Promoted photocatalytic hydrogen evolution over core-shell-structured Co3O4/TiO2 under visible light

Graphical Abstract

1. Introduction

Hole scavenger-assisted hydrogen generation via photocatalysis could be a promising route for clean energy production [1-4]. Numerous researches have been and are still ongoing on H₂ generation through water splitting using photocatalysis because of its sustainability and environmental benefits [3–9]. The water-splitting process takes place by the action of a semiconducting photocatalyst, in which the conduction band (CB) edge will have a negative redox potential for H⁺/H₂ half-reaction (0.0 V vs. NHE) and the valence band (VB) edge will be more positive to the redox potential for O₂/H₂O (1.23 V) half-reaction. TiO₂ has been proved to be a good contender for this type of H₂ generation process due to its stability, low cost, chemical corrosion resistance, and non-toxic nature compared to various oxides [10–13]. However, yields of H₂ over TiO₂ are still low for industrial application due to various reasons such as i) the overpotential for H₂ evolution on TiO₂, ii) rapid recombination of H₂ and O₂ to form H₂O, and iii) the fast recombination of the charge carriers in the reaction vessel [10,14]. To overcome these limitations, alternative approaches can be applied to increase the photocatalytic efficiency of the system. Some approaches include, application of p–n heterojunctions [15-17], doping the semiconductor with a cocatalyst [12,18–20], the addition of sacrificial components, [21], and the application of a Z-scheme combination [22, 23]. For example, metals have been attached to the TiO₂ exterior to develop H₂ generation yield by reducing recombination processes [10]. The presence of a metal nanoparticle, such as Pt, causes the photo-electrons to occupy a Fermi level (E_f) lower in value than that of the CB of TiO₂ subsequently enhancing the charge separation [24,25]. In general, noble metals are helpful because of their higher work function and efficient charge separation in enhancing the photoefficiency of the system [11,15,16,26]. Pt is an extremely effective cocatalyst for H₂ generation upon doping on TiO₂ [27]. Transition metals have also been utilized as cost-effective cocatalysts for visible light response [28, 29]. A major aspect that enhances the photocatalytic ability of the photocatalyst is its specific surface area. High surface area photocatalysts provide adequate responsive spots for photocatalytic progression [30]. The mesoporous hollow-structured spheres represent a morphology that is efficient in light cultivation in the photocatalytic process [31-41].

The Co₃O₄-TiO₂ p-n heterojunction is typically used in photocatalytic water splitting purposes [42-44]. However, the generated hydrogen from these reported structures was comparatively lower than expected (5000 µmolg⁻¹ compared to 22000 µmolg⁻¹ in this work) [43], this could be mainly due to the relatively low specific surface area (40~120 m²g⁻¹) [44-46]. Hollow-structured core-shell photocatalysts, that are usually synthesized via template-assisted methods exhibit higher surface area [47], lower bandgap [48], and controllable shell/core structure [49]. Additionally, cobalt oxide cocatalysts demonstrate an efficient hydrogen evolution compared to other metal supports [42, 43, 50, 51].  In our method of synthesis, we used Poloxamer 407 as a hydrophilic non-ionic surfactant to obtain a high surface area SiO₂ template to guide structuring to increased surface area (400 m²g⁻¹) of hollow TiO₂ nanosphere for the first time. Thus, the combination of p-n junction Co₃O₄@TiO₂ hollow sphere structures demonstrate a significant bandgap tuning due to the controllable Co₃O₄ shell thickness as well as the high surface area (430 m²g⁻¹) that increase the possible number of photoactive sites available for the oxidation-reduction reactions.  The solution process offers an easy step for scaling up the preparation of such photocatalysts compared to the complicated routes presently used e.g. atomic layer deposition [45].

2. Experimentation

2.1. Chemicals

Ti(OC(CH₃)₃)₄, cobalt nitrate hexahydrate, Ammonia solution, C₂H₅OH, HCl, CH₃COOH, hexadecyltrimethylammonium bromide (HTAB), and poloxamer 407 were gained from Sigma-Aldrich.

