Photocatalytic Efficiency of Cerium Dioxide Hollow Spheres Sensitized with Bimetallic Au–Pd Nanoparticles for Enhanced Hydrogen Evolution

Abstract

CeO₂ with a hollow sphere morphology and a surface sensitized with gold (Au) and palladium (Pd) nanoparticles (NPs) was investigated as a photocatalyst for solar-driven hydrogen (H₂) production. Structural characterization by X-ray diffraction (XRD) and morphological analysis by scanning electron microscopy (SEM) confirmed the successful synthesis of the materials and revealed that no significant structural alterations occurred in the CeO₂ support following the surface modification procedures. Raman spectroscopy revealed a shift in the characteristic vibrational band of CeO₂ upon sensitization with metallic NPs, which was attributed to the generation of oxygen vacancies within the oxide lattice. The surface modification of the hollow CeO₂ spheres with metal NPs, exhibiting an average particle size of 6.38 ± 0.08 nm, contributed to an enhanced absorption of visible light and a concomitant increase in photocatalytic H₂ evolution. The modified photocatalysts demonstrated improved performance due to both the tailored morphology of the support, achieving an activity of 19.9 μmol∙g^-1∙h^-1, and its sensitization with metallic NPs, which further elevated the H₂ production rate to 45.4 μmol∙g^-1∙h^-1. These findings underscore the potential of combining morphological control with strategic chemical modification as an effective approach for the design of high-performance photocatalysts for solar H₂ generation.

Key words
hollow spheres; Au-Pd nanoparticles; oxygen vacancies; hydrogen production

INTRODUCTION

The rapid expansion of the global population, coupled with the intensification of industrial activities, has substantially amplified the worldwide demand for energy. According to the 2024 report by the International Energy Agency (IEA), global energy consumption increased by 2.2%, with fossil fuels remaining the dominant source, accounting for approximately 60% of the electricity generation matrix (Global Energy Review 2024 – Analysis - IEA). This situation contributes substantially to the increase in carbon dioxide (CO₂) emissions, which reached levels approximately 50% higher than those recorded in the pre-industrial period in 2024. This increase in atmospheric CO₂ concentrations intensifies the greenhouse effect, resulting in a rise in the global average temperature and triggering multiple adverse environmental implications (Global Energy Review 2025 – Analysis - IEA).

Within this context, the imperative to advance the research and deployment of alternative energy sources becomes increasingly evident. Hydrogen is regarded as a clean and sustainable energy carrier with the potential to substitute fossil fuels (Global Hydrogen Review 2024 – Analysis - IEA). The photocatalytic production of hydrogen represents a sustainable and promising technological approach for the provision of clean energy, offering a viable alternative to fossil fuel combustion and contributing to the mitigation of climate change impacts (Acar & Dincer 2022).

One approach for hydrogen production involves solar-driven water electrolysis, wherein water molecules are dissociated into hydrogen and oxygen gases. This process was first reported by Fujishima & Honda (1972), who employed a titanium dioxide (TiO₂) semiconductor to facilitate the photocatalytic splitting of water. Despite its utility, TiO₂ exhibits inherent limitations, notably its restricted spectral absorption confined to the ultraviolet (UV) region and a high tendency for recombination of photogenerated electron-hole pairs, which significantly reduces its overall photocatalytic efficiency (Fu et al. 2024). Therefore, an effective photocatalyst must exhibit high performance under irradiation in the ultraviolet-visible (UV-Vis) spectral region, as solar radiation comprises approximately 5% ultraviolet (UV) light, 43% visible light, and 52% near-infrared (NIR) light, thereby ensuring optimal overall process efficiency (Jiang et al. 2021).

A conventional photocatalytic process generally comprises three fundamental stages: initially, the photogeneration of electron–hole pairs (e⁻/h⁺); subsequently, the separation and migration of these photoinduced charge carriers to the surface of the photocatalyst; and finally, the occurrence of redox reactions at the catalytically active surface sites (Jiang et al. 2021).

The use of sacrificial agents has been widely adopted in photocatalytic systems to enhance hydrogen evolution efficiency by mitigating the rapid recombination of photogenerated electron–hole pairs. In this context, organic compounds such as alcohols, polyols, and carboxylic acids are frequently employed as electron donors, facilitating hole scavenging and thereby improving charge carrier separation. Among these compounds, glycerol has attracted particular attention as an effective sacrificial agent due to its biodegradability, low toxicity, and high availability as a major byproduct of the biodiesel industry. The utilization of glycerol in photocatalytic processes is therefore especially attractive, as it not only enhances hydrogen production efficiency but also contributes to the valorization of renewable biomass-derived residues, aligning photocatalytic hydrogen generation with principles of sustainability and circular economy (Zhao et al. 2023a, Tahir 2019).

Therefore, it is essential to develop photocatalysts exhibiting properties tailored to meet the specific requirements of the target process, while also demonstrating efficient energy absorption predominantly within the UV-vis spectral region.

Cerium oxide (CeO₂) is regarded as a promising photocatalyst owing to its distinctive physicochemical properties, including a favorable redox potential—arising from the reversible Ce⁴⁺/Ce³⁺ redox couple—which promotes efficient electron transfer processes. Additionally, CeO₂ exhibits a high oxygen storage capacity and a relatively low band gap energy, typically in the range of 2.7 to 2.9 eV, enhancing its photocatalytic performance under visible light irradiation (Fiorenza et al. 2021). Nevertheless, CeO₂ exhibits inherent limitations that hinder its photocatalytic efficiency, including the rapid recombination of photogenerated charge carriers and a restricted ability to absorb visible light. Consequently, various strategies have been explored to overcome these drawbacks, with surface modification and morphological control emerging as the most promising approaches to enhance its photocatalytic performance (Fu et al. 2024).

