Preparation, Characterization, and Photocatalytic Properties of Self-Standing Pure and Cu-Doped TiO2 Nanobelt Membranes.

A traditional hydrothermal method was modified to synthesize ultra-long sodium titanate nanobelts by simultaneously stirring the solution. The ultra-long sodium titanate nanobelts were converted to hydrogen titanate nanobelts through an ion exchanging way. A method was then used to prepare self-standing flexible large-area membranes; they were then subject to post-annealing at different temperatures to obtain a self-standing TiO2 nanobelt membrane with a slight decrease in flexibility. Cu-doped TiO2 membranes were prepared by ion exchanging and post-annealing in the same manner. X-ray diffractions, scanning electron microscopy, field-emission scanning electron microscopy, field-emission transmission electron microscopy, Fourier transform infrared spectrometry, X-ray photoelectron spectroscopy, and UV-vis spectroscopy were used to characterize the samples. Photodegradation of methylene blue (MB) water solutions was used to evaluate the photocatalytic activity. It was seen that the pure sample presented obvious visible-light responding photocatalytic activity, possibly due to the self-sensitization of the MB molecule. The UV-induced photocatalytic activity is higher because of the photoinduced holes and electrons. It was suggested that the Cu dopant induced intra-gap states from electron traps and recombination centers, resulting in the decrease in both of the visible and UV induced photocatalysis.

Introduction

TiO2 still attracts much attention as an important prototype photocatalytic material in these years.[1−7] Photocatalytic effect is its intrinsic property that can be used for pollutant removals and energy generations. Photocatalytic properties of TiO2 are dependent on the electronic structure, crystalline structure, specific surface area, morphologies, pore structure, and exposed surface facets.[8−11] Engineering TiO2 materials by altering their morphologies and electronic structures is the main methodologies to increase the photocatalytic activities. Many works were directed to increase the photocatalytic properties through doping, surface modification,[12−15] and nanostructure construction.[16−22] For example, the TiO2 containing oxygen vacancies or Ti3+ has been widely used in enhancing photocatalytic activity and using visible light.[23] Construction of the TiO2 nanocrystals and nanostructures with highly exposed {001} facets, together integrated with doping and mixing, has also been investigated as an effective ways to increase the photocatalytic activities.[7,24] Grafting the nano-TiO2 materials with graphene, Au, and other co-catalysts, has also drawn people’s interest in photocatalytic H2 generation and environmental remediation due to the increase in charge transfer.[25,26] For the sample, we combined electrospray and hydrothermal post-treatment to prepare hierarchical TiO2 materials with high porous volume, which showed high photocatalytic activities.[27−29]

One-dimensional (1-D) nano-TiO2 materials, including nanotubes, nanorods, and nanofibers, were found to have high photocatalytic activities due to a fast carrier mobility.[30−32] TiO2 nanotube arrays have been widely used in photocatalysis and solar energy conversion, which were mainly prepared from anodization and hydrothermal methods.[33−35] The 1-D TiO2 nano-materials can be also prepared through a hydrothermal treatment of layered titanates. As an example, Yang et al. obtained TiO2(B) nanofibers with a shell of anatase nanocrystals through hydrothermally treating titanates; these materials had high photocatalytic activities because of the special TiO2(B)/anatase structure.[36] In concentrated NaOH solution, sodium titanate nanotubes, sodium titanate nanobelts, and sodium titanate nanoribbons have been obtained via hydrothermal reactions.[37,38] Sodium titanates have a layered structure; the Na ions locate between [TiO6]8 interconnected layers. Sodium titanates can be converted to hydrogen titanate through the exchanging of the Na ions with protons; this can be converted to 1-D TiO2 nano-materials by annealing. Both the titanate and the 1-D TiO2 materials are active in removing pollutants under illumination of UV and visible lights.[39,40] In the present research, a modified hydrothermal reaction was first used to fabricate ultra-long hydrogen titanate nanobelts, which were then used to prepare large-area flexible titanate membranes. The titanate membranes were then changed to TiO2 membranes by annealing at different temperatures, and the photocatalytic activities of them were studied. The Cu ions were also doped in the TiO2 membranes; it was seen that the photocatalytic activities decreased possibly because the Cu dopants might form recombination centers and electron traps. As compared to powder photocatalysts, the self-standing membranes do not need additional supports and can be easily separated from the reaction solution. Therefore, they cannot lead to the secondary particle pollution and can be easily recycled and reactivated.