2.2. Preparation of SiO₂ hollow nanospheres

The SiO₂ hollow nanosphere was formed by the sol-gel technique in the occurrence of poloxamer 407 and hexadecyltrimethylammonium bromide (HTAB) as surfactants according to the following steps: 20 mL of deionized water, 40 mL ethanol, 1.5 g hexadecyltrimethylammonium bromide, 4.5 g of poloxamer 407, 2.35 mL of acetic acid and 0.75 mL of HCl were mixed and the resulting mixture was stirred at 27^○C for 60 min. The 0.4 mL of TEOS was then added to the solution and the obtained combination was stirred for 60 min. Ethanol was evaporated at 40 ^○C for 24 h and the SiO₂ hollow nanospheres were gained.

2.3. Preparation of Hollow mesoporous TiO₂ nanosphere

0.5 g of the formed SiO₂ nanospheres was dissolved in 40 mL of ethyl alcohol.  Then, a mixture of 0.4 mL ammonium hydroxide solution and 0.4 mL hexadecyltrimethylammonium bromide was added to the above suspension with stirring for 15 min. After that, 0.4 mL of titanium sec-butoxide was then introduced and the stirring was kept for another 30 min in ambiance. To produce the hollow mesoporous TiO₂ nanospheres, the SiO₂@TiO₂ sample was dehydrated in the air at 40 ○C for 24 h and then calcinated for 4 h at 500 ○C using a 1^○C/min heating rate. Hollow mesoporous TiO₂ nanospheres were collected by etching SiO₂ core in NaOH aqueous solution at 90 °C.

2.4.  Preparation of Co₃O₄(shell)@TiO₂(core) hollow mesoporous nanospheres

Hollow mesoporous nanosphere Co₃O₄@TiO₂ nanocomposites were prepared as in the following steps: 0.5 g of hollow mesoporous TiO₂ nanosphere was dispersed in 40 mL deionized water and 10 mL of acetic acid. Then, 0.01 g of cobalt nitrate was added and the obtained mixture was left at room temperature for 60 min. The hollow mesoporous nanospheres Co₃O₄@TiO₂ were finally produced by dehydrating the mixture at 40 ^○C for 24 h and then calcination in the air for 4 h at 400 ^○C using 1^○C/ min heating rate. This method with repeated three times to prepare different weight percent of   x% Co₃O₄ to TiO₂ where x=0, 1, 2, 3, and 4 wt.%.

2.5. Characterization

The structure morphology for hollow nanospheres of TiO₂ and TiO₂/Co₃O₄ nanocomposites was investigated using a JEOL-JEM-1230 transmission electron microscope (TEM) and a JEOL-JSM-5410 scanning electron microscope (SEM).  The crystalline structure of x%Co₃O₄@TiO₂ nanocomposites was obtained by Bruker axis D8 X-ray diffractometer utilizing Cu Kα radiation (λ=1.540 Å). The N₂ adsorption/desorption isotherms were observed at 77 K by a Chromatech apparatus (Nova 2000 series) after degassing at 150 °C. The elemental analysis was investigated through the core-level X-ray photoelectron spectra (XPS) measurements via K-ALPHA spectrometer (Thermo Scientific). The diffusive reflectance (DRS) of the obtained materials was chronicled at room temperature by JASCO a V-570 spectrophotometer. The E_g was calculated from the DRS by the Tauc formula. The photocharge recombination studies were investigated through photoluminescence (PL) spectra of gained structures via RF-5301 fluorescence spectrophotometer (Shimadzu). The vibrational spectra of the obtained samples were studied via a Perkin-Elmer spectrophotometer at a resolve of 4.0 wavenumber FTIR spectrometer in the range 4000-400 cm^-1. Raman depiction was done by Horiba Lab RAM instrument applying 523.5 nm from Ar ion laser. Finally, Zahner Zennium electrochemical workstation was cast to measure the photocurrent intensity and to measure transient photocurrent.

