Morphological Modulation of TiO₂ Nanotube via Optimal Anodization Condition for Solar Water Oxidation

Article information

Abstract

1. Introduction

As global energy demand rises, transitioning from fossil fuel-based energy sources to sustainable alternatives is crucial. In this context, photoelectrochemical (PEC) water splitting has emerged as a promising technology for producing hydrogen (H₂) that can reduce the reliance on existing energy sources. However, the current H₂ production mainly depends on gas reforming methods, contributing to CO₂ emissions and global warming [1–3]. On the other hand, H₂ production through solar water splitting is environmentally friendly, producing no harmful byproducts and causing minimal resource depletion. However, the process of solar-based water splitting into H₂ and O₂ requires substantial energy due to the need for redox reactions at each electrode, which are hampered by charge transfer overpotential as well as the fast recombination at each interface [4–6]. To address these challenges, developing efficient photoelectrodes to match the solar spectrum and finding suitable material combinations to suppress charge recombination is paramount.

Titanium dioxide (TiO₂) has long been recognized as a promising candidate for photocatalysis and PEC applications, including solar water splitting, due to its excellent stability and environmental benignity [7–11]. Despite these advantages, TiO₂’s application in PEC water splitting is limited by a large band gap, restricting light absorption to the ultraviolet (UV) range, representing only a small portion of the solar spectrum. In addition, the rapid recombination of photo-generated electron-hole pairs greatly reduces the quantum efficiency of the PEC process [12,13]. Accordingly, various material combinations and engineering approaches have been proposed to enhance TiO₂-based PEC performance.

Anodization has been identified as an effective method to improve the PEC performance of TiO₂ photoanodes, creating self-ordered TiO₂ nanotubes with one-dimensional (1D) nanostructures. This technique enhances charge separation and transfer efficiency at the electrode/electrolyte interface [14–16], significantly increasing the surface area and modifying the electronic properties of TiO₂, which are crucial for optimizing its functionality in PEC applications. For instance, research by Lin et al. highlights the fabrication of high specific surface area TiO₂ nanopowders through the anodization of porous titanium, leveraging TiO₂’s unique electronic and optoelectronic properties and catalytic capabilities essential for efficient water splitting [17]. Similarly, Q. Qian and colleagues demonstrated the potential of using internally anodized porous titanium to create immobilized TiO₂ nanotubes [18], which, with their high specific surface area, enhance electrochemical interactions within the PEC system, boosting overall photoelectrochemical activity. Sun and Yan’s study reveals how varying anodization voltage can alter the length and density of TiO₂ nanotubes, directly impacting hydrogen production efficiency in a two-compartment photoelectrochemical cell [19]. Another study by Zhong et al. shows that larger nanotube diameters, achieved through specific anodization conditions, can significantly enhance the surface area and, consequently, the photocatalytic activity of TiO₂ structures [20].

In this study, we explored the optimal anodization conditions for TiO₂ nanotube (NT) photoanodes to enhance their performance in PEC water splitting, making hydrogen production more efficient and effective. By thoroughly examining photocurrent density, electrochemical impedance spectroscopy (EIS), and Mott-Schottky analysis, we investigated the synergistic effects between anodization parameters and the PEC properties of TiO₂ photoanodes. Particular attention was given to the roles of electrochemically active surface area (ECSA), charge transfer resistance, and onset potential in determining the efficiency of the PEC water-splitting process. This comprehensive analysis provides insights into designing high-efficiency photochemical systems for hydrogen production and elucidates the underlying mechanisms contributing to the enhanced PEC activity of TiO₂ photoanodes, aiming to advance the field of sustainable energy research. This also supports the broader goal of reducing reliance on fossil fuels and mitigating the effects of climate change by transitioning to sustainable energy sources. The enhanced PEC activity of TiO₂ photocathodes developed in this study represents an important step forward in the field of solar-based hydrogen production and provides a promising pathway toward clean and sustainable energy solutions.

2. Experimental

2.1 Growth of TiO₂ nanotubes (NTs)

Initially, Ti plates were cleaned with acetone and washed with ethanol and deionized (DI) water to remove surface impurities. The plates were then dried using a nitrogen gas flow. The anodization process was conducted under various voltage conditions in an electrolytic solution comprising 97 wt% ethylene glycol, 3 wt% DI water, and 0.3 wt% ammonium fluoride. For the reference sample, a constant voltage of 60 V was applied for 20 min. For the TiO₂-6 sample, an on-off anodization procedure was employed: 60 V for 3 min (on), followed by 20 V for 6 min (off), repeated in 10 cycles. Similarly, TiO₂-12 samples were subjected to 60 V for 3 min (on), then 20 V for 12 min (off), repeated in the same cycles. Finally, the TiO₂-18 sample underwent 60 V for 3 min (on), followed by 20 V for 18 min (off), under identical cycles. After anodization, the samples were rinsed with DI water and air-dried. The voltage profiles for each anodization process correspond to TiO₂-ref, TiO₂-6, TiO₂-12, and TiO₂-18 samples, as briefly depicted in Scheme 1, with specific voltage patterns applied over time detailed in Fig. S1 (in supporting information).

