Photogenerated Oxygen Vacancies in Hierarchical Ag/TiO2 Nanoflowers for Enhanced Photocatalytic Reactions.

Oxygen vacancy (Vo) creation and morphology controlling make significant contributions to the electronic and structural regulation of metal oxide semiconductors, yet an investigation about convenient approaches for fabricating hierarchical catalyst with abundant oxygen vacancies still has significant challenges. Here, we report a unique method to create abundant oxygen vacancies in hierarchical Ag/TiO2 nanoflowers during photocatalytic reaction, which is accompanied by light absorption variation and surface plasmon resonance (SPR) enhancement. Its high efficiency of photocatalytic H2 evolution (the highest apparent quantum yield reaches 3.2% at 365 nm) and rhodamine B degradation can be considered as benefits from the synergistic effects of the well-arranged hierarchical structure, the photogenerated oxygen vacancies, and the SPR of cocatalyst Ag. This work proposes an effective strategy to optimize the synthesis of regular hierarchical structures and enriches the research on the vital function of oxygen vacancies in photocatalytic reactions.

Introduction

Photocatalysis, as a renewable and environmentally friendly strategy for solving major energy and environmental issues, has received considerable attention in recent decades.[1−3] Because photocatalytic reactions are complex processes, mass transport of reactant/product, adsorption/desorption of molecules, separation efficiency and transmission rate of photogenerated electron–hole pairs, and reaction thermodynamics all should be taken into consideration. Many efforts have been devoted to the fabrication of photocatalysts and reaction systems for elevating the conversion through different perspectives,[4−7] but it still remains a challenging problem to develop a preponderant catalyst with regards to the above aspects in photocatalytic reaction.

As a classic semiconductor material, TiO2 has excellent application in photocatalysis.[3] Although the research on TiO2 has been going on for decades, its superior catalytic performance and convenient synthesis method ensure its unshakeable status in the field of photocatalysis,[8] and there is still space and value in further study of efficient TiO2-based materials. Recently, the hierarchical structured TiO2 becomes a research hotspot on account of its multiple benefits.[9−13] Various hierarchical TiO2 materials have been developed for photocatalytic reactions, such as nanosheet-coated hierarchical nanotubes,[9] hierarchical urchin-like double-hollow nanospheres,[10] hierarchical round cake assembled with nanosheets,[11] hierarchical nanoporous sphere[12] and hierarchical macro-/mesoporous catalysts.[13] By reason of the unique hierarchical structure, these materials have been appreciated for the larger specific surface area, more light absorption, and enhanced mass transfer. However, morphology engineering can only change the properties of materials to a certain extent, yet electronic structural alteration may bring more substantial changes for catalysts and photocatalytic reactions.

As one of the significant structural modifications, surface Vo has been reported by some experimental instances that it can effectively improve the photocatalytic performance.[14−16] With the advent of Vo, the band gap of TiO2 always becomes narrow, which means the catalyst has higher visible light response.[17] Even more amazing is that the apparent color of catalyst can be substantially changed by Vo, and tuning the amount of Vo can further fine-tune the powder color.[17,18] In addition, the electronic structural change caused by surface Vo can further make Vo the electron mediator or reaction sites.[19,20] Numerous factors of Vo collectively contribute to the elevated photocatalytic activity. According to these discoveries, construction of the structure with abundant Vo is regarded as a new strategy to enhance photocatalytic activity on TiO2.[21]

A variety of methods have been used for Vo creation, such as high/low pressure H2 treatment,[22,23] Ar–H2 treatment,[24] hydrogen plasma,[25] NaBH4 reduction,[26,27] and electrochemical doping.[28] These paths are all effective in creating a certain amount of surface oxygen vacancy but take a lot of works. Zhou et al.[29] reported a kind of hole trapping Vo formed by noble metal Pt loading. The surface Vo exists under the cover of a metal particle under illumination condition. During the photocatalytic process, the specific Vo traps the photogenerated holes whereas allow photogenerated electrons to pass through and reach the Pt surface. This simple operation not only can load the noble metal cocatalyst onto the TiO2 surface but also can create Vo in the photocatalytic process.

