Introduction

Hydrogen (H2) is a green source of energy with no emission of greenhouse gases offering high energy density (ca., 120 MJ/kg) and lower volumetric energy density (ca., 8 MJ/L)1,2,3,4,5,6. It can help to mitigate climate change due to greenhouse gases released from the combustion of fossil fuels7,8. It can be generated via several methods including water electrolysis7,9,10, gas reformation11,12, hydrolysis of hydrides13,14,15,16,17, and biomass gasfication18. Each of these methods exhibits pros and cons19,20,21,22. Overall, they can provide hydrogen with high purity, and techno-economical reasonable costs, and can be used for large production. However, most of these methods require non-renewable sources3. There are several challenges in hydrogen production including high cost, transportation/storage, and insufficient infrastructure23,24. The method for hydrogen production should be environmentally friendly during generation and use a renewable source with minimal energy requirements.

Photocatalytic water splitting is a promising method for hydrogen generation25,26. The topic was reviewed for titanium dioxide (TiO2) semiconductors in several reviews including Ref.27,28,29,30,31. The photocatalytic efficiency of TiO2 is low due to several challenges such as fast charge recombination of the photogenerated species. The optimization of key parameters such as co-catalyst types and loading percentage or procedure may improve the photocatalytic performance for TiO232,33,34. A study comparing different preparation methods of heavy metal co-catalysts for hydrogen generation is rare35,36.

Considering the previous points, this study shed light on the influence of type, loading percentage, and deposition method of metals (e.g., Ru, Co, and Ni) as co-catalysts on the photocatalytic performance of TiO2 for water splitting. Three metal types were investigated including ruthenium (Ru), nickel (Ni), and cobalt (Co). The metal ions were loaded into TiO2 using a loading percentage of 0.05–1 wt.% in the case of Ru and 0.1–1 wt.% in the case of Co and Ni. The loading procedure was investigated using three different procedures namely; impregnation (Imp), photocatalytic deposition (PCD), and hydrothermal (HT). X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), transmission electron microscopy (TEM), high-resolution TEM (HR-TEM), and UV–Vis diffuse reflectance spectroscopy (DRS) were used to characterize the materials. Under UV light irradiation, the hydrogen generation tests were carried out in a closed flow system with argon gas as the carrier gas and methanol serving as the sacrificial electron donor. There are high hydrogen generation rates (HGRs) in the photocatalysts. Electrochemical tests such as cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) were used to investigate the improvement's mechanism.

Experimental

Chemicals and methods

TiO2 was obtained from Acros Organics. Ethanol (99%), methanol (99%), isopropanol (99%), and metal precursors (RuCl3·2H2O, Ni(NO3)2·6H2O, Co(NO3)2·6H2O) were obtained from Fisher Scientific (UK, LTD).

TiO2-based nanocomposite preparation

Different loading percentages of metal e.g., Ru, Co, and Ni were loaded into TiO2 nanoparticles via three different methods; Imp, HT, and PCD respectively. The obtained powder will be denoted as xM/TiO2_y; where x, M, and y represent the weight percent, the metal loading, and the deposition method of the metal co-catalysts on TiO2, respectively. The three procedures can be described below:-

Incipient wet impregnation method (Imp)

TiO2 was suspended in 3 mL using an alumina crucible. After that, the mixture was stirred by a glass rod and ultrasonicated. The suspension was powdered by placing the crucible over a hot water bath with constant stirring till complete dryness. The powder sample was then moved to a tube furnace and purged by Ar (99.99%) using a flow rate of 100 mL min−1 for 15 min. After the purge, H2 (99.99%) gas flowed at a rate of 10 mL min−1, and the furnace temperature was raised to 200 °C (at a rate of 10 °C min-1) and held for 1 h. The powder was cooled naturally under the flow of H2 gas.

Hydrothermal method (HT)

A suspension of TiO2 powder (1 g) was prepared in 90 mL of distilled water and then sonicated for 10 min. For loading the required ratio of metal to TiO2, the studied volume of the metal precursor was mixed with 5 mL ethanol and then added dropwise to the TiO2 suspension, the reaction was maintained under vigorous stirring for one hour. The dispersion was heated at 180 °C for 12 h using a Teflon-lined stainless steel autoclave. The materials were collected and washed with water.