Experimental Section

Synthesis of TiO2 Nanobelt Membranes

TiO2 amorphous materials were first prepared by an electrospray method according to previous studies.[27−29] Then, 0.3 g electro-sprayed (ES) powder was hydrothermally treated in 50 mL 10 M NaOH aqueous solution in a Teflon-lined stainless-steel autoclave at 180 and 200 °C for 24 h. The hydrothermal solutions were continuously stirred during the hydrothermal reactions. Afterward, the products were cleaned by deionized water several times until the pH was 7, which were then immersed in 0.01 M HCl aqueous solution for 24 h at room temperature to exchange Na+ ions with H+ ions. After then, the products were washed repeatedly with deionized water until the pH was neutral. The as-cleaned materials were then dispersed in ethanol and were then spread on a filter paper (ϕ 150 mm) placed upon the bottom of a glass container. The ethanol was slowly removed by a glass pipette, and the material was dried at ∼100 °C for about 1 h to obtain the final product. A piece of membrane can then be peeled off from the filter paper. The as-prepared membranes were annealed at 400, 500, 600, 700, 800, and 900 °C for 2 h, respectively, to obtain TiO2 membranes. Instead of the ES materials, the commercial P25 TiO2 powder was also used as the raw material to prepare membranes according to the same procedure.

For the preparation of the Cu-doped TiO2 paper, the H+-exchanged hydrothermal products were immersed in 500 mL of 0.01, 0.05, and 0.1 mol L–1 CuCl2 water solution at room temperature for 24 h to exchange partial H+ ions with Cu2+ ions. The Cu2+-exchanged products were then washed repeatedly with deionized water to remove physically adsorbed Cu2+ ions. The Cu-doped membranes were obtained according to the same procedure. Lastly, the Cu-doped titanate membranes were annealed at 500, 600, 700, 800, and 900 °C, respectively.

Characterization

Powder X-ray diffractions (XRD) were used to determine the crystalline phases of the membranes using an RINT 1500 diffractometer (made in Japan) with the Cu Kα radiation being used as the X-ray source. The morphologies were observed using a scanning electron microscope (SEM, SEM-5400, JEOL, Japan) with an accelerated voltage of 20 kV and a field emission scanning electron microscope (FE-SEM, S-4800, Hitachi, Japan) with an accelerated voltage of 10 kV. High-resolution field emission transmission electron microscopy (FE-TEM, EM002BF, Japan) was used to observe the sample morphologies with an accelerated voltage of 160 kV. Ultraviolet–visible (UV–vis) absorption spectra were recorded using a UV–vis spectrophotometer (UV-2450, Shimadzu, Japan) in a wavelength range from 300 to 800 nm. The XPS survey spectrum and core-level spectra of Ti2p, Cu2p, Na1s, and O1s were recorded using X-ray photoelectron spectroscopy (VG Multilab2000X) with an X-ray source working with the Al Kα radiation and using the binding energy (284.6 eV) of C1s electrons as the energy reference. The measurement step size is 0.05 eV. Fourier transformation infrared (FT-IR) absorption spectra of the membranes annealed at different temperatures were obtained on an FT-IR spectrometer (FT/IR-6100, JASCO, Japan) from 450 to 4000 cm–1.

