2639.
23. W. Ding, X. Lin, G. Ma and Q. Lu, J. Environ. Chem. Eng., 2022, 8, 104588.
24. (a) I. Othman, J. Hisham Zain, M. Abu Haija and F. Banat, Appl. Catal., B, 2022, 266, 11860; (b) P. A. Bharad, K. Sivaranjani and C. S. Gopinath, Nanoscale, 2015, 7, 11206.
25. (a) E. S. Gnanakumar, J. M. Naik, M. Manikandan, T. Raja and C. S. Gopinath, ChemCatChem, 2014, 6, 3116; (b) S. B. Waghmode, R. Vetrivel, S. G. Hegde, C. S. Gopinath and S. Sivasanker, J. Phys. Chem. B, 2003, 107, 8517.
26. H. Bajpai, K. K. Patra, R. Ranjan, N. Nalajala, K. Prabhakar Reddy and C. S. Gopinath, ACS Appl. Mater. Interfaces, 2020, 12, 30420.
27. (a) D. Yu, Y. Jia, Z. Yang, H. Zhang, J. Zhao, Y. Zhao, B. Weng, W. Dai, Z. Li, P. Wang, J. A. Steele, M. B. J. Roeffaers, S. Dai, H. Huang, J. Long and A. C. S. Sus, Chem. Eng., 2022, 10, 16; (b) S. Rajendran, S. Saju, S. S. Mani, A. K. Asoka, A. Saha, P. S. Arun, B. Ghosh, T. Mathew and C. S. Gopinath, J. Mater. Chem. A, 2025, 13, 2105.
28. W. Jiang, J. Low, K. Mao, D. Duan, S. Chen, W. Liu, C. W. Pao, J. Ma, S. Sang, C. Shu, X. Zhan, Z. Qi, H. Zhang, Z. Liu, X. Wu, R. Long, L. Song and Y. Xiong, J. Am. Chem. Soc., 2021, 143, 269.
29. X. Li, C. Wang, J. Yang, Y. Xu, Y. Yang, J. Yu, J. J. Delgado, N. Martsinovich, X. Sun, X. S. Zheng, W. Huang and J. Tang, Nat. Commun., 2023, 14, 6343.
30. Y. Jiang, S. Li, S. Wang, Y. Zhang, C. Long, J. Xie, X. Fan, W. Zhao, P. Xu, Y. Fan, C. Cui and Z. Tang, J. Am. Chem. Soc., 2023, 145, 2698.
31. Y. Nosaka and A. Nosaka, ACS Energy Lett., 2016, 1, 356.
32. (a) T. J. Mason, J. P. Lorimer, D. M. Bates and Y. Zhao, Sonochemistry, 1994, 1, 2; (b) K. Hirano and T. Kobayashi, Ultrason. Sonochem., 2016, 30, 18.
33. (a) W. Zhao, Y. Shi, Y. Jiang, X. Zhang, C. Long, P. An, Y. Zhu, S. Shao, Z. Yan, G. Li and Z. Tang, Angew. Chem., Int. Ed., 2021, 60, 5811; (b) Y. Wang, X. Song, H. Zhang, J. Ying, Z. Liu, H. Xiong, T. Ji, W. Cao, Y. Bai, T. Ming, J. Zhu, X. Lu, W. Li and L. Mu, Chem. Eng. Sci., 2026, 321, 122744.
34. C. Williams, J. H. Carter, N. F. Dummer, Y. K. Chow, D. J. Morgan, S. Yacob, P. Serna, D. J. Willock, R. J. Meyer, S. H. Taylor and G. J. Hutchings, ACS Catal., 2018, 8, 2567.
35. S. Yoo, E. Lee, G. H. Jang and D. H. Kim, Mol. Catal., 2022, 533, 112791.
36. M. G. Bawendi, M. L. Steigerwald and L. E. Brus, Annu. Rev. Phys. Chem., 1990, 41, 477.

---