Hollow sphere structures have been investigated as alternative morphological supports across various fields, including photodegradation (Zaman et al. 2023), supercapacitors (Shi et al. 2023) and visible-light-driven catalysis (Zhao et al. 2023b). Notably, prior studies have demonstrated enhanced catalytic performance associated with hollow structures, attributed to their increased specific surface area, improved light-harvesting capabilities, and more effective separation and transport of charge carriers.

The integration of noble metal-based cocatalysts through surface modification strategies emerges as a promising approach to enhance the photocatalytic efficiency of semiconductor oxides (Fu et al. 2024). Noble metals such as gold (Au) and palladium (Pd) enhance visible-light absorption by inducing defect-related electronic states and promoting efficient charge transfer across the metal–semiconductor interface, thereby reducing the effective photoactivation energy of the system. This effect arises from their capacity to trap photogenerated electrons, thereby suppressing electron–hole recombination and consequently improving the efficiency of hydrogen (H₂) evolution. Such behavior is intrinsically linked to the work function (ϕ) of the metal nanoparticles, which governs their ability to absorb incident photons and mediate charge transfer to the catalyst surface—a fundamental process for hydrogen production (Fu et al. 2024). Given that the work functions of the metals Au (ϕ = 5.1 eV) (Aghili et al. 2024) and Pd (ϕ = 5.6 eV) (Uddin et al. 2023) are higher than that of the semiconductor CeO₂ (ϕ = 2.3 eV) (Song & Wu 2023), an energy level alignment is established at the metal–semiconductor interface. This alignment effectively suppresses electron–hole recombination and promotes the efficient separation of charge carriers.

In this study, we propose the synthesis of an enhanced photocatalyst composed of hollow CeO₂ spheres sensitized with Au and Pd nanoparticles. The resulting system demonstrated that the synergistic interaction between NPs with the support with hollow spherical morphology and oxygen vacancy properties significantly enhanced photocatalytic H₂ evolution under UV-vis radiation.

EXPERIMENTAL SECTION

All chemicals used in the experiments were of analytical grade commercially obtained from Sigma Aldrich® and were used without any additional purification procedures.

Synthesis _of SiO2 spheres

The synthesis of SiO₂ colloidal particles was carried out following the Stöber et al. (1968) method with some modifications (Bie et al. 2019). Initially, 50 mL of ethanol and 40 mL of deionized water were introduced into a round-bottom flask equipped with a reflux system, and the mixture was stirred at 500 rpm at ambient temperature. Subsequently, 24 mL of ammonium hydroxide solution (28% w/w) and 6 mL of tetraethoxysilane (TEOS) were added dropwise under continuous stirring at 1000 rpm for 1 min. The reaction mixture was maintained under these conditions for 3 h. At the end of the process, the resulting white precipitate was collected by centrifugation, washed five times with deionized water and ethanol, and subsequently dried in an oven at 60 °C for 12 h.

Synthesis _of Si_O2@CeO2 core-shell

The synthesis of the SiO₂@CeO₂ core–shell structure was conducted following the procedure described by Zhang et al. (2016), with minor modifications. Initially, 0.15 g of the previously synthesized SiO₂ spheres were dispersed in 30 mL of ethanol in a beaker. The suspension was subjected to ultrasonic agitation until complete dispersion of the solid was achieved. Concurrently, a solution was prepared by dissolving 0.5 g of urea and 2.6 g of cerium nitrate hexahydrate [Ce(NO₃)₃·H₂O] in 40 mL of deionized water. This solution was then added to the silica suspension, and the mixture was further sonicated for 1 h. At the end of the process, the resulting mixture was transferred to a Teflon-lined stainless-steel autoclave and heated in an oven at 160 °C for 8 h. After cooling to room temperature, the obtained grayish solid was washed thoroughly with deionized water, recovered via vacuum filtration, and dried in an oven at 60 °C overnight. The dried solid was subsequently calcined in a muffle furnace at 600 °C for 3 h, employing a heating rate of 5 °C min^-1. This process yielded a yellow solid, which was stored in a desiccator for subsequent characterization.

Synthesis of holl_ow CeO2 spheres

The removal of the SiO₂ core was carried out following the methodology described by Li et al. (2021). In this procedure, 100 mg of the previously calcined core–shell material was dispersed in a mixture of 5 mL of deionized water and 5 mL of 1 M NaOH solution. The dispersion was maintained under magnetic stirring at 500 rpm in a reflux system at 90 °C for 2 h. Subsequently, the resulting material was isolated by centrifugation, thoroughly washed with deionized water (six cycles), and dried in an oven at 60 °C for 12 h.

Synthesis of metal catalysts

The synthesis of bimetallic (Au–Pd) and monometallic catalysts was performed via the urea deposition–precipitation (UDP) method, as previously described (Sousa et al. 2022) The preparation of the Au–Pd catalyst with a theoretical molar ratio of 1:1 is detailed herein as a representative example. Typically, to achieve a theoretical metal loading of 2 wt%, 500 mg of the support material (H–CeO₂) was dispersed in 50 mL of deionized water and maintained under continuous magnetic stirring at ambient temperature using a sand bath. Subsequently, urea [CO(NH₂)₂, 99%, Sigma-Aldrich®] was introduced at a urea-to-metal molar ratio of 100. After a stirring period of 5 minutes, aqueous solutions of HAuCl₄·3H₂O (99%, Sigma-Aldrich®) and PdCl₂ (5 wt% in 10 wt% HCl, Sigma-Aldrich®) were added sequentially. The resulting mixture was gradually heated to 95 °C and maintained under stirring for 4 h. The solid product was then separated by centrifugation, washed four times with deionized water, and subsequently dried in an oven at 70 °C. Finally, calcination was carried out at 600 °C for 3 h in a muffle furnace, employing a heating rate of 5 °C min^-1.

