Background

In recent years, TiO2 has been widely studied and applied in diverse fields, such as photocatalysis, dye-sensitized solar cell, self-cleaning surface, sensor, and biomedicine [16]. It is well known that TiO2 nanoparticles have the potential to remove recalcitrant organic pollutants in wastewater. However, it is prerequisite to produce immobilized TiO2 photocatalysts with highly efficient activity by scale-up methods. Recently, considerable efforts have been taken to use metallic titanium as the precursor to develop three-dimensional TiO2 films with controllable ordered morphologies, such as nanotubes [7], nanorods [8], nanowires [9], and nanopores [10]. The in situ-generated TiO2 films over titanium substrates possess such advantages as stable with low carbon residual, excellent mechanical strength, and well electron conductivity, which make them suitable to be used as electrodes for photoelectrochemical-related applications [6, 11]. Although a well-defined structural nanotube or nanoporous TiO2 film on metallic Ti can be synthesized by an anodic method [6, 7, 1013], it is still a big challenge to scale up the production of such TiO2 film due to the limitation of electrochemical reactor and the high energy consumption. Chemical oxidation methods by treating titanium substrates in oxidation solutions are more scalable for various applications. By soaking titanium substrates in H2O2 solution followed with calcinations, titania nanorod or nanoflower films can be obtained [8, 14]. However, the film always displays discontinuous structure with many cracks, and its thickness is less than 1 μm [8, 15]. Both of these would result in a low photoelectrochemical performance. With the addition of concentrated NaOH in the H2O2 solution, a porous nanowire TiO2 film can be achieved after an ionic exchange with protons and subsequent calcinations [9]. Employing NaOH and organic solvent as the oxidation solution and elevating the treating temperature, Ti substrate would completely transform into free-standing TiO2 nanowire membranes [16]. However, the disappearance of Ti substrate makes this membrane impossible to serve as an electrode.

Compared to titanium alkoxides or TiCl4, there are much fewer reports on the synthesis of TiO2 nanostructure with the precursor of TiCl3. Normally, anatase TiO2 film can be fabricated via the anodic oxidation hydrolysis of TiCl3 solution [17, 18]. Recently, Hosono et al. synthesized rectangular parallelepiped rutile TiO2 films by hydrothermally treating TiCl3 solution with the addition of a high concentration of NaCl [19], and Feng et al. developed TiO2 nanorod films with switchable superhydrophobicity/superhydrophilicity transition properties via a similar method [20]. Moreover, a hierarchically branched TiO2 nanorod film with efficient photon-to-current conversion efficiency can be achieved by treating the nanorod TiO2 film in TiCl3 solution [21]. However, all of these nanostructural TiO2 films from TiCl3 solution were grown over glass or alumina substrates. Fabricating nanostructral TiO2 films over metallic Ti substrates is a promising way to providing high-performance photoresponsible electrodes for photoelectrochemical applications. The obstacle for starting from Ti substrates and TiCl3 solution must be the corrosion of metallic Ti at high temperatures in the HCl solution, which is one of the components in TiCl3 solution. However, the corrosion could also be controlled and utilized for the formation of porous structures. According to reports, the general method to prepare nanoporous TiO2 film on Ti substrate is through anodic oxidation and post-sonication [10, 12]. In this contribution, we proposed a facile way to fabricate nanoporous TiO2 films by post-treating the H2O2-oxidized TiO2 film in a TiCl3 solution. The as-prepared nanoporous TiO2 film display homogeneous porous structure with enhanced optical adsorption property and photoelectrocatalytic performance, which indicates that the film is promising in the applications of water purification and photoelectrochemical devices.

Methods

Cleansed Ti plates (99.5% in purity, Baoji Ronghao Ti Co. Ltd., Shanxi, China) with sizes of 1.5 × 1.5 cm2 were pickled in a 5 wt% oxalic acid solution at 100°C for 2 h, followed by rinsing with deionized water and drying in an air stream. The nanoporous TiO2 film was prepared by a two-step oxidation procedure. Briefly, the pretreated Ti plate was firstly soaked in a 15 mL 20 wt% H2O2 solution in a tightly closed bottle, which was maintained at 80°C for 12 h. The treated Ti plate was rinsed gently with deionized water and dried. Then, it was immersed in a 10 mL TiCl3 solution (0.15 wt%) at 80°C for 2 h. Finally, the film was cleaned, dried, and calcined at 450°C for 2 h. The obtained nanoporous TiO2 film was designed as NP-TiO2. Two control samples were synthesized, including the one designed as TiO2-1, which was obtained by directly calcining the cleansed Ti plate, and the other named as TiO2-2, which was prepared by one-step treatment of the Ti plate in a TiCl3 solution.

The surface morphology of TiO2 films was observed using a field emission scanning electron microscope (SEM; Zeiss Ultra 55, Oberkochen, Germany). The crystal phases were analyzed using a powder X-ray diffractometer (XRD; D8 Advance, Bruker, Ettlingen, Germany) with Cu Kα radiation, operated at 40 kV and 36 mA (λ = 0.154056 nm). UV-vis diffuse reflectance spectra (DRS) were recorded on a Lambda 950 UV/Vis spectrophotometer (PerkinElmer Instrument Co. Ltd., Waltham, MA, USA) and converted from reflection to absorption by the Kubelka-Munk method.