2.6. Photocatalytic H₂ generation

The photoactivity of the Co₃O₄(shell)@TiO₂ (core)nanospheres for H₂ generation was examined in a 250 mL photocell having a water circulator system. A 500 W Xe light source with a cutoff filter (3O₄@TiO₂ nanospheres by photoreduction during the photoreaction of H₂ generation. Typically, 26 µL of 50 mM of hexachloroplatinic acid was added to the photocell. Before the photoreaction, Ar gas was bubbled for 15 min to eradicate oxygen. The illumination period for the photoreaction was 9 hours. The H₂ progression was followed by the Agilent GC 7890A gas chromatograph.

3. Results and Discussion

The XRD diffractograms of the obtained samples with various contents of Co₃O₄ are described in Fig. 1. The produced TiO₂ and x%Co₃O₄@TiO₂ nanocomposites were mainly of the anatase phase. The characteristic diffraction positions at (2Ө) =25.2°, 37.8°, 48.0°, 53.7°, 54.9°, and 62.2° are in typical coincidence with the (hkl) index planes represented in Fig. 1 [52].

Fig. 1   X-ray diffraction patterns of samples at different wt% of Co₃O₄ source at 0, 1,2,3, and 4 as represented by a, b, c, d, and e, individually.

It should be noticed that the main (101) intensity was reduced by incorporating the 1~4 wt.% of the Co source. There was no indication of any other impurity phases in all samples.Fig. 2 shows the N₂ adsorption/desorption isotherms of the hollow-structured TiO₂ with 1 and 3 wt.% Co₃O₄ as indicated. The isotherms unveiled the characteristic H3 hysteresis type IV loop, [53]. This feature suggests a mesoporous structured composite [54, 55].

Fig. 2 Nitrogen adsorption/desorption isotherms of selected hollow-structured TiO₂ (black) compared to 1.0 and 3.0 wt.% Co₃O₄@TiO₂ nanocomposites (red and blue).

The specific surface texture in terms of surface area and volumetric measure of pure TiO₂ are 401 m² g⁻¹ and 0.490 m³ g⁻¹, correspondingly. the surface texture parameters are progressively augmented by the addition of Co₃O₄-coated the hollow nanosphere (Table 1).  The 4 wt.% Co₃O₄/TiO₂ nanocomposite displayed a 430 m² g⁻¹ of surface area and 0.580 m³ g⁻¹ for the pore volume [52, 56]. The surface characteristics of all the synthesized nanocomposites with varying Co₃O₄ are presented abridged in Table 1. The characteristic structural surface of the Co₃O₄@TiO₂ nanocomposites is predicted to indorse the H₂ evolution.

The morphological structures of the produced samples are depicted through the SEM and TEM images as presented in Fig. 3 and Fig. 4, correspondingly. The SiO₂ template is seen in Fig. 3A with a diameter of about 120 nm showing a flat spherical exterior. While the produced TiO₂ spheres are also extant a flat superficial with little increase of size (130 nm) as seen in both Figs. 3b and 4A. The hollow-mesoporous TiO₂ spheres with a bumpy shallow seen with a similar diameter pure TiO₂ (Fig. 3C, Fig. 4B) upon loading with 1% of Co. The presence of rough nanoparticles is referred to as the small loading of the Co₃O₄. As presented the Co₃O₄ nano-shell with ~20 nm is composed of small nanoparticles.

Figs. 3D~F and Figs.4 C~F are showing that the surface of the hollow-structured TiO₂ spheres having a rough decoration of Co₃O₄ nanoshells. Fig. 4E displays the lattice parameter of the shell and the core at 0.25 and 0.34 nm, which is credited to the (311) plane of Co₃O₄ and the (101) plane of anatase TiO₂ [29,57]. The thin-layered structure of Co₃O₄ is composed of small flake-like particles covering the TiO₂ hollow spheres. This designated structure can offer super active sites for photocharge conduction upon light illumination. Furthermore, the Co₃O₄ (shell)@TiO₂(core) exhibits a close heterojunction interface indicating an improved subsequent photoactivity.