Photocatalytic reactions

Photocatalytic hydrogen evolution experiments were conducted in a calibrated, gas-tight, double-walled quartz photochemical reactor with a total volume of 35 mL. The reactor was equipped with a water-circulation system to maintain the solution temperature at 25 °C and was subjected to continuous magnetic stirring to ensure homogeneity. For each experiment, 7.5 mg of photocatalyst was dispersed in 15 mL of an aqueous glycerol solution (5 wt%) and subsequently introduced into the reactor. Prior to irradiation, the system was purged by applying vacuum cycles under argon flow for approximately 10 min, effectively minimizing the residual oxygen content. The photocatalytic performance of the materials toward hydrogen evolution via hydrogen evolution reaction was evaluated under UV–visible light irradiation over a period of 3 h. Illumination was provided by a 600 W solar simulator (Newport, USA; Power Solar) and a 450 W xenon lamp (Ushio, California, USA), simulating solar intensity. The incident light intensity was monitored using a calibrated silicon photodiode. Optical filters (FSQ-GG400 and FSQ-UG5, Newport Corporation, California, USA) were employed when necessary to select specific irradiation wavelengths. The evolved hydrogen accumulated in the headspace of the sealed quartz reactors was quantitatively analyzed using an Agilent gas chromatograph equipped with a thermal conductivity detector (TCD) and 30 m HP-PLOT/Q and HP-MOLESIEVE columns (Agilent Technologies, Santa Clara, USA). Hydrogen production data were recorded at regular 30-minute intervals up to a total irradiation time of 210 minutes to determine the evolution rates. Gas samples (up to 450 μL) were extracted from the sealed system using a gas-tight Sample Lock syringe (Hamilton).

Characterizations

The crystal structures of the synthesized materials were characterized by X-ray diffraction obtained with a Bruker D8 Advance® X-ray diffractometer with a Cu Kα radiation source (λ = 1.5406 Å) at a copper emission tube voltage of 40 kV and filament current of 40 mA, in a 2θ range from 5° to 90°, a step size of 0.02°, and a measurement time of 5 s per step. The average crystallite size of the materials was determined using the Scherrer equation (Equation 1), based on the most intense diffraction plane – (111) – for both samples. Transmission electron microscopy (TEM) images were obtained using a MORGAGNI 268D® microscope. The samples were prepared in colloidal solutions with isopropanol and deposited on a carbon-coated copper grid. The average nanoparticle size was determined using ImageJ® software considering a count of 40 particles. Scanning electron microscopy (SEM) images were obtained using a Quanta® 200F FEG microscope. Raman spectra were acquired using a Bruker Senterra II® instrument coupled to an optical microscope (Bruker AXS GmbH, Karlsruhe, Germany) with a 50x objective lens and a 785 nm laser as the excitation source. Low laser power density was used to avoid sample overheating. A spectral resolution of 3 cm-1 was used, and the wavelength range was 100-1000 cm-1. UV-Vis diffuse reflectance spectra were obtained considering the reflectance of a standard (BaSO₄) using a UV/Vis spectrophotometer (Agilent, Cary 300) equipped with a diffuse reflectance sampling accessory (Agilent, Santa Clara, USA). The samples were pressed into pellets of 2 g of BaSO₄ and 50 mg of material.

$\mathrm{D}=\mathrm{ }\frac{\mathrm{k·\lambda }}{\mathrm{\beta ·cos\theta }}$ (1)

RESULTS AND DISCUSSIONS

The crystal structures and phase composition of the synthesized materials were characterized by X-ray diffraction analysis. Crystallographic data retrieved from the Inorganic Crystal Structure Database® (ICSD) were employed to identify the principal reflection features of the samples, as illustrated in Figure 1. X-ray diffractogram of (a) the synthesized materials and (b) magnified view between 25 and 35°. Raman spectrum of (c) support (H-CeO2) and catalyst (AuPd/H-CeO2) and (d) magnified view between 300 and 650 cm-1.a, b. The XRD patterns of SiO₂ exhibit only a broad diffraction halo centered around 2θ ≈ 23°, which is indicative of the presence of amorphous silica (Xunwen et al. 2020).

Figure 1
Table SI-SIII.

In the X-ray diffraction patterns of the SiO₂@CeO₂ core–shell material, the hollow CeO₂ spheres (H-CeO₂) support, and the Au–Pd bimetallic catalyst, distinct diffraction peaks are observed at 2θ values of 28.6°, 33.1°, 47.5°, 56.4°, 59.1°, 69.5°, 76.8°, 79.1°, and 88.5°. These peaks can be indexed to the (111), (200), (220), (311), (222), (400), (331), (420), and (422) crystallographic planes, respectively, which confirms that the cerium oxide phase was successfully synthesized with a face-centered cubic fluorite-type structure (space group: Fm3̅m), consistent with the crystallographic data reported in ICSD Nº 155604 (Yashima et al. 2006).

The absence of the characteristic silica diffraction peak in the synthesized materials is attributed to the increased crystallinity resulting from the CeO₂ coating process (Xunwen et al. 2020). Moreover, the lack of a silica-related diffraction pattern in the core–shell XRD profile confirms the effective removal of the SiO₂ core during the synthesis of hollow spheres via the chemical etching method.

The magnified X-ray diffraction pattern in the 2θ range between 25° and 35° (Figure 1. X-ray diffractogram of (a) the synthesized materials and (b) magnified view between 25 and 35°. Raman spectrum of (c) support (H-CeO2) and catalyst (AuPd/H-CeO2) and (d) magnified view between 300 and 650 cm-1.b) demonstrates the absence of any discernible shift in the primary diffraction peaks associated with CeO₂. This observation suggests that the removal of the core did not alter or distort the crystalline structure of the shell, thereby preserving its structural integrity. A similar behavior was observed following the deposition of metal nanoparticles onto the hollow CeO₂ spheres. However, a noticeable narrowing of the diffraction peaks was detected, which is likely attributable to an increased concentration of oxygen vacancy defects—an intrinsic characteristic of the CeO₂ support. Nevertheless, a narrowing of the diffraction pattern was observed, which can be attributed to an increase in the average crystallite size after the deposition of the metallic nanoparticles, as observed in Supplementary Material - Table SI (Yang et al. 2019, Lv et al. 2021).