Photoelectrochemical test systems were composed of a CHI 600D electrochemistry potentiostat, a 500-W xenon lamp, and a homemade three-electrode cell using as-prepared TiO2 films, platinum wire, and a Ag/AgCl as the working electrode, counter electrode, and reference electrode, respectively. A 0.5 M Na2SO4 solution purged with nitrogen was used as electrolyte for all of the measurements.

The photocatalytic or photoelectrocatalytic degradation of rhodamine B (RhB) over the NP-TiO2 film was carried out in a quartz glass cuvette containing 20 mL of RhB solution (C28H31ClN2O3, initial concentration 5 mg/L). The pH of the solution was buffered to 7.0 by 0.1 M phosphate. The solution was stirred continuously by a magnetic stirrer. Photoelectrocatalytic reaction was performed in a three-electrode system with a 0.5-V anodic bias. The exposed area of the electrodes under illumination was 1.5 cm2. Concentration of RhB was measured by spectrometer at the wavelength of 554 nm.

Results and discussion

Figure 1 shows the surface morphologies of films obtained by different procedures. The control sample TiO2-1 is obtained by the calcination of the pickled Ti plate at 450°C for 2 h. The typical coarse surface formed from the corrosion of Ti plate in oxalic solution can be observed (Figure 1A,B). By oxidation at a high temperature, the surface layer of titanium plate transformed into TiO2. However, the surface morphology shows negligible change. The film of TiO2-2, which is synthesized by directly treating the cleansed and pickled Ti plate in TiCl3 solution, displays smoother surface with no observable nanostructure (Figure 1C,D). Moreover, there are discernible TiO2 particles dispersing over the surface. It suggests that in the TiCl3 solution the surface morphology of Ti plate has been modified after dissolution, precipitation and deposition processes. By treating the H2O2 pre-oxidized Ti plate in TiCl3, the film displays a large-scale irregular porous structure, as shown in Figure 1E,F. Moreover, the appearance of NP-TiO2 film is red color (as inset in Figure 1F), which is different from the normal appearance of most anodic TiO2 nanorod or nanotube films [22]. The pores are in the sizes of 50 to 100 nm on the surface and about 20 nm inside; the walls of the pores are in the sizes of 10 nm and show continuous connections. Such hierarchical porous structure contributes to a higher surface area of the TiO2 film. Normally, titanium suffers from corrosion in the hot HCl solution, and the corrosion rate depends on the temperature and the concentration of acid. Without pre-oxidation, the surface layer of Ti plate is exposed to be etched and dissolved in the reaction solution at a medium temperature. Simultaneously, the TiOH2+ and Ti(IV) polymer generated by the hydrolysis of TiCl3 would precipitate and deposit over the surface (Equations 1 and 2) so as to retard the corrosion of Ti plate and avoid the completed dissolution of Ti plate [17, 19]. For the NP-TiO2 film, after the first step of oxidation in H2O2 solution, peroxo complexes coordinated to Ti(IV) have already formed, which cover most parts of the surface and be ready for further growth by the interaction with the oxidation hydrolytic products of TiCl3. However, it is also possible that HCl solution enters the interstitial of the TiO2 nanorod film and induces etching of the substrate Ti. At the experimental temperature, the dissolution of Ti is slow. With the reorganization of Ti(IV) polymer precursor, a porous structure forms over the Ti plate, as shown in Figure 1F.

Ti 3 + + H 2 O TiOH 2 + + H +
(1)
TiOH 2 + + O 2 Ti IV oxo species + O 2 TiO 2
(2)
Figure 1
figure 1

FE-SEM images of TiO 2 films over Ti plates. (A, B) TiO2-1, (C, D) TiO2-2, and (E, F) NP-TiO2 (the inset in (F) shows the digital picture of the NP-TiO2 film).

Figure 2A is the XRD pattern of NP-TiO2 film. The strong diffraction peaks at about 35.2°, 38.7°, 40.4°, 53.3°, and 63.5° can be assigned to the metallic Ti (JCPDS 44-1294). At the same time, the peak at 25.3° corresponds to the (101) plane of anatase phase TiO2 (JCPDS 83-2243). Diffraction peaks of rutile or brookite cannot be found, indicating that the titania film is composed of exclusively anatase. DRS spectra were measured to analyze the optical absorption properties of the films, as shown in Figure 2B. There is almost no optical adsorption for the TiO2-1 film, indicating that only a very thin layer of metallic Ti transforms into TiO2 after the calcination of Ti plate, and this contributes a poor photoresponse performance. TiO2-2 film displays a typical semiconductor optical absorption with the adsorption edge at about 380 nm, corresponding to the band gap of anatase TiO2. However, the absorption is relatively low, indicating that only few of TiO2 nanoparticles deposit over the surface of TiO2-2 film. The strong optical absorption appearing below 400 nm for NP-TiO2 film suggests a full growth of TiO2 layer over the Ti plate. Moreover, several adsorption bands centered at about 480, 560, and 690 nm can be observed in the spectrum of NP-TiO2 film. They possibly originated from the periodic irregular nanoporous structure. Such nanoporous structure is favorable to increase the photoresponsible performance, because the incident light that entered the porous structure would extend the interaction of light with TiO2 and result in an enhanced absorption performance, which can be observed in other nanotube or photonic crystal structural TiO2 films [22, 23].