Photocatalytic Activity Measurement

A small piece of the TiO2 membrane (∼ 0.04 g) was used to degrade the methylene blue (MB) aqueous solution (1.0 × 10–5 mol L–1) under visible light and UV light illuminations, respectively. The TiO2 membrane was placed in a ϕ 50 mm glass container, with the 20 mL MB water solution being added. For the photocatalytic reactions under visible light illumination, two 15 W daylight lamps, equipped with a 420 nm short-wavelength cutting filter to remove UV light, were used. Four 15 W UV 365 nm fluorescent lights were used to conduct the UV-induced photocatalysis. The light intensity of UV light was ∼3 mW/cm2. The concentrations of reacted MB water solutions were recorded with a UV–visible spectrometer (UV-1601, Japan) after a certain time interval. The photocatalytic activities were evaluated from the decrease of the MB absorbance at 660 nm during the photocatalytic reactions.

Results and Discussion

Figure A-1 shows the digital picture of the hydrothermally reacted product. The product presents a flocky form that almost filled the half of a beaker (200 mL). Figure A-2 shows the digital picture of the as-prepared pre-annealed membrane with a diameter of ϕ 150 mm and a weight of ca. 0.2 g. This membrane was flexible and could be bent repeatedly several times without damage, so it can be fabricated as various shapes. The SEM image (Figure A-3) shows that the membrane is composed of ultra-long (>100 μm) uniform 1-D nanofibers, which entangled together to form a large-area membrane. Figure B-1 shows the digital picture of the hydrothermal product prepared from the commercial P25 TiO2; the product becomes precipitates on the beaker bottom, which is greatly different from Figure A-1. We cannot obtain a complete membrane, but some small fragments (Figure B-2), although the product also consists of 1-D nanofibers (Figure B-3).

Figure 1

(A-1) Digital picture of the hydrothermal product prepared from the ES materials, dispersed in water. (A-2) Digital picture of the free-standing membrane (ϕ 150 mm) prepared from the H+-exchanged product. (A-3) SEM image of the membrane composed of entangled ultra-long uniform nanofibers. (B-1) Digital picture of the water dispersion of the hydrothermal product prepared form the commercial p25 powder. (B-2) Digital picture of the cracked membrane. (B-3) SEM image of the membrane composed of nanofibers.

Figure A shows the cross-sectional SEM image of the membrane prepared from the ES materials. It is seen that the thickness of the membrane is ca. 20 μm. Figure B shows the SEM image of the cracked segments prepared from the commercial P25. Although the nanofibers also entangled together, they broke into small pieces after drying. This indicates that the flexibility of the nanofibers should be lower than that prepared from the ES materials, so a complete membrane cannot be obtained. The FE-SEM image (Figure C) shows that the 1-D nanofibers in the membrane prepared from ES materials are in the form of nanobelts. The single nanobelt has a width of ∼50 nm. Some of them compact together layer by layer to form polymerized nanobelts, finally presenting in nanofiber morphology. The TEM image (Figure D) also shows the 1-D nanobelt morphology. The nanobelts cannot be obtained if the hydrothermal solutions were not stirred during reactions. Because the ES materials used in the hydrothermal reactions are amorphous TiO2-polymer (PVP) aggregates, stirring was necessary to break the TiO2-polymer aggregates to react with the highly concentrated NaOH water solutions in order to obtain the nanobelt morphology.

Figure 2

(A) Cross-sectional SEM image of the membrane prepared from the ES materials. (B) Low-magnified SEM image of the membrane prepared from the commercial P25 powders. (C) High-resolution FE-SEM image of the membrane prepared from the ES materials. (D) TEM image of the 1-D nanobelts.

