In the past three decades, titanium dioxide (TiO2) and its nanocomposites have been widely investigated as promising photocatalysts [1-4]. Among various TiO2 nanostructures, TiO2 nanotube (NT) arrays synthesized by a simple electrochemical anodization provide the advantages of adjustable structures, high surface area, and excellent charge transport properties. Various applications such as photocatalytic decomposition, water splitting, photovoltaics, batteries, and sensors have been explored. TiO2 NTs have been found to have better photocatalytic properties of organic pollutants compared to commonly used TiO2 nanoparticles, although the NT surface area is much smaller [5-7]. This is due to the excellent light trapping, enhanced electron-hole separation, and much slower deactivation of NTs during the photocatalytic reaction . To further enhance the photocatalytic activity of NTs, various strategies have been used such as the improvement of morphology, crystal structure, and surface area [9,10], introduction of heterogeneous structures by decoration [11,12], and metal or non-metal doping [13,14]. Among these, the enhancement of NT crystallinity has been considered as an important and widely investigated approach to achieve better photocatalytic performances.
NT crystallinity can be simply enhanced by increasing the annealing temperature. More importantly, the stability of structures and geometries has to be considered during annealing. As to aligned NT array, the retention of its porous structure during annealing is essential to preserve its large specific surface area, which is desired for photocatalytic reaction. However, for NT arrays fabricated on metallic Ti foils, the substrate effect would result in crystallite growth in the tube walls during the high-temperature annealing process, leading to thicker tube walls and a significant decrease of the surface area. Therefore, the discussion of optimal annealing temperature and crystallinity is difficult. The highest photodegradation efficiency has been reported using NT films annealed at different temperatures such as 450°C , 500°C , 550°C , and 600°C . These earlier studies have concluded that the tubular structure was stable at or below the optimal temperatures. On the other hand, at higher temperatures, the NT structure collapses and undergoes deformation, dramatically decreasing the surface area and thus photocatalytic activity. Another disadvantage caused by the substrate effect is the phase instability of NTs under high-temperature treatment. TiO2 NTs with a stable anatase phase have excellent performance in photocatalytic and photoelectrochemical reactions [17,19-21]. However, after annealing of the as-formed NTs on their metallic substrate above approximately 550°C, a crystal phase transformation from anatase to rutile developed [15,22], which would play a role in lowering the photocatalytic performance.
The synthesis of free-standing NT films is a promising way to generate NTs without the presence of a Ti metal substrate, resulting in both structure and phase stability up to a high temperature of approximately 600°C  or 700°C , although the fabrication process is tedious and inefficient. By rapid breakdown anodization [25,26], NT bundles in powder form can be prepared. However, these bundles featured broken tubes and also contained impurities of remnants from the anodization electrolyte. Therefore, the bundles maintained stable phase and structure specifically at temperature <450°C. Therefore, the development of alternative methods is warranted. In the present study, we evaluated TiO2 NTs in powder form by dispersing as-anodized NT arrays. The morphology, structure, and crystal phase composition of the synthesized NT powders were characterized. The photocatalytic property of the powders annealed at different temperatures was also investigated to achieve optimized photocatalytic activity.
Preparation of the NT powders
The preparation details of anodic NT arrays were similar to our previous reports . Self-organized TiO2 NTs were grown on a Ti substrate (10 × 10 cm, 0.89 mm thickness, 99.7% purity, Alfa Aesar, Ward Hill, MA, USA). Then, the as-anodized irregular surface oxide layer was stripped off the Ti substrate by ultrasonication in deionized (DI) water. The cleaned Ti foil was anodized again for 2 h to grow aligned NT arrays in the same electrolyte. After the two-step anodization, ultrasonication in ethanol or DI water was carried out for 5 to 10 min to disperse the obtained oxide tube layer. Then, the Ti foil was reused for anodization formation of NTs. The above process was repeated until the Ti foil was completely consumed. The NT powder was collected by centrifugation and dried in a vacuum drying oven at 120°C overnight. The powder product was heat-treated at 450°C, 550°C, 650°C, and 750°C for 2 h in air to crystallize the NTs.