Herein, we fabricated hierarchical Ag/TiO2 nanoflowers with photogenerated oxygen vacancies for the photocatalytic H2 evolution and degradation of rhodamine B (RhB) with good efficiency and stability. The nanoflower morphology of TiO2 can be tuned by acidity of the mixture in hydrothermal reaction. The regular TiO2 nanoflowers (T-5) with higher crystalline and more rutile content show elevated photocatalytic activity because of its superior optical absorption and separation efficiency of photogenerated electron–hole pairs. With the addition of cocatalyst Ag, the obtained H2 evolution and RhB degradation rate were further improved for Ag/T-5. During the reactions, abundant Vo generated, which simultaneously improve the visible light absorption and enhance the surface plasmon resonance (SPR) of Ag. The directionally electron transfer driven by the synergistic effect of Vo and SPR can provide high concentrations of electrons and reduce photogenic carrier recombination, thus improving the photocatalytic performance. This study is beneficial to explore simple synthetic method of Vo containing hierarchical oxide materials for a better photocatalytic application.

Experimental Section

Materials

All chemicals are of analytical grade and used without further purification, unless otherwise stated. Titanium(IV) sulfate [Ti(SO4)2, CP, ≥96.0%] and silver nitrate (AgNO3, AR) are purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Hydrochloric acid (HCl, AR, 36–38%) is from Yantai Sanhe Chemical Reagent Co., Ltd. (Yantai, China). Distilled water is used in all experiments.

Synthesis

The novel three-dimensional flower-like anatase-rutile TiO2 materials were prepared by the following typical procedure. Ti(SO4)2 (CP, ≥96.0%) (0.12 g) was dissolved in 5 mL of deionized water (solution-A). An aqueous solution (solution-B) was prepared by adding 10 mL of de-ionized water into 5 mL of HCl (AR, 36–38%). Solution-A and solution-B were mixed together to yield a transparent solution (solution-C). Solution-C was transferred to a 25 mL Teflon-lined stainless-steel autoclave. The autoclave was sealed well, and the hydrothermal reaction was carried out at a temperature of 180 °C for 120 min. After the reaction, the autoclave was allowed to cool down naturally. The white colored sediments were centrifuged and washed several times with deionized water and alcohol. The pure sample obtained after centrifugation was dried in an oven at 70 °C and was denoted as T-5. Following the same procedure, sample T-2 was prepared by changing the addition of HCl (AR, 36–38%) to 2 mL.

The Ag/T-5 catalyst was synthesized by an incipient wetness impregnation method, and silver nitrate was used as a precursor. Then, the sample was dried at 70 °C overnight and was treated in air flow at 500 °C for 2 h. The nominal amount of silver in the as prepared catalyst was 1 wt %.

Characterization

The powder X-ray diffraction (XRD) patterns were recorded on a Rigaku D/max-2500 X-ray diffractometer using Cu Kα radiation (40 kV and 150 mA) at a scanning rate of 10°/min. The morphologies of the prepared samples were observed by field emission scanning electron microscopy (SEM, JSM-6700F) and high-resolution transmission electron microscopy (TEM, JEOL JEM-2010) at 200 kV. UV–vis diffuse reflectance spectra (DRS) were obtained on a Model c spectrophotometer furnished with an integrating sphere with a reflectance standard of BaSO4. Electron paramagnetic resonance (EPR) signal of oxygen vacancy was recorded on a Bruker ESP 300 EPR spectrometer at 77 K, and the signals of 5,5-dimethyl-1-pyrroline N-oxide (DMPO)–•O2– and DMPO–•OH were recorded in DMPO solution at 273 K. The photoluminescence spectrum (PL) was obtained on a PerkinElmer LS-55 spectrophotometer with an excitation wavelength of 325 nm. X-ray photoelectron spectroscopy (XPS) were performed on a Quantum 2000 Scanning ESCA Microprobe (Physical Electronics) using Al Kα radiation (1846.6 eV) as the X-ray source. The binding energy of the C 1s line at 284.6 eV was used as an internal standard reference. The concentration of Ag+ in the reaction solution was measured using an inductively coupled plasma optical emission spectrometer (ICP-OES) (Ultima 2, HORIBA Jobin Yvon Co., France).