Figure A shows the XRD patterns of the as-prepared hydrothermal products prepared from the ES materials at 180 and 200 °C, respectively; the XRD pattern of the sample prepared from the commercial P25 powder at 180 °C is also shown. It is seen that the crystalline phase is Na2Ti3O7. Clearly, the products prepared from the ES materials have better crystallinity than that prepared from the P25 powder, which prefer to grow along the [200] (2θ:11°) direction. Because of the better crystallinity of the Na2Ti3O7 nanobelts, they can have stronger mechanical intensity and are longer to form nanobelt membranes. Figure B shows the XRD patterns of the membranes annealed at the temperatures from 400 to 900 °C. It is seen that the crystalline phase is converted to anatase TiO2 when the temperature is below 800 °C. If the temperature is higher than 900 °C, then the titanate was changed to rutile TiO2. The higher transition temperature of anatase-to-rutile (A-R) transformation is due to the suppression of A-R transition in the course of calcination by the residual Na+ ions in the TiO2 nanobelt membrane as revealed by the XPS analysis below. The residual Na+ ions should limit the interface diffusion that is necessary for phase transformation. Figure C shows the FT-IR spectra of the H+-exchanged nanobelt membrane, and the membranes annealed at 400, 500, 600, and 700 °C. The IR absorption bands labeled as A and B are ascribed to the physically absorbed water and chemically bonded hydroxyl groups (OH–) in hydrogen titanates, respectively. After calcination, the chemically bonded OH– disappears due to phase transition to anatase TiO2. The IR absorption peak at 469 cm–1 is from the Ti–O–Ti vibrations of the anatase lattice. For the titanate nanobelts, there are three IR peaks at 673, 780, and 958 cm–1, arising from the Ti–O–Ti vibrations in the Ti–O layer. After calcination, the peak at 673 cm–1 disappeared due to the phase transition.

Figure 3

(A) XRD patterns of the hydrothermal products prepared from the ES and the commercial P25 materials. (B) XRD patterns of the membranes after post-annealing. (C) FTIR spectra of the H+-exchanged membrane and the membranes after post-annealing at different temperatures.

Figure A shows the FE-SEM image of the nanobelt membrane after 2 h of annealing at 500 °C. It is seen that the annealing process did not destroy the 1-D nanobelt morphology. Figure B shows the corresponding TEM image, which also shows the form of nanobelts and several of nanobelts fused together. Therefore, the post-annealing does not change the morphology of the nanobelts although it is transformed to anatase TiO2, so the sample still presents the shape of the membrane. The flexibility decreases because of the increase in fragility.

Figure 4

(A) FE-SEM image of the hydrogen titanate membrane subjected to 2 h post-annealing at 500 °C. (B) Corresponding TEM image.

Figure A shows the UV–vis absorption spectra of the hydrogen titanate membrane, pure TiO2 nanobelt membrane annealed at 600 °C, and Cu-doped TiO2 membranes annealed at 600 °C prepared by the ion exchanging with different concentrations of CuCl2 solutions. The titanate membrane (labeled as “not annealed”, black line) has no visible light absorption; the absorption edge is located near 360 nm. The absorption edge of the pure TiO2 membrane shifts to 420 nm; this agrees well with the XRD analysis that the titanate phase was transformed to anatase TiO2. The introduction of Cu ions in TiO2 generates obvious absorption in the visible region from 400 to 600 nm; this increases with the increase in CuCl2 solution concentrations. It is seen that the Cu doping almost does not affect the absorption edge of the TiO2 membrane; this means that the Cu dopants form the gap states and do not affect the conduction band (CB) and valence band (VB) of TiO2. The effect of annealed temperature on the UV–vis absorptions was also studied, as shown in Figure B. The annealing below 800 °C did not change the absorption edge of Cu-doped TiO2 samples, indicating that the crystalline phase of anatase is unchanged. When the annealing temperature increased to 900 °C, the absorption edge shows an obvious redshift due to the anatase-to-rutile transition, in good accordance with the XRD analysis. It is also seen that the Cu doping leads to a wide absorption from 400 to 800 nm in the rutile phase, different from the anatase case.