The elemental and chemical composition of the prepared photocatalyst was revealed through XPS analysis of the selected 3%Co₃O₄@TiO₂ sample as displayed in Fig. 5. The Ti 2p band exposed in Fig. 5a discloses two core peaks at 458.2 and 464.1 eV for the chemical states of Ti³⁺ and Ti⁴⁺ [57]. The Co 2p XPS core level (Fig 45B) shows two doublets at 796.5 eV (2p_1/2) -780.2eV (2p_3/2) and 789.0 eV(2p_1/2)-803.6 eV(2p_3/2) assigned to Co³⁺ and Co²⁺ states, individually.

The cohabitation of the Co²⁺ and Co³⁺ species is also confirmed from the spin-orbit splitting of 14.6 eV and the satellite peaks around 790 and 805 eV [58]. The co-occurrence of Co³⁺ and Co²⁺ agrees with the existence of Co₃O₄ on the exterior of the TiO₂/Co₃O₄-3wt% nanocomposites. As well, the O 1s band in Figure 5C shows one peak at 531.4 eV, which is deconvoluted to the oxide-structured TiO₂ or Co₃O₄ (531.4 eV), and the ^-OH assemblies adsorbed onto the sample’s surface (532.6 eV) [57,58].

Supplementary investigation for the structure of the x%Co₃O₄@TiO₂ was attained by Raman spectroscopy as in Fig. 6A. The Raman spectra present discrete bands situated at 145, 420, and 516 cm^-1, which signify the vibrant Eg, B1g, and A1g modes of the anatase phase [59]. The addition of the Co₃O₄ shell into the TiO₂ did not alter the Raman spectra except for the intensity of the vibrational modes. The Raman bands of Co₃O₄ lies in the same range of the anatase one as 146 cm^-1 ascribed to Co lattice vibrations are overlain with the TiO₂. The observed Raman bands for the Co₃O₄ at 147, 387, and 515 cm^-1 for tetrahedral F2g symmetry of CoO₄ according to the literature [60]. The functional groups within x% Co₃O₄@TiO₂ hollow spheres were analyzed via FT-IR spectroscopy as in Fig 6B. The spectra display discrete bands at 3565, 3385, 2340, 1627, 1498, 1341, 775, and 610 cm^-1 which are all indicating the typical spectra of functionalized TiO₂ [51]. The band located at 1627 cm^-1 the broader centered at 3385 cm^-1 are linked to chemisorbed or physisorbed water molecules [61]. Also, the wideband located at 3565 cm^-1 could be ascribed to the –OH group. While the 2340 cm^-1 vibration belongs to the physical attachment of CO₂. The lower intensity features around 1341 and 1498 cm^-1 may be ascribed to superficial carbonate type designed by the presence of Co₃O₄@TiO₂ in ambiance [60,62,63]. Lastly, the extensive band at 610~775 cm^-1 is situated within the Ti-O-H bending mode [63].

Fig. 3   SEM images of pure SiO₂ hollow spheres (A) and x% Co₃O₄@TiO₂ nanocomposites at x=0.0 (B), 1.0 (C), 2.0 (D), 3.0 (E), and 4.0 (F).

Fig. 4   TEM images of pure TiO₂ hollow spheres (A) and x% Co₃O₄(shell)@TiO₂(core) nanocomposites at x=1.0(B), 2.0 (C), 3.0 (D) and 4.0 (F). The high-resolution TEM image of (D) showing lattice parameters for Co₃O₄ and TiO₂ represented in (E).

Fig. 5   High-resolution XPS of 3.0Co₃O₄@TiO₂ nanocomposite showing Ti2p (A), Co2p (B), and O1S in (C).