The average crystallite size of H-CeO₂ and AuPd/H-CeO₂ was determined using the Scherrer equation (Equation 1), based on the most intense diffraction plane - (111) - for both samples. In this equation, D represents the average crystallite size, k is the shape factor (0.9), λ is the X-ray wavelength (1.5406 Å), β denotes the full width at half maximum (FWHM) of the selected diffraction peak, and θ is the corresponding Bragg angle (Table SI). The calculated average crystallite size for the CeO₂ hollow spheres was 7.9 nm. Upon deposition of Au-Pd nanoparticles, the average crystallite size increased to 13.9 nm.

No diffraction peaks corresponding to Au nanoparticles were observed, for which typical reflections occur at approximately 2θ = 38°, 44°, and 64°, nor to Pd, which generally exhibits diffraction peaks at around 2θ = 40°, 46°, and 68° (Balasurya et al. 2023, Qian et al. 2023). This behavior may be related to their low concentrations (2 wt%) relative to the support, as well as to their small particle size (approximately 6 nm), as determined by transmission electron microscopy (Figure 2. AuPd/H-CeO2 catalyst: a) SEM image, b) TEM image and c) particle size distribution histogram.b) (Sousa et al. 2022).

Figure 2
X-ray diffractogram of (a) the synthesized materials and (b) magnified view between 25 and 35°. Raman spectrum of (c) support (H-CeO₂) and catalyst (AuPd/H-CeO₂) and (d) magnified view between 300 and 650 cm-1.

Raman spectroscopy was employed to investigate the crystalline structure, vibrational properties, and defects, particularly oxygen vacancies, of the materials. Figure 1. X-ray diffractogram of (a) the synthesized materials and (b) magnified view between 25 and 35°. Raman spectrum of (c) support (H-CeO2) and catalyst (AuPd/H-CeO2) and (d) magnified view between 300 and 650 cm-1.c, d presents the Raman spectra of the H-CeO₂ and AuPd/H-CeO₂ samples.

In Figure 1. X-ray diffractogram of (a) the synthesized materials and (b) magnified view between 25 and 35°. Raman spectrum of (c) support (H-CeO2) and catalyst (AuPd/H-CeO2) and (d) magnified view between 300 and 650 cm-1.c, it can be observed that both analyzed samples exhibit a single band of significant intensity. Figure 1. X-ray diffractogram of (a) the synthesized materials and (b) magnified view between 25 and 35°. Raman spectrum of (c) support (H-CeO2) and catalyst (AuPd/H-CeO2) and (d) magnified view between 300 and 650 cm-1.d highlights the spectral region between 300 and 650 cm^-1 for a more detailed analysis of the vibrational features. In the core–shell spectrum, the prominent band at 462 cm^-1 is assigned to the F_{2 g} vibrational mode, characteristic of the fluorite-type structure within the Fm3̅m space group, corresponding to a symmetric stretching of Ce–O bonds (Rao et al. 2018). Upon incorporation of metallic nanoparticles into the support matrix, a red shift of this band to 455 cm^-1 (lower frequency region) is observed. This shift is attributed to a decrease in the lattice parameter, resulting from a shortened Ce–O bond length, which in turn is induced by the formation of oxygen vacancies within the support (Zhu et al. 2019, Jayakumar et al. 2019).

The quantitative determination of defect-site concentrations in the materials can be achieved by evaluating the relative intensities of the D and F_{2 g} bands (I_D/I_{F 2 g}), which allows the estimation of oxygen vacancy concentrations (Xue et al. 2022, Zhang et al. 2025). By analyzing the Raman spectra shown in Figure 1. X-ray diffractogram of (a) the synthesized materials and (b) magnified view between 25 and 35°. Raman spectrum of (c) support (H-CeO2) and catalyst (AuPd/H-CeO2) and (d) magnified view between 300 and 650 cm-1.c, the oxygen vacancy concentrations of the investigated materials can be estimated. The I_D/I_{F 2 g} ratios are 0.029 and 0.163 for the support and the Au–Pd/H–CeO₂ catalyst, respectively. This indicates that the catalyst exhibits a higher concentration of oxygen vacancies than the H–CeO₂ support, which is associated with metal–support interactions and the synergy between the equimolar nanoparticles, thereby enhancing the catalytic performance.

The morphology of the AuPd/H-CeO₂ catalyst was examined using scanning electron microscopy (SEM). Following the removal of the SiO₂ template, a more distinct contrast between the shell and the core of the spheres was observed, evidencing the successful formation of hollow CeO₂ spheres and the effective chemical etching of the SiO₂ template, as illustrated in Figure 2. AuPd/H-CeO2 catalyst: a) SEM image, b) TEM image and c) particle size distribution histogram.a.

The synthesized spheres exhibited a rough surface morphology, which is likely attributable to the self-assembly process. Additionally, the particles were uniform in both size and shape, presenting an average diameter of 352 ± 37 nm and a shell thickness of approximately 48 ± 5 nm (Figure 2. AuPd/H-CeO2 catalyst: a) SEM image, b) TEM image and c) particle size distribution histogram.a). These characteristics were considered satisfactory for catalytic applications, as a high surface area combined with thicker shells contribute to enhanced stability and improved catalytic performance (Li et al. 2021, Zhang et al. 2016).