Figure 5

(A) UV–Vis diffusion absorptions of the hydrogen titanate membrane, pure anatase TiO2 membrane, and Cu-doped anatase TiO2 membranes. (B) UV–Vis diffusion absorptions of Cu-doped TiO2 membranes prepared through the calcination of titanate membranes at different temperatures.

XPS was employed to explore the chemical states of surface elements of the Cu-doped TiO2 membrane prepared from the 0.1 mol L–1 CuCl2 solution and the post-annealing at 600 °C. Figure shows the XPS survey spectra of the Cu-doped sample, which shows that it contains C, Ti, O, and Cu elements. The C element comes from the oil contaminants in the XPS measurement. The peak near 1070 eV might be ascribed to Na1s or Ti Auger peaks. As no Na Auger peak around 470 eV was detected, this peak is attributed to the Ti (A) peak. This result indicates that almost all of Na ions in the samples can be removed by the ion exchanging, in accordance with the XRD analysis. The peak around 970 eV is ascribed to the O Auger peak. Figure shows the core-level XPS spectra of the Cu2p, O1s, and Ti2p, respectively. The Cu2p3/2 was fitted with two Lorentz–Gaussian peaks, showing the presence of Cu0/Cu+ and Cu2+ species.[13,41−43] The Cu2+ species are also indicated by the satellite peak in the spectrum. The atomic ratio of Cu0/Cu+ and Cu2+ is about 3:1, so Cu dopants mainly exist in the Cu0/Cu+ state. Figure B shows that the O1s spectrum can be also fitted with two peaks, showing the presence of hydroxyl groups in addition to the lattice oxygens. Figure C shows the Ti2p XPS peaks at 463.9 and 458.2 eV that can be well fitted with two Lorentz–Gaussian peaks. The 5.7 eV splitting of spin-orbital coupling between them indicate the +4 valence of Ti ions. The presence of residual Na+ ions might increase the A-R transition temperatures. Combining the UV–vis analysis and XPS analysis, we suggested that the Cu ions should enter the TiO2 lattice. Corresponding oxygen impurities should also be generated in the sample in order to keep electric neutrality. The Cu species and oxygen impurities may lead to the exponential tail of the Cu-induced absorption, as shown in Figure A.

Figure 6

XPS survey spectrum of the Cu-doped TiO2 membrane.

Figure 7

(A) Cu2p core-level XPS spectrum. (B) O1s core-level XPS spectrum. (C) Ti2p core-level XPS spectrum

Figure shows the change of normalized MB concentrations in the dark, under visible light illumination, and under UV light illumination. The pure TiO2 and the Cu-doped membranes, which were obtained by annealing at 600 °C, were used as the photocatalysts. After the dark absorption reached equilibrium, the visible light was first turned on. It is seen that the pure TiO2 membrane shows an obvious photocatalytic activity, which tends to disappear after Cu doping. This result seems to contradict with the common knowledge that the visible light activity should increase because the Cu-doped TiO2 absorbs visible light. Afterward, the visible light was turned off, and the UV light was turned on. It is seen that the UV-induced photocatalysis becomes faster than the visible-induced photocatalysis. The Cu doping also greatly decreases the UV-induced photocatalytic activity.

Figure 8

Dependences of normalized MB concentrations (C/C0) on time in the dark and during the photocatalytic reactions over the pure TiO2 membrane and Cu-doped TiO2 membranes prepared from 0.01 and 0.05 mol L–1 CuCl2 water solutions

