Evaluation of Photocatalytic Hydrogen Evolution in Zr-Doped TiO₂ Thin Films

Abstract

Doping titanium dioxide has become a strategy for enhancing its properties and reducing its recombination issues, with the aim of increasing its efficiency in photocatalytic processes. In this context, this work studied its deposition over glass substrates using a sol–gel dip coating methodology. The effect of doping TiO₂ with Zirconium cations in low molar concentrations (0.01, 0.05, 0.1%) in terms of its structural and optical properties was evaluated. The structural characterization confirmed the formation of amorphous thin films with Zr introduced into the TiO₂ cell (confirmed by XPS characterization), in addition to increasing and defining the formed particles and their size slightly. These changes resulted in a decrease in the transmittance percentage and their energy band gap. Otherwise, their photocatalytic properties were evaluated in hydrogen production using ethanol as a sacrificial agent and UV irradiation. The hydrogen evolution increased as a function of the Zr doping, the sample with the largest Zr concentration (0.1% mol) being the most efficient, evolving 38.6 mmolcm⁻² of this gas. Zr doping favored the formation of defects in TiO₂, being responsible for this enhancement in photoactivity.

1. Introduction

It is known that titanium dioxide as a photocatalyst has been widely used in different environmental applications mainly due to its chemical stability, low cost, and its adequate band position to perform simultaneous red-ox reactions [1]. On the other hand, some of its optical properties make it an ideal candidate to be used as a thin film [2]. In this context, using it as a thin film can maximize light absorption because of the maximization of the surface area used in the material [3]. Unfortunately, one of its most common issues is the fast recombination of photo-generated charges after its illumination. For this purpose, many strategies have been carried out to avoid or minimize this phenomenon, improving its physicochemical properties to enhance its reaction yields.

Metal doping has resulted as an efficient alternative to achieve these goals; moreover, in most cases, its band gap energy can be reduced so it can be activated in the visible-light region. Many reports have shown the effect of metal doping in photocatalytic processes, highlighting the use of some elements such as Cu- [4,5], Ag- [6,7], Au- [8,9], Pt- [10,11] Ni- [12,13], among others.

In this context, Zr is a good candidate for doping TiO₂ lattices because both elements are in the same group, have the same valence state, and the anatase phase supports Zr incorporation, forming the solid solution Ti_1−xZr_xO₂ [14]. Some reports have studied the effect of Zr introduction into the TiO₂ anatase lattice on its physicochemical properties. For instance, Bolbol et al. [14] deposited Zr-doped TiO₂ thin films over glass substrates, varying the Zr dopant concentration from 0.5 to 10% mol using the sol–gel spin coating technique. They found changes in the structural and optical properties owing to a micro-strain increase created by Zr addition. This feature caused a reduction in electron–hole recombination compared to pristine TiO₂. Similarly, Juma et al. [15] and Oluwabi et al. [16] explored Zr-TiO₂ samples deposited by spray pyrolysis. In this context, Zr introduction in the TiO₂ lattice suppressed the anatase in the rutile phase transformation process, reduced the film roughness, decreased the film’s crystallinity, and strongly increased the dielectric constant. In photocatalytic applications, Mbiri et al. [17] evaluated the Zr dopant content effect on TiO₂ thin films in the degradation of persistent organic pollutants, such as Chlorisazon, Phenol, and 4-Chlorophenol, finding a reduction in the recombination rate and achieving degradation percentages higher than 80%. Similar findings were reported by degrading formaldehyde [18], methylene blue [19,20,21], 4-Nitrophenol [22], bismark brown red [23], methylene orange [24], and 4-chlorophenol [25]. On the other hand, fewer reports have been observed in the case of hydrogen evolution; for instance, Chattopadhyay et al. [26]. prepared Ti_1−xZr_xO_2-y nanocrystals in different Zr compositions. They found that Zr⁴⁺ incorporation into the TiO₂ lattice modified the surface chemistry, caused lattice strain and increased the amount of Ti³⁺ species that favored the electron transference, reaching a superior hydrogen production compared with that of pristine TiO₂. Some other works have reported the effect of Zr dopant in TiO₂; however, most of them focus on the use of powder particles [27,28,29,30].

Considering the above, this work reports the findings of the effect of Zr doping on TiO₂ thin films deposited over a glass substrate by sol–gel dip coating in terms of their optical and structural properties. Also, the impact of these changes on the photocatalytic efficiency in the hydrogen evolution reaction using ethanol as sacrificial agent is described.