Figure 2. AuPd/H-CeO2 catalyst: a) SEM image, b) TEM image and c) particle size distribution histogram.b presents a representative transmission electron microscopy (TEM) image of the AuPd/H-CeO₂ catalyst, along with the corresponding particle size distribution histogram of the metallic nanoparticles deposited on the support surface. The image reveals a homogeneous dispersion of the nanoparticles, exhibiting an average diameter of 6.38 ± 0.08 nm (Figure 2. AuPd/H-CeO2 catalyst: a) SEM image, b) TEM image and c) particle size distribution histogram.c).

The actual Au and Pd loadings on the H-CeO₂ support were quantified by ICP–OES, and the results are summarized in Table SIII. Approximately 97% of the targeted metal content was successfully deposited onto the support. For the catalyst synthesized with a nominal Au:Pd molar ratio of 1:1, ICP–OES analysis revealed an experimental Au:Pd ratio of 1:1.4. In general, the measured metal loadings were slightly lower than the nominal values, particularly for Au, suggesting a preferential deposition and stronger interaction of Pd species with the H-CeO₂ surface.

The size of metal nanoparticles significantly influences their light absorption characteristics due to the phenomenon of surface plasmon resonance (SPR), particularly in the case of gold nanoparticles (AuNPs). SPR refers to the collective oscillation of conduction electrons at the surface of the nanoparticles upon excitation by incident light, resulting in the generation of a localized electromagnetic field. This field enhances light confinement at the metal–semiconductor interface, thereby improving photocatalytic activity (Hu et al. 2020). Previous studies have demonstrated that metal nanoparticles with diameters ranging from 3 to 7 nm exhibit a redshift in the light absorption peak toward longer wavelengths (within the visible region), a behavior attributed to the SPR effect, which contributes to enhanced catalytic performance. This phenomenon was observed in the present work upon deposition of the nanoparticles onto the H-CeO₂ support (Figure 3. (a) UV-Vis spectra (Kubelka-Munk function) of H-CeO2 support and (b) AuPd/H-CeO2 catalyst; (c) Diffuse reflectance spectrum of H-CeO2 support and (d) AuPd/H-CeO2 bimetallic catalyst.c, d) (Li et al. 2024a, Wang et al. 2018).

Figure 3
AuPd/H-CeO₂ catalyst: a) SEM image, b) TEM image and c) particle size distribution histogram.

Figure 3. (a) UV-Vis spectra (Kubelka-Munk function) of H-CeO2 support and (b) AuPd/H-CeO2 catalyst; (c) Diffuse reflectance spectrum of H-CeO2 support and (d) AuPd/H-CeO2 bimetallic catalyst.a, b presents the diffuse reflectance spectra of the H-CeO₂ support and the AuPd/H-CeO₂ bimetallic photocatalyst, derived from the Kubelka-Munk function and analyzed using Tauc plot extrapolation as a function of photon energy (eV). Indirect electronic transitions were evaluated by plotting (F(R)hν)n as a function of photon energy (hν), where n=1/2 corresponds to an indirect transition. A noticeable reduction in the optical band gap was observed, decreasing from 2.67 eV for the support to 2.46 eV for the photocatalyst. This band gap narrowing is attributed to the generation of oxygen vacancies in the presence of metallic nanoparticles, which promote electron transfer processes between the support and the deposited metal particles.

The absence of significant variations in the band gap energies (Eg) between the photocatalyst (AuPd/H-CeO₂) and its support (H-CeO₂) is primarily attributed to structural and electronic modifications occurring at the metal–support interface, where the formation of Schottky barriers promotes charge separation and enhances photocatalytic activity, without inducing direct alterations in the intrinsic Eg of the material (Hu et al. 2020).

The diffuse reflectance spectra of the H-CeO₂ support and the AuPd/H-CeO₂ bimetallic catalyst are presented in Figure 3. (a) UV-Vis spectra (Kubelka-Munk function) of H-CeO2 support and (b) AuPd/H-CeO2 catalyst; (c) Diffuse reflectance spectrum of H-CeO2 support and (d) AuPd/H-CeO2 bimetallic catalyst.c, d. Analysis of the spectra reveals that both materials predominantly absorb electromagnetic radiation in the ultraviolet region (200 –400 nm). Furthermore, the deposition of Au and Pd metallic nanoparticles onto the support resulted in an enhanced absorption of radiation within the visible region of the spectrum. This increase in visible light absorption was sufficient to drive significant hydrogen production rates, considering that sunlight, utilized as the energy source for hydrogen evolution reaction, contains a relatively low proportion of ultraviolet radiation and a substantially higher proportion of visible and infrared radiation.

Photocatalytic Activity

The behavior of H–CeO₂ and AuPd/H–CeO₂ was investigated as photocatalysts for H₂ production for hydrogen evolution reaction, using glycerol as a sacrificial agent to donate electrons and facilitate the scavenging of photogenerated holes upon light absorption by the semiconductor. This process suppresses e⁻/h⁺ recombination, thereby increasing the efficiency of free-electron generation for water reduction to form hydrogen (Marques et al. 2017).

Figure 4. Photocatalytic H2 Evolution from AuPd/H-CeO2 and H-CeO2. illustrates the photocatalytic hydrogen (H₂) evolution in the presence of the support material alone (H-CeO₂) and the bimetallic catalyst (AuPd/H-CeO₂) under UV-Visible irradiation. It is evident that the hollow support (represented in black), despite the absence of metallic species, exhibits notable photocatalytic activity within the studied times, achieving a maximum hydrogen production of 19.9 μmol·g^-1 after 180 minutes. This enhanced performance can be attributed to the morphological characteristics of the material, which promote increased internal reflection of incident radiation, thereby augmenting light absorption and facilitating a higher generation rate of electron-hole pairs essential for the photocatalytic process. These findings demonstrate that hollow CeO₂ spheres outperform alternative morphologies, such as rods and solid spheres, which exhibited significantly lower activity levels, with hydrogen evolution rates below 2.0 μmol·g^-1 (Yan et al. 2023, Hao et al. 2017).