Photocatalytic Degradation of RhB

The photocatalytic RhB degradation activity of the samples was evaluated in aqueous solution. Catalyst powder (0.01 g) was put into 100 mL of 1.0 × 10–5 M RhB aqueous solution, and it is worth noting that the concentration is the same for all investigated catalysts. Prior to being exposed to the Xe lamp irradiation, 30 min magnetic stirring was applied to make the dye molecules adsorbed on the catalyst surface to reach the adsorption–desorption equilibrium. Then, it was irradiated using a Xe lamp and the samples were drawn out at suitable intervals. Centrifugal separation was used to obtain the upper clear solution. The absorption spectrum of the RhB solution was detected by UV–visible spectroscopy (a CARY 500 UV/vis/NIR spectrometer) in the wavelengths range from 200 to 800 nm. The concentration of RhB was determined by monitoring the changes in the absorbance maximum at about 553 nm. According to the Lambert–Beer law, the absorbance is proportional to the concentration of a dilute solution, from which the photodegradation percentage is calculated using the formulawhere η is the degradation rate, C0 the initial concentration of the RhB solution, and C the concentration at a certain irradiation time. Following the photodegradation studies, reaction kinetics of these processes were pursued. In general, the degradation of RhB can be considered as first-order of the dynamical reaction as followswhere k is the reaction rate constant and t the reaction time. It should be emphasized that photocatalytic tests were implemented under the stable conditions.

Photocatalytic H2 Evolution

Generally, 50 mg of the as-prepared catalysts was ultrasonically dispersed into a mixed solution of 90 mL of deionized water and 10 mL of methanol. The pyrex reactor with a quartz upper cover was used in the photocatalytic reaction. A 300 W xenon lamp (light intensity was 100 mW/cm2) was placed on the top of photoreactor as the light source. H2 evolution process lasts 10 h and keeps stirring during this period. Gas chromatography (GC-7920) was used to analyze the amount of hydrogen production, and high-purity nitrogen was used as carrier gas.

Photoelectrochemical Measurements

After ultrasonically cleaned in deionized water, acetone, and ethanol for 10 min in proper order, the indium-tin oxide (ITO) glass pieces were used to prepare working electrodes. The slurry, which was prepared by 5 mg of photocatalyst ultrasonically dispersed into 5 mL of dimethylformamide for 1 h, was drip-coated on the conductive surface of ITO glass to form a uniform photocatalyst film with an area of 1 cm2. All electrochemical and photoelectrochemical (PEC) properties were measured with a three-electrode system by Princeton Applied Research ParStat 2273. Typically, a catalyst-coated ITO glass was used as the working electrode, a platinum sheet was used as the counter electrode, and a saturated Ag/AgCl was used as the reference electrode. The Mott–Schottky plots were measured in a 0.5 M KPi buffer (pH 7) at a frequency of 1000 Hz and amplitude of 10 mV under the dark condition. Electrochemical impedance spectroscopy (EIS) was performed in an alternating current voltage amplitude of 10 mV with a 0.5 Hz to 105 KHz frequency range, and 0.5 M Na2SO4 solution was used as electrolyte. The photocurrent density versus time curves was obtained by intermittent illumination using a 100 W xenon lamp and tracking several cycles at a voltage of 0.0 V versus reference electrode.