2. Materials and Methods

2.1. Thin Film Deposition

Pristine and Zr-doped TiO₂ thin films were deposited by the sol–gel chemical method. In this context, a Titanium butoxide (97% Sigma Aldrich, St. Louis, MO, USA) 3 M solution was prepared in isopropanol (DEQ), adding the appropriate %mol of Zirconium butoxide (97% Sigma Aldrich) to obtain 0.01, 0.05, or 0.1 mol of Zr in the media. All solutions were deposited over glass substrates (previously washed in separate isopropanol-acetone-water washings) using a dip coating technique. For this purpose, the glass substrates were placed vertically in the system and immersed three times into the solution at a constant speed. After each immersion cycle, the solvent was evaporated instantly by subjecting the substrates to a hot temperature (165 °C). Additionally, the films were calcined at 400 °C for 2 h to promote phase crystallization.

2.2. Characterization

The structural characterization of the films was evaluated using an X-ray PANanalytical diffractometer with Cu Ka 1.54 Å radiation at a grazing incidence angle. Surface images were taken in ASYLUM RESEARCH MFP3D-SA AFM equipment in tapping mode. Transmittance spectra were obtained with a UV–VIS/Cary 5000, running the samples at a 200–800 nm wavelength interval, and PL spectra were measured in an Agilent Cary Eclipse (excitation wavelength—325 nm) fluorescence spectrophotometer. XPS measurements were analyzed using a VG Multilab 2000 (Thermo VG Scientific equipment, Waltham, MA, USA) with a monochromatic MG-Kα (1253.6 eV) irradiation source.

2.3. Photocatalytic Reactions

Photocatalytic hydrogen production reactions were performed under UV (254 nm) irradiation using a cylindrical Pyrex batch reactor. Four films (30 cm² of active area) were pasted inside the reactor, adding 200 mL of deionized water. The system was vented with Argon for 15 min to promote an anoxic media, and then a UV lamp was immersed through a quartz tube and turned on. Gas samples were taken every 30 min using a syringe and injected in a gas chromatograph Varian GP-3380 with a thermal conductivity detector using Argon as the mobile phase and RESTEK REST-19808 (RESTEK, Centre County, PA, USA) column as a stationary phase. The acetaldehyde concentration in the remaining reaction liquid was measured by High-Resolution Liquid Chromatography using a Shimadzu Nexcol C18 (Shimadzu, Columbia, MD, USA) column as the stationary phase, and a mixture of acetonitrile/water 45:55 solution as the mobile phase.

3. Results and Discussion

Figure 1 presents the XRD patterns of all the samples. As seen, the pristine TiO₂ and the Zr-doped films were amorphous, and no diffraction peaks were detected. Similar results have been reported in the literature for TiO₂ films [31,32,33,34], where peaks with very low intensities or no peaks are detected in similar sol–gel deposition conditions.

Figure 2 shows the elemental composition of the TiO₂ and Zr-doped TiO₂ films according to the XPS spectra for the C 1s, O 1s, Ti 2p, and Zr 3d levels. The C 1s core level has been deconvoluted into three peaks related to C-C, C-O-C, and O-C=O bonds (Figure 2a). On the other hand, the O1s spectra were deconvoluted into three curves for the TiO₂ and 0.01 Zr films and four curves for the 0.05 Zr and 0.1 Zr films. These curves correspond to the M-O, M-OH, C-OH, and C-O bonds [35,36], with only the 0.05 Zr and 0.1 Zr films showing the contribution of the C-OH species [35]. Due to the Zr being incorporated, the peak related to the M-O bond increased; likewise, the 0.01 Zr film is the only one that has less area in the peaks related to the M-OH and C-O bonds (Figure 2b).

Additionally, the Ti2p region is contributed to by the Ti³⁺, Ti⁴⁺, and Ti(OH)₂ (marked with *) species [36,37,38], while the Zr3d core level shows the presence of Zr⁴⁺ species [17,36]. In this context, the binding energy values obtained in the Zr 3d spectra are slightly lower to the ones reported in the ZrO₂ spectra, which suggests that the atoms are incorporated into the TiO₂ structure instead of in the ZrO₂ phase in low proportions [15]. Additionally, as the electronegativity of the Ti (1.54) is higher than that of the Zr (1.33), the peaks in the Ti2p and Zr3d spectra shifted toward a lower binding energy between the samples with Zr⁴⁺, which is related to the partial substitution of Ti⁴⁺ by Zr⁴⁺ ions [15,36]. This feature can be related to the formation of a Ti_1−xZr_xO₂ phase.