Figure 2
figure 2

XRD pattern of NP-TiO 2 (A) and the DRS spectra of various films (B).

Using TiO2 films as the working electrodes in a three-electrode system, photocurrents under irradiation with full spectrum of light source were measured and compared, as shown in Figure 3. From the current transients (inset in Figure 3), all films show anodic photocurrents upon illumination, corresponding to the n-type photoresponse of TiO2. For TiO2-1 film, the initial anodic photocurrent spike is very strong and subsequently decays quickly. Simultaneously, a cathodic overshoot appears immediately when the light is switched off. The anodic current spike and cathodic overshoot are occasionally observed in many cases, and which is generally regarded as the indication of the surface recombination of photogenerated charges [2426]. A decay of anodic current immediately after the initial rise of the signal when the light is switched on is attributed to photogenerated electron transfer to the holes trapped at the surface states or the intermediates which originated from the reaction of holes at the semiconductor surface. With the accumulation of the intermediates, the electrons are trapped by the surface states, resulting in an anodic current spike. Owing to the same reason, the intermediates or trapped holes would induce a cathodic overshoot when switching off the light. The obvious strong spike for the illuminated TiO2-1 film suggests the slow consumption of holes and the corresponding oxidation process, which is related to the activity of the surface TiO2 layer. The poor crystallinity, large TiO2 particles, and the small amount of TiO2 in the directly oxidized film would result in the poor photoelectrochemical performance. However, the transient of NP-TiO2 film is different, displaying much smaller anodic current spike and more stable photocurrent. The photocurrent density is calculated as the difference of the current density upon illumination at the center time and in the dark, which is shown as a graph in Figure 3. NP-TiO2 film possesses the highest photocurrent density, which is about 1.2 mA/cm2, significantly higher than the corresponding TiO2-1 and TiO2-2 films. The efficient photoelectrochemical performance can be attributed to the porous structure of NP-TiO2 film, in which the interaction time between TiO2 and light would be increased due to the trapped photons inside the pores, corresponding to its enhanced optical absorption.

Figure 3
figure 3

A comparison of photocurrent density of various films. The inset shows a comparison of the current transients (applied potential: 0.2 V vs. Ag/AgCl).

The performance of the NP-TiO2 film was further tested by photoelectrocatalytic degradation of RhB solutions. The decolorization of RhB by photolysis is low, only 5.2% reduction observed after 2 h of irradiation (Figure 4). Without an applied bias, by illuminating the solution with the NP-TiO2 film, the decolorization efficiency only improved to about 11%. This low photocatalytic efficiency of the film could be attributed to the too small active area of the film and the phosphate in the buffered solution, which is regarded as the scavenger of radicals [27]. However, with a bias of 0.5 V vs. Ag/AgCl, the decolorization of RhB has been significantly improved, about 52.8% decolorization of RhB solution after 2 h of irradiation. Photoelectrocatalysis is a combination of photocatalysis and electrooxidation using the semiconductor films. By this method, an anodic bias on NP-TiO2 film is used to drive photogenerated electrons and holes moving toward different direction, so as to suppress the recombination and promote the organic degradation [11, 28]. Moreover, besides the improved optical absorption, the porous structure also contributes to a short diffusion path for RhB molecules to the active surface area. Therefore the NP-TiO2 film displays efficient photoelectrocatalytic activity for organic degradation. It can be expected that the chemical oxidation method for NP-TiO2 films is scalable for practical applications. With a larger active area, the NP-TiO2 film is potential to be used as an efficient electrode for energy conversion and organic pollutant removal.

Figure 4
figure 4

RhB decolorization as a function of time under various conditions.

Conclusions

A nanoporous TiO2 film on Ti substrate was synthesized by treating the initially H2O2-oxidized Ti plate in hot TiCl3 solution and followed by calcinations. The pre-oxidation in H2O2 solution is necessary to form such porous structure, indicating that the formation process is a combination of the corrosion of Ti substrate and the oxidation hydrolysis of TiCl3. The film possesses exclusively anatase phase and hierarchical porous morphology, with the diameter of the inside pores as small as 20 nm. The porous TiO2 film displays enhanced optical absorption, photocurrent generation, and efficient photoelectrocatalytic activity for RhB decolorization. The generated photocurrent density can reach as high as 1.2 mA/cm2. The chemical oxidation method for the nanoporous TiO2 film is possible to be scaled up and developed into a strategy to provide efficient TiO2 electrodes for diverse applications.