Photocatalytic deposition (PCD)

In the photo-deposition method, the reduction can be achieved via the photon energy-assisted methanol method. Briefly, the metal precursor was added to 200 mL of 20% methanol aqueous suspension containing TiO2 and then transferred into a Pyrex cell. After the purge with Ar gas, the cell was subjected to a UV-LED lamp for three hours. The precipitate was collected, washed using distilled H2O, and dried.

Photocatalytic H2 production experiments

The photocatalytic hydrogen generation was measured via an online system. As shown in Fig. S1, a Pyrex glass reactor connected to a flow system was set for photocatalytic evaluation of hydrogen gas. Typically, 50 mg of the photocatalyst was suspended in the reactor containing 200 mL of 20 vol.% methanol aqueous solution as a hole scavenger and subjected to constant stirring (800 rpm). Before irradiation, the suspension was ultrasonicated for 5 min and purged for 30 min with Ar gas. After that, the Ar gas flow rate was decreased to 10 mL min−1, and the suspension cell was situated at a 1 cm distance from the illumination UV–LED light source (25 W, 365 nm, NVMUR020A, NICHIA, Japan). The amount of gas evaluated was detected every 15 min using a gas-chromatography (GC) system (Shimadzu, GC-2014, Shin Carbon ST 80/100 with Length: 2 m and ID:2 mmID column, thermal conductivity detector, and argon gas as the carrier). The experiment lasted for 300 min at room temperature.

Photoelectrochemical measurements were performed using a potentiostat workstation (CorrTest® Instruments, model CS350) an electrolyte of Na2SO4 aqueous solution (40 mL, 0.1 M). The working electrode was prepared by casting the materials into FTO substrates with an active area of ca. 1.0 cm2. The counter and reference electrodes were a Pt wire and Ag/AgCl, respectively. Cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) measurements were recorded with and without light.

The working electrodes were prepared using the electrophoretic deposition method. Typically, the photocatalyst (20 mg) was dispersed in 1 mL isopropanol and sonicated for 15 min to make a homogenous slurry or paste. 50 µL of the slurry was deposited into an FTO conducting glass substrate that was then dried in air. The materials were deposited via layer-by-layer procedures. Finally, the electrodes were dried in an oven and calcinated at 120 °C for 60 min.

Characterization instruments

XRD patterns were collected using Bruker D8 advance equipment (Cu Kα radiation). XPS spectra were recorded using Al Kα radiation (Thermo Scientific, USA). TEM and HR-TEM images were collected using JSM-2100 (JEOL, Japan). The ultraviolet–visible (UV–Vis) DRS of powder samples were collected using an Evolution 220 spectrophotometer (Thermo Fisher Scientific, UK). Tauc’s plots were used to determine the materials band gap. Nitrogen adsorption–desorption isotherms were determined using Quantachrome (USA, at 77 K). Pore size distribution (PSD) was determined using the Barrett-Joyner-Halenda (BJH) method.

Results and discussion

Photocatalysts synthesis and characterization

Three different methods; namely Imp (Fig. S2), HT (Fig. S3), and PCD (Fig. S4), were used for the synthesis of metal-loaded TiO2 (Fig. 1). Metal ions of Ru3+, Co2+, and Ni2+ ions were used for loading offering composites of Ru/TiO2, Co/TiO2, and Ni/TiO2, respectively. The Imp method involves metal ions adsorption into the TiO2 substrate via sonication and stirring (Fig. S2). The metal ions were then reduced via hydrogen gas at 200 °C (Fig. S2). HT method includes the same procedure of loading, except that the metals loaded TiO2 were then subjected to hydrothermal treatment at 280 °C for 2 h (Fig. S3). Following the same loading procedure, PCD was performed (Fig. S4). PCD involves the reduction of the metal loading using a light source (Fig. S4). All procedures offer the synthesis of Ru/TiO2, Ni/TiO2, and Co/TiO2. The materials are fully characterized using XRD (Fig. S5), XPS (Fig. 2, Figs. S6, S7), TEM/HR-TEM (Fig. 3), N2 adsorption–desorption isotherms (Fig. S8), pore size distribution (PSD, Fig. S9), and UV–Vis DRS (Fig. 4, Figs. S10, S11). These techniques characterize the material's crystallinity, phase, oxidation state, particle size, morphology, porosity, pore size distribution, and optical properties including bandgap.