Figure 3 presents the AFM images taken in a contact mode. As seen, all films were deposited uniformly; however, the pristine TiO₂ sample presented the formation of fine particles with the appearance of some cracks potentially generated during solvent evaporation. In contrast, the incorporation of Zr promoted the formation of densely packed particles, and additionally it reduced the formation of cracks, promoting a good coverage of the substrate. According to some authors, Zr incorporation as a dopant can retard the TiO₂ densification, reducing the formation of cracks and pinholes in the layer [39,40]. On the other hand, similarity can be seen between the TiO₂, 0.01 Zr, and 0.05 Zr samples; however, a slight reduction in the surface roughness was observed (Figure 3). Similar findings were reported by Naumenko et al. [39], where low Zr loads promoted the formation of densely packed particles as a result of the reduced crystallization effect that Zr promotes, or the increase in the nucleation centers during the film growth which inhibits grain growth, as Juma et al. mentions [15].

On the contrary, a larger Zr concentration (0.1 Zr sample) promotes an increase in the film surface roughness; in this context, this drastic change can be associated with the formation of a more viscous sol due to the addition of a larger Zr precursor concentration [17]. An increase in the surface roughness can suggest an increase in the surface area, resulting in the presence of more active sites to enhance their photoactivity [41].

The transmittance percentage obtained from the deposited films is presented in Figure 4a. All the films are transparent in the visible light region (% T > 70%); however, when the Zr⁴⁺ concentration increases, the transmittance percentage slightly decreases, this behavior being related to high light dispersion over the film surface due to the increased roughness (more remarkable in the 0.1 Zr sample) [42,43]. The optical band gap of the films was calculated by the Tauc plot using the following equation [44]:

$$\left(\alpha h\nu \right)=A{\left(h\nu -E_g\right)}^n$$

where α is the absorption coefficient, $h\mathsf{\nu }$ is the photon energy, A is a proportionality constant, n is the Tauc exponent (n = 1/2 for direct transitions and n = 2 for indirect transitions), and $E_g$ is the band gap of the material [44,45]. Figure 4b shows a Tauc plot, where the band gap values slightly decrease with the incorporation of Zr (TiO₂ = 3.88 eV; 0.01 Zr = 3.81 eV; 0.05 Zr = 3.85 eV; 0.1 Zr = 3.76 eV). This decrease in the band gap value is related to shifts in the absorption edge toward a higher wavelength, which is due to an increase in the doping carrier concentrations. These carriers interact with free carriers and ionized impurities, causing a decrease in the band gap value [46]. In our films, as the Zr⁴⁺ concentration increases, the absorption edge shifts to higher wavelengths, decreasing the band gap value due to the increase in different impurities from the Zr⁴⁺. According to Bolbol et al., this phenomenon is known as the Burstein–Moss effect [14].

The PL spectra of the films excited with a wavelength of 320 nm are shown in Figure 4c. As seen, all the samples presented a broad emission between 390 and 540 nm. As this band displays a broad emission, this region could be deconvoluted into three bands, which are related to the TiO₂ band gap absorption (∼420 nm), self-trapped excitons in TiO₂ (∼450 nm), and oxygen vacancies of TiO₂ and of ZrO₂ and Zr⁴⁺ (∼430 to 530 nm) [47,48,49] (Figure 4d). On the other hand, from Figure 4c, there is an evident decrease in the PL intensity in the samples with a larger load of Zr⁴⁺; in this context, it is well known that the PL emission reduction is associated with a minimization of the e⁻/h⁺ recombination, which translates into a greater availability of free charges to carry out red-ox reactions [50]. Additionally, it is known that the doping process produces extra free electrons in the TiO₂ lattice, which also reduces the emission efficiency by creating a non-radiative channel [14].

The photocatalytic hydrogen evolution rates obtained from the deposited films using ethanol as a sacrificial agent are presented in Figure 5a. As seen, all the samples present photoactivity, which is enhanced because of the increase in the Zr⁴⁺ concentration, reaching almost double the TiO₂ production (22 mmolcm⁻²) in the sample loaded with 0.1%mol of Zr (38.6 mmolcm⁻²). In this context, this observed improvement in the photocatalytic activity can be related to different facts; for instance, a larger Zr concentration causes a greater roughness on the film surface, which, as mentioned previously, increases the surface area and the active sites where ethanol molecules can be adsorbed and react to produce hydrogen. Additionally, the 0.1 Zr sample presented reduced electron–hole recombination, evidenced by the PL analyses, mainly associated with the formation of energy levels below the TiO₂ conduction band as self-trapped excitons, and defects such as oxygen vacancies.