Scheme 1

Schematic diagram to describe how to modulate the applied voltage to develop (a) TiO₂-ref, (b) TiO₂-6, (c) TiO₂-12, (d) TiO₂-18 film.

3. Results and Discussion

The surface and cross-sectional morphological properties of TiO₂ nanotube arrays were compared using FE-SEM images summarized in Fig. 1. All samples displayed a one-dimensional nanotubular structure. The control sample, TiO₂-ref, exhibited a pore diameter of approximately 53 nm and a length of about 8.5 m, with uniform nanopores through the surface area and no observed inter-spacing between the nanopores, indicating a strong interaction. Conversely, TiO₂-6, TiO₂-12, and TiO₂-18 nanotube arrays, fabricated by pulsed anodization through a repeated on (60 V)–off (20 V) cycle, showed segmented bamboo-like structures. Herein, the several key parameters to characterize the TiO₂ nanotubes can be defined. Thickness means the wall thickness of the nanotubes. Spacing indicates the distance between adjacent nanotubes. Pore size denotes the diameter of the openings at the top of the nanotubes. Segment length (node) is the length of each segment within the nanotube structure. To elucidate the differences in nanostructures induced by pulse duration, the pore size, segment length (node), and nanotube wall thickness were measured at 50 different points for each sample, and the average values were summarized in Table 1. Longer off-time anodization resulted in longer TiO₂ nanotubes, extended node lengths, and larger pore sizes. This bamboo-like TiO₂ nanotube structure enhances the active surface area compared to planar TiO₂ nanotubes, facilitating easier electrolyte insertion and improving charge carrier diffusion during the solar water splitting process. This quantitative analysis provides clear insight into the impact of pulse duration on the structural properties of TiO₂ nanotubes, confirming that anodization parameters significantly affect morphology and potentially the PEC performance of TiO₂ nanotube-based systems.

Fig. 1

FE-SEM images showing the cross-section and surface morphology of TiO₂ NTs, with (a, a′) cross-sectional views of TiO₂-ref, (a″) surface view of TiO₂-ref, (b, b′) side views of TiO₂-6, (b″) surface view of TiO₂-6, (c, c′) cross-sectional views of TiO₂-12, (c″) surface view of TiO₂-12, and (d, d′) cross-sectional views of TiO₂-18, (d″) surface view of TiO₂-18.

Table 1

Summary of the geometric parameters of TiO₂ NTs shown in Fig. 1

The crystal structure of TiO₂ photoanodes fabricated using various anodization methods was characterized through XRD analysis. The resulting XRD patterns, displayed in Fig. 2, confirm that all four samples comprise the ana tase TiO₂ phase [22,23]. Notably, prominent peaks were observed at 25.3° and 48.1°, corresponding to the major (101) and (200) planes of anatase TiO₂, respectively [24]. As the off-time during electrochemical anodization increased, the peak intensity of the (101) plane slightly increased, indirectly indicating a longer growth of the TiO₂ nanotube arrays. This result demonstrates that all samples possess the anatase TiO₂ crystal structure, confirming the successful synthesis of anatase phase TiO₂ photoanodes through anodization.

Fig. 2

XRD patterns of TiO₂ NTs, including * peaks from the Ti substrate.

The optical properties of the TiO₂ nanotube arrays were analyzed using UV-visible (UV-Vis) spectroscopy (Fig. 3), revealing that the strongest absorption occurred predominantly in the UV region. Specifically, light absorption began at 400 nm, with intensity steadily rising with longer off-time anodization. For the sample with an 18-min off-time anodization, a slight visible absorption around 420 nm was observed, potentially resulting from the formation of defect or trap-based sub-bands [25,26].

Fig. 3

Optical characterization of TiO₂ NTs: (a) absorbance spectra and (b) Tauc plots.

To compare the optical bandgap of each sample, Tauc plots were derived using the following equation (2) [27–29]. The TiO₂ bandgap was calculated to be approximately 3.2 eV. Compared to the control TiO₂ sample, the TiO₂ nanotube arrays produced through pulse anodization exhibited a slightly increased bandgap, likely due to the removal or passivation of defect or trap sites [30–32].

In conclusion, all samples demonstrated optical properties consistent with TiO₂ [29,33,34].