Figure 4
(a) UV-Vis spectra (Kubelka-Munk function) of H-CeO₂ support and (b) AuPd/H-CeO₂ catalyst; (c) Diffuse reflectance spectrum of H-CeO₂ support and (d) AuPd/H-CeO₂ bimetallic catalyst.

Upon evaluating the catalytic performance of the bimetallic system (depicted in red), we observed an enhancement in hydrogen (H₂) production rates of up to 184% relative to the bare support, achieving a value of 45.39 μmol·g^-1 over a 180-minute reaction period. This marked improvement is attributed to the incorporation of AuPd nanoparticles, which function as co-catalysts by facilitating the efficient capture of photogenerated electrons and thereby inhibiting their recombination with holes (Li et al. 2024b, Hu et al. 2020).

The presence of oxygen vacancies in the photocatalyst leads to a reduction in the band gap energy, thereby enhancing the absorption of visible light (Fiorenza et al. 2021, Hu et al. 2020). This property accounts for the superior performance of the bimetallic catalyst, which demonstrates enhanced absorption of visible light compared to the unmodified support, as evidenced by UV-vis spectroscopy analyses (Figure 3. (a) UV-Vis spectra (Kubelka-Munk function) of H-CeO2 support and (b) AuPd/H-CeO2 catalyst; (c) Diffuse reflectance spectrum of H-CeO2 support and (d) AuPd/H-CeO2 bimetallic catalyst.c, d).

These results show an improvement compared with other CeO₂-based materials with different morphological structures (Table SII), indicating that hollow spherical architectures combined with metallic nanoparticles effectively suppress electron–hole recombination, thereby providing higher hydrogen production rates from water electrolysis reactions.

Photocatalytic Mechanism

Based on the experimental results obtained in this study, it is evident that the investigated materials exhibit significant hydrogen production rates, with AuPd/H-CeO₂ demonstrating the highest photocatalytic activity under visible light irradiation.

The presence of oxygen vacancies within the photocatalyst, in conjunction with the incorporation of bimetallic AuPd nanoparticles and the unique morphology of the hollow CeO₂ support, extends the light absorption capability into the visible region. Accordingly, a charge transfer mechanism is proposed for the AuPd/H-CeO₂ system, as depicted in Figure 5. Schematic representation of the proposed mechanism for H2 production under sunlight irradiation on AuPd/H-CeO2 photocatalyst.. The hollow architecture of the support functions as a cavity for multiple internal reflections of incident photons, thereby enhancing light harvesting efficiency and promoting the generation of electron–hole pairs (e⁻/h⁺) essential for driving redox reactions (Wang et al. 2021).

Figure 5
Photocatalytic H₂ Evolution from AuPd/H-CeO₂ and H-CeO₂.

Upon irradiation with ultraviolet (UV) light, the high-energy photons promote electrons from the valence band (VB) to the conduction band (CB) of H-CeO₂ (equation 2). In contrast, under visible light illumination, which possesses lower photon energy, the excitation of VB electrons occurs toward the oxygen vacancy states in AuPd/H-CeO₂, located approximately 0.26 eV below its conduction band (equation 3) (Skorodumova et al. 2002).

$$\mathrm{Ce}{O}_2+h\nu \to e_{\mathrm{CB}}^{-}+h_{\mathrm{VB}}^{+}$$ (2)

$$e_{\mathrm{CB}}^{-}\to {\mathrm{Vo}}_{}$$ (3)

$$e_{\mathrm{CB}}^{-}(H-\mathrm{Ce}{O}_2)\to \mathrm{Au}-\mathrm{Pd}$$ (4)

With the enhanced absorption of visible light by the photocatalyst, photoexcited electrons are promoted to oxygen vacancy states present in the semiconductor. These electrons are subsequently transferred to the surface of noble metal nanoparticles, which function as co-catalysts by forming a Schottky junction, resulting in efficient charge separation (equation 4). This junction facilitates the accumulation of photogenerated electrons and promotes the reduction of protons (H⁺) to molecular hydrogen (H₂), while simultaneously enhancing charge separation by suppressing electron–hole recombination, thereby favoring oxidative processes (Liu et al. 2017, Hu et al. 2020).

Under visible light, localized surface plasmon resonance (LSPR) occurs, where the generated hot electrons directly participate in proton reduction. On the Pd surface, hydrogen adsorption occurs (equation 5) followed by hydride formation (equation 6), acting as an active hydrogen reservoir, where the recombination of adsorbed hydrogen (H*) leads to the formation of H₂ (equation 7). However, the photogenerated holes are consumed by the oxidation of glycerol, which acts as a sacrificial electron donor and is preferentially oxidized compared to water, generating protons and suppressing charge recombination (equation 8).

$$H^{+}+e^{-}\to H\ast$$ (5)

$$\mathrm{Pd}+H^{\ast }\to \mathrm{PdHx}$$ (6)

$$2\mathrm{PdHx}\to H_2+\mathrm{Pd}$$ (7)

$$C_3{H}_8{O}_3+H_{2}O\to \text{oxidized products}+H^{+}+e^{-}$$ (8)

Consequently, the synergistic integration of a hollow structural architecture, oxygen vacancies, and noble metal nanoparticles results in an increased specific surface area and a higher density of active sites, ultimately leading to improved photocatalytic performance in H₂ evolution.