Fig. 7A displays the UV-vis DRS of the x%Co₃O₄@TiO₂ nanocomposites compared to the bare mesoporous TiO₂ hollow spheres. The bare TiO₂ exhibited a sizable enhancement of efficiency in light-harvesting by the introduction of Co₃O₄ nanoshells. The optical density in the visible range is enhanced as well. Thus, the close contact between the p-type Co₃O₄ and n-type TiO₂ amended the interfacial band edges resulting in the acceleration of photocharge production [64]. The estimated E_g of mesoporous TiO₂ is 3.43 eV, in contrast, the Co₃O₄@TiO₂ provides a wider light absorption capability due to its narrower E_g of ~2.57 eV (Table 1, Fig. 7B). The reduction of  E_g and the upsurge of the visible-light absorption for the Co₃O₄@TiO₂ nanocomposites might be clarified by the impacts of nanoscale surface plasmon resonance [65,66] or the charge allocation in optical transitions between the TiO₂ core and the Co₃O₄ shell.

Photocatalytic H₂ evolution

The p–n heterojunction is fabricated by the amalgamation of p-type Co₃O₄ nanoshell and n-type TiO₂ core fashions in an electric field with band orientation. These core-shell nanocomposite hollow spheres powerfully ease the parting of photocharges and surging the photoefficiency [34,63,64]. The photocatalytic action of hollow-structured TiO₂ spheresandx%Co₃O₄@TiO₂ nanocomposites was utilized for H₂ generation under the illumination of visible light using glucose (10 vol% in H₂O) as a hole scavenger. The photosystem contains the H₂PtCl₆ deposits Pt nanoparticles on the surface of Co₃O₄@TiO₂.

As seen in Fig. 8A, if TiO₂ was only used as a sole photocatalyst, trivial H₂ generation was observed (~10 µmol/g, Table 1). The poor photocatalytic H₂ generation utilizing the only TiO₂ is due to the accelerated photocharge recombination and the large overpotential for H₂ generation. The x%Co₃O₄@TiO₂ heterojunction showed enhanced photoefficiency toward hydrogen evolution (Fig. 8a, Table 1). The mesoporous Co₃O₄@TiO₂ nanocomposites produced cumulative amounts of H₂ in an exponential trend through the photocatalytic reaction due to the higher photoactivity under visible light.

Fig. 6   Raman (A) and FTIR spectra investigation of samples of various wt.% of Co source at 0, 1,2,3, and 4 % as signified by a, b, c, d, and e, correspondingly.

Fig. 7   UV-vis DRS of hollow-structured TiO₂ spheres compared to Co₃O₄@TiO₂ nanocomposites as designated in (A). The assessed bandgap using the Tauc plot in (B).

Fig. 8   Photocatalytic hydrogen progress against illumination time by applying diverse Co₃O₄-loaded TiO₂ core-shell structures as indicated in (A). The impact of the photocatalyst amount of the best 3%Co₃O₄@TiO₂ photocatalyst is presented in (B).