To calculate the Electrochemically Active Surface Area (ECSA), Cyclic Voltammetry (CV) measurements were conducted at scan rates ranging from 25 mV s⁻¹ to 150 mV s⁻¹, increasing in increments of 25 mV s⁻¹. The measurements were performed within a voltage range of –0.3V to 0.5V versus Ag/AgCl, being in the non-faradaic region and ensuring that the measured currents are primarily due to double-layer charging rather than faradaic processes [26,35,36]. These values were then converted to V versus RHE using the Nernst Equation (Eq. 1). Two key parameters, C_s and C_dl, are used to determine the ECSA. C_s represents the specific capacitance of an ideal flat surface of the electrocatalyst, with a commonly used mean value of 0.04 mF cm⁻² in calculations [37,38]. C_dl, the bilayer capacitance, is measured experimentally and indicates the charge storage capacity at the interface between the electrode surface and the electrolyte [39], excluding any Faradaic reactions. A higher C_dl value suggests a larger active surface area, reflecting the surface roughness or electroactive surface area of the electrode. In contrast, C_s provides a standard electric capacity per unit area, serving as a reference for the expected bilayer electric capacity of a known surface area under similar conditions. The ratio of C_dl to C_s allows for the estimation of the active surface area of the electrode, providing insight into the electrochemical performance of the TiO₂ photoanodes produced. This analysis highlights the significance of surface area and roughness in enhancing the electrochemical properties of electrodes for applications such as photocatalysis and PEC water splitting. The CV results are presented in Fig. 4 for various scan rates, and Fig. 5 depicts the calculated ECSA derived from Fig. 4 using Eq. 3 and 4 stated below [40]. CV measurements were repetitively conducted three times for each scan rate, and ΔJ/2 was calculated as the average value of three measurements.

Fig. 4

CVs of TiO₂ nanotube arrays at various scan rates from 25 mV s⁻¹ to 150 mV s⁻¹: (a) TiO₂-ref, (b) TiO₂-6, (c) TiO₂-12, and (d) TiO₂-18.

Fig. 5

(a) Half the difference in current density (J) at the center of the CV potential window vs. scan rates for TiO₂ nanotube arrays, and (b) calculated ECSA values.

The C_dl value of the TiO₂-6 nanotube arrays was the highest, at 0.241 F cm⁻², representing an increase of approximately 34.4 times compared to that of the TiO₂-ref photoanode (0.007 F cm⁻²). For TiO₂-12 and TiO₂-18 nanotubes, relatively lower C_dl values were observed compared to TiO₂-6 nanotubes despite the greater length of the TiO₂ nanotubes. This indicates that the bamboo-type nodes can significantly provide more specific active sites. Therefore, it can be confirmed that the sample with the shortest pulse length exhibits the largest surface area.

The comprehensive PEC performance of TiO₂ nanotubes is evaluated and illustrated in Fig. 6. Fig. 6a displays the chopped on-off photocurrent density as a function of the applied potential for TiO₂ nanotubes. The TiO₂-6 photoanodes achieve photocurrent densities up to 3.5 times higher than TiO₂-ref photoanodes. Furthermore, the onset potential (V_on) of bamboo-type TiO₂ nanotubes shifts negatively from 0.4 V_RHE to 0.2 V_RHE, possibly due to passivation from defect or trap sites in the interfacial region. This marked increase in photocurrent density and the negative shift in V_on for TiO₂-6 photoanodes underscore the critical impact of nano structuring and surface modification strategies on enhancing the PEC properties of TiO₂.

Fig. 6

Comprehensive PEC characterization of TiO₂ nanotubes: (a) photocurrent density measurements, (b) applied bias photon-to-current efficiency (ABPE) curve, (c) EIS Nyquist plots, and (d) Mott-Schottky analysis.

The ABPE was calculated to elucidate the photo-conversion efficiency of the TiO₂-based photoanodes, using Eq. 5 for this calculation [41,42]. The results are presented in Fig. 6b, highlighting the efficiency of each photoanode under applied bias versus RHE.