CONCLUSIONS

In this study, the influence of morphological modification of CeO₂, achieved via the formation of hollow spheres, and surface sensitization with Au and Pd metallic nanoparticles was investigated with respect to the efficiency of photocatalytic hydrogen evolution reaction for H₂ production under solar irradiation. SEM analyses showed the formation of hollow spheres with an average diameter of 352 nm and an average shell thickness of 48 nm. The photocatalyst with only H-CeO₂ exhibits remarkable photocatalytic activity, reaching a hydrogen production of 19.9 μmol·g^-1. The AuPd/H-CeO₂ photocatalyst exhibited enhanced photocatalytic activity, reaching 45.39 μmol·g^-1 during a reaction period of 180 minutes. The AuPd/H-CeO₂ photocatalyst exhibited enhanced photocatalytic activity, which is attributed to increased radiation absorption through multiple internal reflections within the hollow architecture, as well as to the presence of Au and Pd nanoparticles. These nanoparticles promoted the generation of oxygen vacancies and extended the optical absorption of the material into the visible region. The synergistic effect between morphological engineering, surface functionalization, and the induction of structural defects appears to play an important role in development of an efficient and sustainable photocatalytic system for solar-driven hydrogen evolution.

SUPPLEMENTARY MATERIAL

Table SI-SIII.

Table SI-SIII.

Acknowledgements

The authors are grateful for the financial support from FAPEPI (05/2021) and technical support from the Northeast Strategic Technology Center (CETENE-PE). The authors declare no conflict of interest regarding the publication of this article.

• Data availability
  All data generated or analyzed during this study are included in this published article.