The H₂ generation reached 18200 µmol/g after 9 h of light radiation by adding the Co₃O₄ nanoshell up to 3 wt%. The total H₂ was 1820 times superior to the bare TiO₂. The presence of Co₃O₄ leads to the creation of suitable valance and conduction band positions for H₂ formation reaction. Fig. 8B illustrates the impact of 3%Co₃O₄@TiO₂ concentration from 0.4~2.4 g/L on the photogeneration of H₂ within the 9h irradiation period. At the lowest dose of 0.4 g/L of 3%Co₃O₄@TiO₂nanocomposite, the photogenerated H₂ evolution was ~ 7700 µmol/g. By increasing the dose to 1.6 g/L, the photogenerated H₂ enhanced to 22400 µmol/g compared to 9250 and 18200 µmol/g for 0.8 and 1.2 g/L, singly.  The further increase of the optimized photocatalyst to 2.4 g/L reduced the amount of generated H₂ to 14700 µmol/g. Thus, the optimal dose of the 3%Co₃O₄@TiO₂ nanocomposite was fixed at 1.6 g/L. The possible reasons for dropping the H₂ photogeneration could be the lessening of photoactive sites or the inefficient photon scattering due to the opacity of 3%Co₃O₄@TiO₂ that inhibits light photons [3,15,26]. The existence of the mesostructured hollow 3%Co₃O₄@TiO₂ heterojunctions enhances light collecting and reflection, the diffusion of glucose molecules scavenges the holes, and the high surface area of the material result in the observed superior photocatalytic efficiency. The reusability of the spent Pt/3%Co₃O₄@TiO₂ photocatalyst was investigated in Fig. 9. The generation of H₂ evolution over the reused photocatalyst was slightly decreased keeping ~98% of the original H₂ amount after the fifth cycle.

To understand why the 3%Co₃O₄@TiO₂ isthe optimal photocatalyst, we investigated the PL spectra of x%Co₃O₄@TiO₂ compared to the pure TiO₂ hollow spheres as in Fig. 10A. The pure TiO₂ displayed a PL featurearound 388 nm with relatively higher power than other samples. Nevertheless, the PL features of x%Co₃O₄@TiO₂ nanocomposites unveiled a redshift to the wavelength of 557 nm for the 3%Co₃O₄@TiO₂ with the lowest intensity (table 1). The reduction in the PL signal in this sample specifies the movement of electrons from the CB of TiO₂ to the CB of Co₃O₄ [66].

Fig. 9 The constancy of the improved dose of 3%Co₃O₄@TiO₂ hollow core-shell-structured photocatalyst with maintainable hydrogen production level after five consecutive cycles.

The subordinate emission of the 3%Co₃O₄@TiO₂ is due to the recombination suppression of the photocharges upon irradiation. Afterward, the electrons easily move to bend the Fermi level to less value. This results in a higher reductive ability of 3%Co₃O₄@TiO₂. Concurrently, the modified trapped states imply an efficient electron-hole separation that subsequently enhances the evolution of H₂ [57, 66].These outcomes were further confirmed by measuring the photocurrent intensity during light irradiation, as shown in Fig. 10B. According to the photocurrent results, the photocurrents for x%Co₃O₄@TiO₂ were increased by increasing the Co₃O₄ nanoshells compared to a negligible photoresponse by only TiO₂. The intense photocurrent indicates that 3%Co₃O₄@TiO₂ possess the highest ability to transfer the photogenerated carriers upon illumination [67].

The proposed mechanism regarding the photocatalytic generation of H₂ utilizing Pt/ 3%Co₃O₄@TiO₂ is presented in Fig. 11. The Co₃O₄-coated TiO₂ hollow sphere’s surface works to reduce the recombination of the charge carriers by advancing the holes from TiO₂ to Co₃O_4­ and electrons to the photo-deposited Pt. The holes at the Co₃O₄ are captured by the glucose scavenger to produce protons. The electrons on the Pt surface can then oxidize the protons to produce H₂ on Pt particles [11,26,66]. The substantial narrowing of the E_g of the 3%Co₃O₄@TiO₂ heterostructures was a reason for the favorable response to light illumination. Furthermore, the modified energy levels due to the close connection of the core-shell structure functionalize the hole-trapping spots that balance the potential of the H⁺/H₂ reaction. This eventually moves the electrons to the Pt nanoparticle which meets the glucose dispersed through the pores of the Co₃O₄@TiO₂ heterojunction. the glucose itself, like a scavenging agent, eats the holes. Thus, the placid electrons by Pt are transported to H⁺ to form H₂ (Fig. 11).

Fig. 10   PL spectra (A) and transient photocurrent intensity (B) of pure TiO₂ hollow spheres compared to x%Co₃O₄@TiO₂.