Characterization of the NT powders
The morphology of the resulting NT powders was characterized by a field emission scanning electron microscope (FE-SEM; Sirion 200, FEI, Hillsboro, OR, USA). The relative Brunauer-Emmett-Teller (BET) surface area was evaluated by adsorption-desorption isotherms using nitrogen gas at 77 K (ASAP 2010, Micromeritics, Norcross, GA, USA). Samples were degassed at 200°C for 4 h under high vacuum prior to measurement. The pore size distribution was analyzed by Barrett-Joyner-Halenda (BJH) adsorption differential pore volume. TiO2 paste was prepared by adding of the powder (20 wt%) to a mixture of isopropanol:n-butyl alcohol = 1:4 (v/v) solution. Thereafter, the paste was mixed under magnetic stirring for 24 h. The paste was deposited on fluorine-doped tin oxide (FTO) glasses by using a doctor blade technique and then air-dried, forming a porous NT film. The film was reinforced by annealing again at 450°C. The thickness of the NT layer was approximately 10 μm, which was controlled by tapes (Scotch Magic Tape, 3 M, St. Paul, MN, USA). The crystal structures were verified by X-ray diffraction (XRD; Cu Kα radiation, Rigaku 9KW SmartLab, Rigaku, Tokyo, Japan) patterns.
The photocatalytic properties of the samples annealed at four different temperatures were evaluated by the photodecomposition of methylene blue (MB). An NT-coated substrate with an active area of 9 cm2 was immersed in 100 mL of MB aqueous solution (initial concentration: 10 mg/L) for 1 h in the dark to reach the adsorption-desorption equilibrium at the TiO2 surface. Ultraviolet (UV) light (λ = 365 nm) was provided by a UV lamp assembly (8 W × 3) with the distance between lamps and the photocatalyst film of approximately 15 cm (irradiation intensity of 5 mW/cm2). All the samples were tested under the same condition. The absorbance of the MB solution was measured using a UV-vis/NIR spectrophotometer (UV-3600, Shimadzu, Kyoto, Japan) at a wavelength of 664 nm to determine variations in MB concentration with the UV irradiation time.
Results and discussion
It has been previously reported that for NTs grown on a Ti substrate, direct annealing to crystallinity results in severe substrate effects [24,28], resulting in the destruction of tubes and a largely decreased surface area. Herein, by ultrasonication in ethanol, the as-anodized NTs could be detached from the Ti substrate prior to usage (Figure 1a), and thus, the influence of the Ti substrate was successfully eliminated. After the removal of surface NTs, the Ti foil could be reused until the foil was completely consumed. The reaction yield each time (approximately 2 h) per cm2 area of the Ti foil was approximately 0.01 g. The as-prepared NT powder was amorphous and showed a gray color (Figure 1b). After the crystallization at 450°C, the color of the NT powder changed to white (Figure 1c). Compared to NT arrays, dispersed NT powders showed high quality, and thus, the dimension and phase of NTs could be stabilized during high-temperature crystallization, as will be discussed in the next section.
The SEM images of the as-formed NT powders showed a tubular morphology. The tube arrays were separated into isolated tubes, and the entire tube was broken into several segments. Figure 2a shows that the tube wall consisted of TiO2 nanocrystallites. After annealing at 450°C, loosely packed tube agglomerates were observed (Figure 2b). By further increasing the crystallization temperature of the NT powders, the tubular structure was well preserved, with slight structure changes. The SEM image of the samples annealed at 750°C is shown in Figure 2c. However, when NTs were dispersed in DI water, changes in structural integrity were more apparent. Heat treatment at 750°C resulted in the collapse of the NT structure (Additional file 1: Figure S1). The tubes were severely sintered and consolidated, and the tubular architecture gradually disappeared. The changes might probably be due to the water-induced tube wall modification as has been previously reported, leading to a hybrid structure . In addition, to recycle NT powder as a photocatalyst, it was immobilized onto an FTO substrate. From the cross-sectional view in Additional file 1: Figure S2, a uniform, dense, and randomly packed TiO2 film was tightly connected to the substrate, promising for use in photocatalytic reactions.