Figure 1
figure 1

Schematic representation for Imp, HT, and PCD methods used for loading Ru, Co, and Ni into TiO2.

Full size image
Figure 2
figure 2

XPS for Ru/TiO2, (a) survey, (b) O1s, (c) Ti2p, and (d) Ru3d.

Full size image
Figure 3
figure 3

TEM and HR-TEM images for Ru/TiO2 for loading methods of (a,b) HT), (c,d) Imp and (e,f) PCD.

Full size image
Figure 4
figure 4

(a) DRS spectra and (b) the band gap energies for 0.1 wt.%Ru_TiO2 (Imp), 0.3 wt.%Co_TiO2 (Imp), and 0.3 wt.%Ni_TiO2 (Imp) loaded into TiO2.

Full size image

XRD patterns for the materials were recorded to confirm the crystallinity and phase purities (Fig. S5). They display characteristic diffraction peaks at Bragg angles (2θ) of 25.2°, 36.1°, 48.0°, 53.9°, and 55.0° corresponding to the Miller indexes of (101), (004), (200), (105), and (211), respectively for anatase TiO2 (JCPDS No. 21-1271). The other peaks at Bragg angles of 27.4°, 36.1°, 41.2°, 54.3°, 56.6°, 62.7°, and 69.0° were assigned to (110), (101), (111), (211), (220), (002) and (301), respectively for Rutile TiO2 (JCPDS No. 21-1276). The TiO2 is a homophase junction of rutile and anatase phases. There is no strong diffraction of the loaded metals indicating that these species are amorphous or their loadings are undetectable in the XRD patterns. Thus, XPS for Ru/TiO2 was recorded (Fig. 2). XPS survey confirms the presence of Ti, Ru, and O at binding energies 458.3 eV, 280.2 eV, and 529.9 eV indicating the formation of Ru/TiO2 composite (Fig. 2a). Data analysis for Ru3d shows the peak at 280.27 eV corresponding to Ru0 indicating that the Imp method offers the synthesis of metallic Ru (Fig. 2d). XPS analysis was also performed for Co/TiO2 and Ni/TiO2 prepared by the Imp method (Figs. S6, S7). The Co2p3/2 peak in the XPS analysis at 780.9 eV corresponds to Co2+ in CoO (Fig. S6d). While XPS for Ni2p3/2 shows a peak at 857.2 eV corresponding to Ni2+ in NiO (Fig. S7d).

TEM and HR-TEM images for Ru/TiO2 were recorded for loading methods of Imp, HT, and PCD (Fig. 3). Two different particles are observed in TEM images corresponding to TiO2 and Ru (Fig. 3). Irregular particles of TiO2 with a particle size of 10–50 nm (Fig. 3). The particles of TiO2 can be confirmed from the lattice fringes observed using HR-TEM images (Fig. 3). TiO2 displays lattice fringes of 0.33 nm (101). The high electron density in Ru causes the observation of dark particles located in the plane of (101) for TiO2 (Fig. 3). N2 adsorption–desorption isotherms and PSD of the materials were recorded as shown in Figs. S8 and S9, respectively. Data analysis shows BET-specific surface areas of 70–203 m2/g with a pore size of 1.5–2 nm (Figs. S8, S9). It is important to mention that most of these porosities refer to the interparticle pores formed between metal-loaded and TiO2 crystals i.e., interparticle porosity according to TEM images (Fig. 4).

The optical absorption of the materials was determined using DRS (Fig. 4, Figs. S10, S11). The band gap values were determined using Tauc’s plots. UV–Vis spectra of all materials exhibit the characteristic absorption peak for TiO2 at maximum absorption at wavelength 300 nm (Fig. 4, Figs. S10, S11). TiO2 shows a bandgap of 2.97 eV (Fig. 4b). Transition elements cause a decrease in the bandgap values to a lower value of 2.6–2.8 eV. The drop in the bandgap of TiO2 indicates the formation of heterojunction between the transition elements and TiO2 nanoparticles. This effect can tune the material’s photocatalytic performance.