References

• ACAR C & DINCER I. 2022. Selection criteria and ranking for sustainable hydrogen production options. Int J Hydrogen Energy 47: 40118-40137.
• AGHILI B, RAHBARPOUR S, BERAHMAN M & HORRI A. 2024. Influence of Surface Roughness on the Work Function of Gold: A Density Functional Theory Study. J Phys Chem C 128: 8077-8084.
• BALASURYA S, AL-GHAMDI AA, OKLA MK, AL-AMRI SA, ABDEL-MAKSOUD MA, AUFY M & KHAN SS. 2023. Efficient photocatalytic activity of Au@Mg nanospheres on mineralization of polystyrene: A sustainable remediation strategy on sunlight-induced photodegradation, environmental toxicity, and sensing of cefixime. J Water Process Eng 51: 103350.
• BIE C, ZHU B, XU F, ZHANG L & YU J. 2019. In Situ Grown Monolayer N-Doped Graphene on CdS Hollow Spheres with Seamless Contact for Photocatalytic CO₂ Reduction. Advanced Materials 31.
• FIORENZA R, BALSAMO SA, CONDORELLI M, D’URSO L, COMPAGNINI G & SCIRÈ S. 2021. Solar photocatalytic H₂ production over CeO₂-based catalysts: Influence of chemical and structural modifications. Catal Today 380: 187-198.
• FU W, ZHANG Y, ZHANG X, YANG H, XIE R, ZHANG S, LV Y & XIONG L. 2024. Progress in Promising Semiconductor Materials for Efficient Photoelectrocatalytic Hydrogen Production. Molecules 29: 289.
• FUJISHIMA A & HONDA K. 1972. Electrochemical photolysis of water at a semiconductor electrode. Nature 238: 37-38.
• HAO Y, LI L, ZHANG J, LUO H, ZHANG X & CHEN E. 2017. Multilayer and open structure of dendritic crosslinked CeO₂-ZrO₂ composite: Enhanced photocatalytic degradation and water splitting performance. Int J Hydrogen Energy 42: 5916-5929.
• HU H, QIAN D, LIN P, DING Z & CUI C. 2020. Oxygen vacancies mediated in-situ growth of noble-metal (Ag, Au, Pt) nanoparticles on 3D TiO₂ hierarchical spheres for efficient photocatalytic hydrogen evolution from water splitting. Int J Hydrogen Energy 45: 629-639.
• JAYAKUMAR G, ALBERT IRUDAYARAJ A & DHAYAL RAJ A. 2019. A comprehensive investigation on the properties of nanostructured cerium oxide. Opt Quantum Electron 51.
• JIANG L, YANG J, ZHOU S, YU H, LIANG J, CHU W, LI H, WANG H, WU Z & YUAN X. 2021. Strategies to extend near-infrared light harvest of polymer carbon nitride photocatalysts. Coord Chem Rev 439: 213947.
• LI H, WANG S, CHI H & LI C. 2024a. Photocatalytic Water Splitting Driven by Surface Plasmon Resonance. ChemPhotoChem 8: e202300049.
• LI L, CHEN X, XIONG X, WU X, XIE Z & LIU Z. 2021. Synthesis of hollow TiO₂@SiO₂ spheres via a recycling template method for solar heat protection coating. Ceram Int 47: 2678-2685.
• LI R, GAO T, WANG Y, CHEN Y, LUO W, WU Y, XIE Y, WANG Y & ZHANG Y. 2024b. Engineering of bimetallic Au-Pd alloyed particles on nitrogen defects riched g-C₃N₄ for efficient photocatalytic hydrogen production. Int J Hydrogen Energy 63: 1116-1127.
• LIU L, ZHANG X, YANG L, REN L, WANG D & YE J. 2017. Metal nanoparticles induced photocatalysis. Natl Sci Rev 4: 761-780.
• LV L, HAO YR, GUO N, SUN J, SONG T, XUE H, HUANG K & WANG Q. 2021. The Synergistic Effect of Oxygen Vacancy and Carbon Interface Engineering in Hollow Cerium Oxide to Achieve Enhanced Oxygen Reduction Performance. ACS Appl Energy Mater 4: 5339-5347.
• MARQUES FC, STUMBO AM & CANELA MC. 2017. Estratégias e Materiais Utilizados em Fotocatálise Heterogênea para Geração de Hidrogênio Através da Fotólise da Água. Quim Nova 40: 561-571.
• QIAN A ET AL. 2023. Ultrathin Pd metallenes as novel co-catalysts for efficient photocatalytic hydrogen production. Appl Surf Sci 618: 156597.
• RAO BG, SUDARSANAM P, NALLAPPAREDDY PRG, YUGANDHAR REDDY M, VENKATESHWAR RAO T & REDDY BM. 2018. Selective allylic oxidation of cyclohexene over a novel nanostructured CeO2-Sm₂O₃/SiO₂ catalyst. Research on Chemical Intermediates 44: 6151-6168.
• SHI Z, LIU Y, ZHANG Y, SUN J, ZHENG J, WEI C, DU W, LIU L & CHENG C. 2023. Designed synthesis of yolk-shelled NiCo₂O₄/MnCo₂O₄ hollow sphere with boosted performance for supercapacitors. Appl Surf Sci 611: 155758.
• SKORODUMOVA NV, SIMAK SI, LUNDQVIST BI, ABRIKOSOV IA & JOHANSSON B. 2002. Quantum origin of the oxygen storage capability of ceria. Phys Rev Lett 89: 166601/1-166601/4.
• SONG J & WU F. 2023. Ceria-based quantum dots/nanorods supported metallic cobalt for efficient photocatalytic hydrogen production. Int J Hydrogen Energy 48: 20705-20716.
• SOUSA R, DA SILVA J, COSTA J, DE MOURA C & DE MOURA E. 2022. Bimetallic Au-Pd/α-MoO₃ Catalyst with High Oxygen Vacancies for Selective Oxidation of Cinnamyl Alcohol. J Braz Chem Soc: 10.21577/0103-5053.20220091.
• STOBER W, FINK A & ERNST BOHN D. 1968. Controlled growth of monodisperse silica spheres in the micron size range. J Colloid Interface Sci 26: 62-69.
• TAHIR M. 2019. La-modified TiO₂/carbon nanotubes assembly nanocomposite for efficient photocatalytic hydrogen evolution from glycerol-water mixture. Int J Hydrogen Energy 44: 3711-3725.
• UDDIN I, PHAN NAN, LE THI HY, KIM H, WHANG D & KIM GH. 2023. MoTe₂-Based Schottky Barrier Photodiode Enabled by Contact Engineering. ACS Appl Nano Mater 6: 445-452.
• WANG D, YIN FX, CHENG B, XIA Y, YU JG & HO WK. 2021. Enhanced photocatalytic activity and mechanism of CeO₂ hollow spheres for tetracycline degradation. Rare Metals 40: 2369-2380.
• WANG S, ZENG B & LI C. 2018. Effects of Au nanoparticle size and metal-support interaction on plasmon-induced photocatalytic water oxidation. Cuihua Xuebao/Chinese Journal of Catalysis 39: 1219-1227.
• XUE L, ZHANG C, WU J, FAN QY, LIU Y, WU Y, LI J, ZHANG H, LIU F & ZENG S. 2022. Unveiling the reaction pathway on Cu/CeO₂ catalyst for electrocatalytic CO₂ reduction to CH₄ Appl Catal B 304: 120951.
• XUNWEN S, LIQUN Z, WEIPING L, HUICONG L & HUI Y. 2020. The synthesis of monodispersed M-CeO₂/SiO₂ nanoparticles and formation of UV absorption coatings with them. RSC Adv 10: 4554-4560.
• YAN K, WEN C, LI R, ZHANG B, LIU T, LIU Q & ZHOU Z. 2023. Morphological optimized CeO₂ and Cu-doped CeO₂ nanocrystals for hydrogen production by solar photo-thermochemical water splitting based on surface photoinduced oxygen vacancies. Appl Surf Sci 636: 157779.
• YANG F, BAO X, LI P, WANG X, CHENG G, CHEN S & LUO W. 2019. Boosting Hydrogen Oxidation Activity of Ni in Alkaline Media through Oxygen-Vacancy-Rich CeO₂/Ni Heterostructures. Angew Chem Int Ed 58: 14179-14183.
• YASHIMA M, KOBAYASHI S & YASUI T. 2006. Crystal structure and the structural disorder of ceria from 40 to 1497 °c. Solid State Ion 177: 211-215.
• ZAMAN S, KHAN I, ZHANG FM, KHAN S, KHAN A, KHAN S, SADIQ S, RAFIQ M, SAGHIR S & SUN XJ. 2023. Synthesis of mediator free hollow BiFeO₃ spheres/porous g-C₃N₄ Z-scheme photocatalysts for CO₂ conversion and Alizarin Red S degradation. Mater Sci Semicond Process 162: 107534.
• ZHANG B, HUANG S, LI Y, SHEN J, TIAN X & DING M. 2025. Highly dispersed Cu on hollow spherical CeO₂: An efficient and stable catalyst for the RWGS reaction. Applied Catalysis B: Environment and Energy 366: 125003.
• ZHANG J, GONG M, TIAN C & WANG CA. 2016. Facile synthesis of well-defined CeO₂ hollow spheres with a tunable pore structure. Ceram Int 42: 6088-6093.
• ZHAO H, WANG C, ZHONG W, PENG S, SHU R, TIAN Z & CHEN Y. 2023a. Insights into morphology-dependent Au/TiO₂ catalyst in glycerol aqueous solutions towards photothermal reforming hydrogen production. Int J Hydrogen Energy 48: 15586-15599.
• ZHAO X, WANG S, YANG K, YANG X & LIU X. 2023b. Controlled gold-palladium cores in ceria hollow spheres as nanoreactor for plasmon-enhanced catalysis under visible light irradiation. J Colloid Interface Sci 633: 11-23.
• ZHU J, ZHANG G, XIAN G, ZHANG N & LI J. 2019. A High-Efficiency CuO/CeO₂ Catalyst for Diclofenac Degradation in Fenton-Like System. Front Chem 7: 495180.

Edited by

• Handling editor
  Maria Inês Tavares

Data availability

All data generated or analyzed during this study are included in this published article.

Publication Dates

• Publication in this collection
  18 Sept 2026
• Date of issue
  2026

History

• Received
  10 Jan 2026
• Accepted
  24 Mar 2026