The effect of annealing temperature on crystal structure is then examined. For NTs on the Ti substrate, rutile developed from the oxidization of underlying Ti metals and was gradually transmitted to the upper tube region. Therefore, in the absence of the Ti metal substrate, no rutile phase was observed up to 750°C, and the anatase phase was maintained (Figure 3). This indicates the NT structure and crystal phase of powders were highly stable. The anatase (101) diffraction peak was more prominent with increasing annealing temperatures, indicating an enhanced crystallinity of the anatase phase. However, the unstable hybrid structure that was dispersed in water could be destroyed during high-temperature annealing, showing a phase transition from anatase to rutile at 650°C, as indicated by the XRD patterns shown in Additional file 1: Figure S3. Furthermore, at 750°C, most of the anatase have been converted to the rutile phase.
The amount of pollutant molecules adsorbed onto the TiO2 surface is an important factor that influences photocatalytic activity. For a larger surface area, the amount of adsorption can be high, resulting in improved photocatalytic activity. For example, when the NT film thickness increased, the rate constant of photocatalytic reaction was significantly improved [5,7,30]. Thus, the surface area and pore size distribution of NT powders were investigated through nitrogen sorption analysis. Figure 4a shows the nitrogen adsorption-desorption isotherms of the NT powders annealed at various temperatures. The BET surface area shows a decreasing tendency with increasing annealing temperatures. The specific surface areas of the NTs annealed at 450°C, 550°C, 650°C, and 750°C were 26.7, 24.9, 21.0, and 8.0 m2/g, respectively. The NT powders possessed a very limited surface decrease up to 650°C, indicating that crystallite growth and aggregation of the NT powder were restricted to the tube walls. However, at 750°C, although the anatase phase was well preserved as previously described, due to the enlargement of TiO2 nanocrystallites, the surface area presented a sharp decrease, which in turn reduced the photocatalytic degradation rate. Figure 4b shows a bimodal BJH average pore size distribution with two peaks, which was analyzed based on adsorption branch. The average pore size increased with higher temperatures, from 16.22 to 38.98 nm. The larger pores were formed during the sintering of the nanocrystallites in the NTs at higher temperatures. Therefore, a few pores were present inside the film, and a reduction in internal surface area was observed. The peaks located at the value of approximately 2 nm corresponded to the mesopores in the tube walls.
To investigate the photodegradation of MB solution by the annealed NT powder electrodes, variations in MB concentration as a function of illumination time were monitored (Figure 5a). In the absence of the NT photocatalyst, the MB solution showed negligible self-degradation. MB molecules decomposed faster with increasing annealing temperatures within the specific range of 450°C to 650°C, indicating that NTs of higher crystallization possessed better photocatalytic activity. In addition, the 750°C sample showed a lower photocatalytic activity. The linear plot in Figure 5a indicates that the photocatalysis process followed a quasi-first-order kinetics and thereby can be presented as ln(C 0/C) = ln(A 0/A) = kt. Here, C 0 is the initial MB concentration, and C is the concentration after irradiation for time t. A 0 and A are the corresponding absorbances. k is a first-order rate kinetic constant, which well represents the photocatalytic activity. A comparison of the rate constant values at different temperatures is presented in Figure 5b.
For the 550°C and 650°C samples, after photocatalytic reaction, the aqueous solutions were almost completely decolorized, showing complete degradation of MB. Although the 650°C sample presented a relatively small surface area, its photocatalytic activity was the highest, with the rate constant of 0.414/h. During the photocatalytic oxidation, the photogenerated electrons and holes either recombined inside TiO2 nanocrystallites or reacted with adsorbed species on the TiO2 surface. The crystallinity of NT powders gradually improved with increasing annealing temperatures, resulting in (a) decreased density of residual elements and thus diminished crystalline defects in anodic tubes , which in turn led to lower electron-hole recombination probability in bulk, and (b) effective diffusion of electrons and holes to adsorbed reactants on the TiO2 surface to participate in the decomposition reaction. Both factors lead to a more efficient electron-hole separation and thus enhanced photocatalytic activity. On the other hand, for the 750°C sample, a significantly smaller surface area and, in turn, a lower photocatalytic activity were observed.