Photocatalytic water splitting

The photocatalytic activity of TiO2-based composite is tested. The effects of metal types (e.g. Ru, Co, and Ni), loading procedures (Fig. 5), and loading percentage (Fig. S12) were investigated. The HGR can be arranged in the sequence of Ru > Co > Ni (Fig. 5). HGR values for Ru/TiO2, Co/TiO2, and Ni/TiO2 were 23.91, 16.55, and 10.83 mmol/g h, respectively. Electron-rich elements such as Ru exhibit high HGR.

Figure 5
figure 5

Time course of H2 evolution over 50 mg samples of (a) 0.1 wt.%Ru_TiO2, (b) 0.3 wt.%Co_TiO2 and (c) 0.3 wt.%Ni_TiO2 photocatalysts with different metal loading methods, (d) shows the relation between initial H2 production rates over 50 mg of 0.1 wt.%Ru_TiO2 (Imp), 0.3 wt.%Co_TiO2 (Imp), and 0.3 wt.%Ni_TiO2 (Imp) nanocomposites relative to TiO2 in 20 vol% methanol aqueous solution subjected to UV_light for 5 h.

Full size image

The initial and cumulative hydrogen rate was observed for 0.1, 0.3, and 0.3 wt.% of Ru, Ni, and Co metals, respectively. High loading of these metals causes a decrease in the initial and cumulative rates. These observations could be due to the light block caused by high loading that prevents light radiation from reaching the external surface of TiO2 semiconductors.

The effect of the loading method was investigated using Imp, HT, and PCD (Fig. 5, Fig. S12). The methods of metal loading affect the distribution of the co-catalyst e.g. Ru, Ni, and Co. Thus, they affect the composite’s catalytic performance. Among the three loading methods, Imp exhibits a high catalytic performance (Fig. 5, Fig. S12). The high photocatalytic performance of the materials synthesized using Imp could be due to the homogenous distribution of the loaded metals into TiO2 according to TEM images (Fig. 3). The decrease of the bandgap of TiO2 using the Imp procedure is another explanation of this observation (Fig. 4). Further investigations were recorded using electrophotochemical measurements via EIS (Fig. 6) and CV curves (Fig. 7 and Fig. S13).

Figure 6
figure 6

(a–c) Nyquist plots of the EIS data in dark conditions for Ru_TiO2, Co_TiO2, Ni_TiO2 nanocatalysts of different metal loading methods and (d–f) represent Nyquist plots of the EIS data under UV light illumination in 0.1 M Na2SO4 electrolyte.

Full size image
Figure 7
figure 7

(a,b) Nyquist plots of the EIS data in (a) dark conditions and (b) under UV light illumination respectively, whereas (c) CV for 0.3 wt.%Ni_TiO2 (Imp), 0.3 wt.%Co_TiO2 (Imp), and 0.1 wt.%Ru_TiO2 (Imp) using 0.1 M Na2SO4 electrolyte.

Full size image

EIS spectra using the Nquists plot of all materials were recorded with and without light radiation (Fig. 6). Based on the analysis of Nquists plots, the small circle indicates low impedance i.e., high conductivity. All metal types i.e., Ru, Co, and Ni exhibit small circles for Imp compared to other loading procedures The same observation can be noticed without (Fig. 6a–c) and with light radiation (Fig. 6d–f). Furthermore, the size of all circles is smaller under UV radiation compared to without light.

Based on EIS analysis using Nyquist plots (Fig. 7), the co-catalysts of Ru, Ni, and Co exhibit small circles compared to bare TiO2. The presence of these cocatalysts increases the conductivity of the semiconductor TiO2. This observation were noticed without light radiation (Fig. 7a) and with light radiation (Fig. 7b). There is a dramatic decrease in the circle size indicating high conductivity for the composites compared to the bare TiO2. CV curve for bare TiO2 and Ru_TiO2 exhibit only cathodic reduction profile (Fig. 7c). On the other side, Co_TiO2 and Ni_TiO2 exhibit redox properties i.e. reduction–oxidation peaks. All prepared photocatalysts by the Imp method showed higher cathodic current than the others prepared by HT and PCD methods (Fig. S13).