Figure 5b presents the possible mechanism of hydrogen production through ethanol oxidation. In this context, the use of this organic compound has been highlighted by its capacity to act as an electron donor to the conduction band, enhancing the hydrogen production compared with using pure water [29]. Additionally, the presence of the different defects can capture light-induced electrons; for instance, in the case of oxygen vacancies, they play an important role acting as electron (e⁻) traps, which, in consequence, avoid their recombination with the holes (h⁺) [51]. Moreover, Zr⁴⁺ doping also introduces alternative defect levels close to the conduction band of TiO₂, which also can act as electron trap centers because the presence of Zr metal is used to produce a Schottky barrier to facilitate electron capture [52].

On the other hand, ethanol is oxidated in the Zr-TiO₂ film valence band, forming acetaldehyde. The formation of this compound was evidenced in the remaining liquid of the reaction, detecting the following concentrations: TiO₂ = 750 μmol, 0.01 Zr = 627 μmol, 0.05 Zr = 612 μmol, 0.1 Zr = 530 μmol. Finally, the protons (H⁺) formed from ethanol oxidation react with the electrons (e⁻) accumulated in the formed defects and the conduction band to efficiently produce hydrogen.

Finally,

Table 1. Summary of the photocatalytic hydrogen production studies using TiO₂ thin films.

Photocatalyst	Dopant	Deposition Method	Illumination	Sacrificial Agent	Hydrogen Production	Ref.
TiO2	Cu-Ni	Drop casting	300 W Xe lamp	Methanol 25%	41,690 μmol/gh	[53]
TiO2	--	DC sputtering	UV 254 nm lamp	--	38 μmol	[54]
TiO2	--	Hydrothermal	UV 254 nm lamp	--	132 μmol	[55]
TiO2	--	RF sputtering	300 W Xe lamp	--	0.55 μmol/hcm2	[56]
TiO2	Au-Pd	Spin coating	300 W Xe lamp	Glycerol 5%	0.014 mL/min	[57]
TiO2	Ag	Sol–gel/dip coating	5000 W Xe lamp	KOH 1 M	580 μmol	[58]
TiO2	Ag	Drop casting	420 W Hg lamp	Water/methanol 1:1	148 μmol/gh	[59]
TiO2	Cr	RF sputtering	250 W W lamp	NaOH 1 M	24 μmol/h	[60]
TiO2	Ag	Hydrothermal	16 W Hg lamp	Ethanol 10%	8.1 μmol/cm2	[61]
TiO2	Pt	Dip coating	Black light lamps	Ethanol 50%	9 μmol/min	[62]
TiO2	Pt	RF sputtering Sol–gel/spin coating	250 W W lamp	NaOH 1M	12.5 μmol/h 4.3 μmol/h	[63]
TiO2	N-NiO N-CuO	Sol–gel/dip coating	UV 254 nm lamp	--	62,000 μmol/g	[64]
TiO2	Pt	Dip coating	13 W UV lamp	Water/methanol 1:1	349.6 μmol/gh	[65]
TiO2	Zr	Anodization method	500 W Xe lamp	Artificial sea water/ethilenglycol	15 μmol	[30]
TiO2	Zr	Sol–gel/dip coating	UV 254 lamp	ethanol	38,600 μmol/cm2	This work

presents a summary of the hydrogen evolution reaction via photocatalysis or photo-electrocatalysis using TiO₂ thin films deposited under different methodologies. In this context, some studies focus on analyzing the different deposition conditions related to the used methodology, while others analyze the effect of an added dopant concentration (added in situ or as a multi-layer). As seen, the hydrogen production values obtained in this work are comparable with those that present higher production values, which suggests that doping TiO₂ with this metal (Zr) is an alternative to enhance its photocatalytic performance.

4. Conclusions

Zr-doped TiO₂ amorphous thin films were successfully grown over glass substrates using the sol–gel dip coating methodology, varying the Zr concentration in the film precursor solution (0.01, 0.05 and 0.1 mol%). The incorporation of Zr⁴⁺ into the TiO₂ cell was confirmed using XPS, which resulted in the possible formation of the Ti_1−xZr_xO₂ phase. The increase in the Zr concentration caused changes in the structural and optical properties of the films such as an increase in the film roughness, and a slight reduction in the transmittance percentage and their band gap. All the films exhibited photoactivity, evolving hydrogen under UV irradiation and ethanol as a sacrificial agent, with the hydrogen accumulation being increased by almost double in the samples with the largest Zr load (0.1 Zr = 38.6 mmolcm⁻²) compared with that of pristine TiO₂ film. This enhancement in the photocatalytic activity was associated with the increase in the roughness of these films, which resulted in an increased surface area and was favorable for the reaction, and with the formation of different defects near to the TiO₂ conduction band that acted as electron trap centers and minimized the e⁻/h⁺ recombination rate.