We present a simple and cost-effective approach for the growth, dispersion, annealing, and immobilization of TiO2 NT powders. The structure and crystal phase of the NT powders at low and high temperatures were investigated. The NTs in powder form showed improved structure and crystal phase stability under high-temperature treatment compared to the conventional NTs on a Ti substrate. By increasing the temperature from 450°C to 650°C, the highly crystallized NT photocatalyst showed a significantly enhanced photocatalytic activity, mainly due to the low density of defects. Furthermore, an increase in temperature from 650°C to 750°C led to worse photocatalytic activity due to the aggregation of NTs and thus a largely diminished surface. Further utilization of high-crystallized NTs in photoelectrocatalytic degradation is currently in progress.
Competing interests The authors declare that they have no competing interests.
Authors’ contributions XC and YL proposed the idea and presided over the study. JL, XL, and SZ conceived and designed the experiment. JL wrote the paper. All authors read and approved the final manuscript.
The work was supported by the National Natural Science Foundation of China (Grant Nos. 61125503, 61404081, and 11374204), the Shanghai Municipal Natural Science Foundation (Grant No. 14ZR1417700), the “Shu Guang” project of the Shanghai Municipal Education Commission and Shanghai Education Development Foundation (No. 13SG52), and the Science and Technology Commission of Shanghai Municipality (Nos. 12JC1404400 and 14520501000).
Fujishima, A; Rao, TN; Tryk, DA: Titanium dioxide photocatalysis, J Photoch Photobio C, 2000, 1, S. 1–21Crossref
Gao, M; Peh, CKN; Ong, WL; Ho, GW: Green chemistry synthesis of a nanocomposite graphene hydrogel with three-dimensional nano-mesopores for photocatalytic H2 production, RSC Adv, 2013, 3, S. 13169–77Crossref
Wong, TJ; Lim, FJ; Gao, M; Lee, GH; Ho, GW: Photocatalytic H2 production of composite one-dimensional TiO2 nanostructures of different morphological structures and crystal phases with graphene, Catal Sci Technol, 2013, 3, S. 1086–93Crossref
Zhang, G; Huang, H; Zhang, Y; Chan, H; Zhou, L: Highly ordered nanoporous TiO2 and its photocatalytic properties, Electrochem Commun, 2007, 9, S. 2854–8Crossref
Liu, Z; Zhang, X; Nishimoto, S; Murakami, T; Fujishima, A: Efficient photocatalytic degradation of gaseous acetaldehyde by highly ordered TiO2 nanotube arrays, Environ Sci Technol, 2008, 42, S. 8547–51CrossrefPubMed
Xue, C; Narushima, T; Ishida, Y; Tokunaga, T; Yonezawa, T: Double-wall TiO2 nanotube arrays: enhanced photocatalytic activity and in situ TEM observations at high temperature, ACS Appl Mater Interfaces, 2014, 6, S. 19924–32CrossrefPubMed
Nishanthi, S; Subramanian, E; Sundarakannan, B; Padiyan, DP: An insight into the influence of morphology on the photoelectrochemical activity of TiO2 nanotube arrays, Sol Energy Mater Sol Cells, 2015, 132, S. 204–9Crossref
Paramasivam, I; Macak, J; Schmuki, P: Photocatalytic activity of TiO2 nanotube layers loaded with Ag and Au nanoparticles, Electrochem Commun, 2008, 10, S. 71–5Crossref
Xiao, F: Self-assembly preparation of gold nanoparticles-TiO2 nanotube arrays binary hybrid nanocomposites for photocatalytic applications, J Mater Chem, 2012, 22, S. 7819–30Crossref
Lv, X; Zhang, H; Chang, H: Improved photocatalytic activity of highly ordered TiO2 nanowire arrays for methylene blue degradation, Mater Chem Phys, 2012, 136, S. 789–95Crossref