The TiO₂-6 photoanodes exhibit the highest ABPE, demonstrating their superior PEC activity. Specifically, at an applied potential of 0.7 V_RHE, TiO₂-6 achieved an ABPE of 0.43%, significantly surpassing other photoanodes. In contrast, the TiO₂-ref, TiO₂-12, and TiO₂-18 photoanodes displayed maximum efficiency at 0.35 V_RHE, with efficiencies of 0.16% and 0.19%, respectively. This indicates that TiO₂-6 not only enhances charge separation and transfer, as evidenced by its higher photocurrent density, but also converts a larger fraction of incident photon energy into electrical energy, underscoring its effective PEC water-splitting capability. Fig. 6c shows a Nyquist plot from EIS measurements, which is crucial for understanding the charge transfer properties of the TiO₂ photoanodes. These plots are essential for our detailed investigation to optimize key electrochemical kinetics to boost PEC water-splitting performance. The charge transfer resistance (R_CT) results are systematically summarized in Table S1, offering a quantitative comparison of the photoanodes. The TiO₂-ref photoanode exhibits a notably high resistance of 240,000 Ω cm⁻², indicating significant barriers to charge movement across the electrode/electrolyte interface. This high resistance likely leads to increased charge recombination rates, thus hindering overall PEC performance. In stark contrast, the TiO2-6 photoanode shows a much lower resistance of 2,615 Ω cm⁻², indicating enhanced electrochemical activity and improved charge carrier mobility, favorable for efficient PEC operations. Similarly, the TiO₂-12 demonstrates a resistance of 3,196 Ω cm⁻², showing better charge transfer efficiency than the reference model. Meanwhile, TiO₂-18 exhibits a resistance of 29,000 Ω cm⁻², lower than the reference but higher than that of TiO₂-6 and TiO₂-12, reflecting its intermediate electrochemical performance.

The semicircular diameters in the Nyquist plots are directly correlated with the charge transfer resistance (Rct) [43–45]. A clear trend of decreasing Rct is observed from the TiO₂-ref photoanodes to the TiO₂-6 photoanodes, signifying a systematic enhancement in electrochemical reactivity. The TiO₂-6 photoanode, in particular, exhibits the most favorable charge transfer conditions due to possessing the lowest Rct among the variants studied. Reducing charge transfer resistance is critical for facilitating efficient transfer from the photoanode to the electrolyte, which is essential for effective PEC water splitting. This enhanced charge transfer efficiency underscores the success of the modifications made to the TiO₂-6 photoanode, highlighting the potential for further optimizations in photoanode design to improve PEC water splitting.

The Mott–Schottky analysis was conducted under dark conditions concerning the applied potential (vs. RHE). Fig. 6d presents the Mott-Schottky plot, deriving the slope of the curves and calculating the charge carrier density (N_d) for TiO₂ nanotube photoanodes. The N_d was determined from the slope and intercept of the Mott-Schottky plot using the following Eq. 6 [46,47].

where ɛ_o is the permittivity of a vacuum, ɛ is the dielectric constant of the semiconductor, and kT/q is the temperature-dependent term. These key parameters, including the N_d and flat band potential (E_fb), are summarized in Table S2, providing a comprehensive overview of the TiO₂ nanotube photoanodes for PEC water splitting [48]. Specifically, the TiO₂-6 photoanode, with a slope of 9.07×10¹⁰, shows a higher N_d of 7.77×10¹⁹ cm⁻³ compared to the TiO₂ control sample. This increased N_d implies a higher concentration of charge carriers, which can significantly improve PEC performance by promoting more effective charge separation and minimizing charge recombination [36,49]. Furthermore, the TiO₂-6 exhibits a notably higher E_fb of 0.83 V, indicating a more favorable energetic alignment for charge transfer processes [36,50]. These N_d and E_fb changes reflect variations in PEC properties across the different TiO₂ NT photoanodes, directly influencing their efficiency in PEC applications.

4. Conclusions

This study thoroughly explains TiO₂ photoanodes fabricated under various anodization conditions to enhance PEC water-splitting efficiency. The 1D TiO₂ nanotubes were developed using a facile electrochemical anodization process by adjusting pulse anodization parameters. A bamboo-type nanostructure was formed by repeatedly applying an on-off voltage cycle of 60 V and 20 V. The 1D TiO₂ nanotube arrays exhibited different morphological, optical, and PEC properties depending on the off-time, which ranged from 6 min to 18 min. Notably, the TiO₂-6 photoanode, which had the highest ECSA, exhibited excellent PEC performance, achieving a J_ph up to 3.5 times higher, compared to the TiO₂-ref photoanode, with a negative shift in V_on of approximately 0.2 V. This result underscores the critical role of ECSA in enhancing light absorption, charge separation, and transfer at the electrode/electrolyte interface. EIS measurements further support this finding, showing that TiO₂-6 had the lowest charge transfer resistance, indicating better electrochemical reactivity. Additionally, the Mott-Schottky analysis offers valuable insights into the energy band alignment of the photoanodes, with TiO₂-6 displaying the highest carrier density. These results emphasize the significance of electrode surface properties and energy band positions, such as ECSA and charge transfer resistance, for high-performance PEC photoanodes. Furthermore, this study demonstrates that PEC water-splitting efficiency can be significantly enhanced through controlled anodization conditions, paving the way for the development of efficient and sustainable hydrogen production methods.

Supplementary Information

Article information Continued