Figure A shows the normalized concentrations of the MB solutions in the dark and under visible light illumination for the TiO2 membranes obtained by annealing at different temperatures. It is seen that the dark absorption amount of the MB on the TiO2 membranes increases with the increase in annealing temperatures. Nearly 70% of MB can be adsorbed by the sample prepared by annealing at 900 °C. Under visible light illumination, all of the samples show the photocatalytic activities. The photocatalytic activities of the anatase samples (annealed below 800 °C) are slightly better than that of the rutile sample (annealed at 900 °C); this is consistent with the common knowledge that the anatase has higher activity than the rutile TiO2.[44] Because the TiO2 membranes do not absorb the light below 420 nm, it is suggested that the visible-induced photocatalysis may be caused by the self-sensitization of the MB molecule adsorbed on the TiO2 surface. Figure B shows the dark adsorption and the photocatalytic degradation of MB solutions under visible light illumination by the Cu-doped TiO2 membrane prepared by the post-annealing at 900 °C. The crystalline phase of TiO2 is rutile. Independent on the doped Cu amount, it is seen the Cu-doped membranes also have strong dark absorption. They do not show visible activity, the same as the anatase cases.

Figure 9

(A) Time dependence of the normalized concentrations (C/C0) of the MB solutions in the dark and under visible light illumination for the pure TiO2 membranes prepared by post-annealing at different temperatures. (B) Time dependence of the C/C0 of the MB solutions in the dark and under visible light illumination for the Cu-doped TiO2 membranes prepared from the 0.01 and 0.1 mol L–1 CuCl2 solutions.

Figure shows the proposed mechanism of self-sensitized photocatalysis and the effect of Cu dopants. Under visible light illumination, the electrons can be excited from the highest occupied molecular orbitals (HOMOs) of the MB molecules to their lowest unoccupied molecular orbitals (LUMOs). The electrons in the LUMOs can transfer to the CB of the TiO2, and they can then transport along the 1-D TiO2 materials. The pure TiO2 membranes are well crystalline and contain less electrons traps, so the electrons from the MB molecule can transfer to the O2 molecule and cause the formation of super oxygen free radicals (O2–). The O2– is an important reactive oxidative species (ROS) that can induce the formation of H2O2 and ·OH (hydroxyl free groups). These ROSs lead to the MB photodegradation under visible light illumination. The UV–vis absorptions indicate that the Cu dopants should mainly form deep gap states (defect levels) in the forbidden gap of the TiO2. The deep gap states can trap the electrons from the CB of TiO2 materials, which then become immobile and need a long time to be detrapped to the CB, so the transfer of the electrons to O2 is limited and the photocatalysis is thus stopped. The mechanism of the UV light-induced photocatalytic effect is different, which arises from the generation of photoinduced holes and electrons in the CB and the VB of the TiO2 samples. The photocatalytic results also indicate that the Cu dopants mainly form the recombination centers for the photoinduced holes and electrons, which results in the decrease in photocatalytic activities.

Figure 10

Proposed mechanism of self-sensitized photocatalysis and the photocatalytic effect prohibited by the Cu dopants.

Conclusions

A facile method was used to fabricate ultra-long nanobelt sodium titanates, which were then converted to hydrogen titanates by H+ exchanging. The ultra-long hydrogen titanates were used to construct self-standing large-area membranes. The TiO2 nanobelt membranes were obtained by the post-annealing of the hydrogen titanate membranes. Cu-doped TiO2 membranes were also prepared by exchanging the H+ in hydrogen titanate with Cu2+ ions before membrane construction. It was seen that the TiO2 membrane showed a high temperature for A-R phase transition due to the existence of residual Na+ ions. The valence state of Cu dopants includes both Cu+ and Cu2+, which causes obvious visible absorption ranging from 400 to 600 nm for anatase and 400 to 800 nm for rutile. The UV–visible absorptions show that Cu dopants created gap states in the TiO2. In additional to the UV-induced photocatalytic effect, the pure TiO2 membranes also showed the self-sensitized photocatalytic effect for photodegrading the MB under visible light illumination. Independent on the Cu doping, the TiO2 membrane also showed good ability for the MB dark adsorption. It was seen that the Cu dopants reduced the photocatalytic activities for both UV and visible light illuminations, possibly because they formed electron traps and recombination centers.