The mechanism of H2 generation via water splitting requires a photocatalyst with a negative conduction band potential higher than the values for water reduction potential (− 0.41 eV). The photocatalyst should also display a positive valence band value than the water oxidation potential i.e., 1.23 eV. TiO2 semiconductor displays suitable band gap for water splitting. However, the high rate for the recommendation of the created electron–hole. Thus, bare TiO2 exhibits lower HGRs (Fig. 5). Moreover, we can improve the photocatalytic performance via the addition of co-catalysts such as Ru, Co, and Ni (Fig. 5). These co-catalysts improved also the conductivity (Figs. 6, 7) of pristine TiO2 enabling high electrochemical performance that can be reflected in high HGR values (Fig. 5).

A summary of some photocatalysts used for water splitting is tabulated in Table 1. Our study reported several parameters that can enhance the photocatalytic performance of well-known semiconductors i.e., TiO2. The co-catalysts properties such as metal types, loading, and deposition procedures play significant roles in the materials' performance. Our photocatalyst composites exhibit high HGRs compared to several reported materials (Table 1). Aqueous solutions of the chloride salts of Ru, Pd, and Ag were impregnated into TiO2 anatase with a content of 30 wt.% (Table 1)37. RuO2\TiO2 exhibited higher photocatalytic activity compared to PdO\TiO2 and Ag2O\TiO237. Ni\Pt@TiO2 (110) offered high photocatalytic HGRs from methanol alcohols under ultrahigh vacuum (UHV)38. Ni (2.21 wt.%)/Pt/black TiO2−x was synthesized via photodeposition (Table 1)39. The composite contains metallic Ni and Pt (i.e., 2 at.%)39. The material exhibited high charge separation efficiency. It showed HGRs of 69 and 3.1 mmol g–1 h–1 under UV–Vis and visible light, respectively (Table 1). The presence of Pt was essential to obtain metallic Ni instead of Ni(OH)2 (Table 1)39. The prepared photocatalysts in our study using inexpensive metals e.g., Co and Ni exhibit comparable HGRs to expensive co-catalysts such as Ru (Table 1). The high performance of these transition elements is mainly due to their high electrochemical performance. Co-catalysts such as Ni or Co nanoparticles enhanced the charge migration and separation of photogenerated species on TiO240. These metal species i.e., Ni2+ improved the electron transfer inside TiO2 due to their low potential (Ni2+ + 2e− = Ni, Eo = − 0.23 V)41. Thus, different forms of Ni were reported including Ni42,43, Cu–Ni44, Ni–Pd45, NiO46,47,48,49, and hydroxides (Ni(OH)2, Table 1)50. The presence of oxygen vacancy inside the composite improved the photocatalytic performance of TiO2. A study showed that CoO/h-TiO2 exhibited Z-scheme heterostructures with oxygen vacancy51. The authors observed that these vacancies were created during the composite formation and enhanced the photocatalytic H2 generation of TiO2 with an HGR value of 129.75 μmol h–1 (Table 1)51. However, the synthesis procedure required several steps and calcination at high temperatures. On the other side, our synthesis protocols are simple. Our catalysts exhibit high HGR values (Table 1).

Table 1 Summary of TiO2-based photocatalysts used for water splitting.
Full size table

Conclusions

This work presents an effective approach was reported to improve the photocatalytic performance of TiO2 nanoparticles for the production of hydrogen. The use of several metal co-catalysts e.g., Ru, Co, Ni applied in diverse ways (incipient wet impregnation, hydrothermal, and photo-deposition) resulted in a substantial enhancement in hydrogen generation compared to the pristine TiO2. Ru–TiO2, synthesized using the incipient wet impregnation method, exhibited the highest initial rate of hydrogen evolution (i.e., 23.9 mmol h−1 g−1), exceeding both Co and Ni counterparts. The results emphasize the crucial significance of the co-catalyst type, loading quantity, and deposition process in enhancing the photocatalytic efficiency of TiO2 for generating hydrogen using solar energy. Moreover, the similar performance of redox transition metals (Co and Ni) compared to the pricier Ru indicates potential opportunities for cost-efficient photocatalysts in the production of hydrogen.