Xu, Z; Yang, W; Li, Q; Gao, S; Shang, JK: Passivated n–p co-doping of niobium and nitrogen into self-organized TiO2 nanotube arrays for enhanced visible light photocatalytic performance, Appl Catal B, 2014, 144, S. 343–52Crossref
Sun, Y; Yan, KP; Wang, GX; Guo, W; Ma, TL: Effect of annealing temperature on the hydrogen production of TiO2 nanotube arrays in a two-compartment photoelectrochemical cell, J Phys Chem C, 2011, 115, S. 12844–9Crossref
Kang, X; Chen, S: Photocatalytic reduction of methylene blue by TiO2 nanotube arrays: effects of TiO2 crystalline phase, J Mater Sci, 2010, 45, S. 2696–702Crossref
Oh, H-J; Lee, J-H; Kim, Y-J; Suh, S-J; Lee, J-H; Chi, C-S: Synthesis of effective titania nanotubes for wastewater purification, Appl Catal B, 2008, 84, S. 142–7Crossref
Yu, JG; Wang, B: Effect of calcination temperature on morphology and photoelectrochemical properties of anodized titanium dioxide nanotube arrays, Appl Catal B, 2010, 94, S. 295–302Crossref
Lai, Y; Zhuang, H; Sun, L; Chen, Z; Lin, C: Self-organized TiO2 nanotubes in mixed organic–inorganic electrolytes and their photoelectrochemical performance, Electrochim Acta, 2009, 54, S. 6536–42Crossref
Acevedo-Peña, P; Carrera-Crespo, JE; González, F; González, I: Effect of heat treatment on the crystal phase composition, semiconducting properties and photoelectrocatalytic color removal efficiency of TiO2 nanotubes arrays, Electrochim Acta, 2014, 140, S. 564–71Crossref
Huo, K; Wang, H; Zhang, X; Cao, Y; Chu, PK: Heterostructured TiO2 nanoparticles/nanotube arrays: in situ formation from amorphous TiO2 nanotube arrays in water and enhanced photocatalytic activity, Chem Plus Chem, 2012, 77, S. 323–9
Bauer, S; Pittrof, A; Tsuchiya, H; Schmuki, P: Size-effects in TiO2 nanotubes: diameter dependent anatase/rutile stabilization, Electrochem Commun, 2011, 13, S. 538–41Crossref
Pan, Z; Li, Y; Hou, X; Yan, J; Wang, C: Effect of annealing temperature on the photocatalytic activity of CdS-modified TNAs/glass nanotube arrays, Phys E, 2014, 63, S. 1–7Crossref
Lin, J; Guo, M; Yip, CT; Lu, W; Zhang, G; Liu, X: High temperature crystallization of free-standing anatase TiO2 nanotube membranes for high efficiency dye-sensitized solar cells, Adv Funct Mater, 2013, 23, S. 5952–60Crossref
Fahim, NF; Sekino, T: A novel method for synthesis of titania nanotube powders using rapid breakdown anodization, Chem Mater, 2009, 21, S. 1967–79Crossref
Antony, RP; Mathews, T; Dasgupta, A; Dash, S; Tyagi, A; Raj, B: Rapid breakdown anodization technique for the synthesis of high aspect ratio and high surface area anatase TiO2 nanotube powders, J Solid State Chem, 2011, 184, S. 624–32Crossref
Albu, SP; Ghicov, A; Aldabergenova, S; Drechsel, P; LeClere, D; Thompson, GE: Formation of double-walled TiO2 nanotubes and robust anatase membranes, Adv Mater, 2008, 20, S. 4135–9
Lin, J; Liu, X; Guo, M; Lu, W; Zhang, G; Zhou, L: A facile route to fabricate an anodic TiO2 nanotube-nanoparticle hybrid structure for high efficiency dye-sensitized solar cells, Nanoscale, 2012, 4, S. 5148–53CrossrefPubMed
Zhuang, H-F; Lin, C-J; Lai, Y-K; Sun, L; Li, J: Some critical structure factors of titanium oxide nanotube array in its photocatalytic activity, Environ Sci Technol, 2007, 41, S. 4735–40CrossrefPubMed
Mirabolghasemi, H; Liu, N; Lee, K; Schmuki, P: Formation of ‘single walled’ TiO2 nanotubes with significantly enhanced electronic properties for higher efficiency dye-sensitized solar cells, Chem Commun, 2013, 49, S. 2067–9Crossref