Results and Discussion

Crystal Structure

XRD patterns of the obtained TiO2 samples demonstrate that almost all the diffraction peaks can be well indexed to the TiO2 rutile phase (JCPDS no. 65-1119), as shown in Figure . A small peak at 2θ = 25.3° assigned to a TiO2 anatase phase (JCPDS no. 21-1272) exists in both T-2 and T-5. Comparing with T-2, T-5 with more HCl in hydrothermal reaction shows improved crystallinity and less TiO2 anatase phase content, which infers that relatively low pH is beneficial to form rutile TiO2.[30] After the Ag loading process along with heat treatment, the peak of anatase phase cannot be found, and Ag/T-5 keeps the rutile phase. In addition, no obvious diffraction of Ag species is observed, which indicates that Ag species (consist of Ag0 and Ag+, see the XPS spectrum in Figure b) highly disperse on the TiO2 surface.

Figure 1

XRD patterns of T-2, T-5, and Ag/T-5.

Figure 9

XPS spectra of (a) Ti 2p and (b) Ag 3d in Ag/T-5 before and (c) Ti 2p and (d) Ag 3d in Ag/T-5 after the photocatalytic H2 evolution reaction.

Morphology

The SEM images of T-2 and T-5 were shown in Figure a,b. It can be seen that these two samples reveal hierarchical flower-like structures with an average size of 3 μm. Because both flower-like structures were assembled by numerous nanorods, the nanorods of T-5 in Figure b have higher uniformity with an average length of 1–2 versus 0.1–0.2 μm in thickness. The obvious prism of these uniform nanorods may be attributed to the variation of the hydrolysis rate of the Ti(SO4)2 precursors. During the synthesis process of T-5 in a fairly strong acidic aqueous medium, extremely low pH suppresses the hydrolysis rate of the titanium source which benefits the growth of oriented TiO2 nanorods.[31] Cocatalyst Ag was further loaded on the surface of T-5, and the close contact of Ag particle and TiO2 can be observed in the HRTEM of Ag/T-5 (Figure c). In addition, the clear lattice fringes of TiO2 and Ag are also found in Figure c, indicating good crystallinity of Ag/T-5 sample. The lattice fringes with 0.32 and 0.24 nm of spacing are in accordance with the d-spacing of the (110) facet of rutile TiO2 and the (111) facet of Ag, which corroborates the existence of Ag species on the TiO2 surface.

Figure 2

SEM images of (a) T-2 and (b) T-5; (c) HRTEM and (d) TEM images of Ag/T-5; (e) histogram of particle size distribution of Ag particles; (f) EDS-mapping images of Ag/T-5.

In order to investigate the particle size and distribution of Ag cocatalyst, the TEM image with smaller magnification is shown in Figure d. The Ag particles on a Ag/T-5 sample are circled in white, and histogram of particle size distribution of Ag particles is shown in Figure e. The Ag particle size ranges from 8 to 18 nm, and the average size is calculated as 13 ± 2.5 nm. The morphology of TiO2 nanorod corresponds with its nanorod structure in the SEM image of T-5 (Figure b), which means the Ag loading process along with heat treatment did not seriously affect the morphology of TiO2. To make sure that the Ag element is highly dispersed, the EDS-mapping images of Ag/T-5 are provided in Figure f. In the selected area, Ag has the same distribution region as that of Ti and O, which demonstrates that Ag particles are highly dispersed on the surface of TiO2. The lighter color of Ag compared with that of Ti or O can be attributed to its low loading amount (1%) on the Ag/T-5 sample.

In order to clarify the intermediate morphologies that transforms to the final flower-like structures, the time-dependent morphological evolution of T-5 sample is examined by SEM in Figure at different hydrothermal reaction times for 10, 30, and 120 min. As shown in Figure a, only some spherical particles with diameters of 30–50 nm consisting of smaller particles are obtained when the reaction time is 10 min, which can be ascribed to the nucleation at the beginning of the reaction. Prolonging the reaction time to 30 min, in Figure b, these particles turn to some microspheres with an average size of 300 nm, which are covered with several nano-pins on the surface. The emergence of these nano-pins portends the oriented growth of the crystal grain. Moreover, the rough sphere surface with small sharp points indicates that the oriented growth will continue and more nano-pins may appear in subsequent reaction. The morphology of sample obtained after the reaction for 120 min is shown in Figure c, and it can be seen that the three-dimensional nanoflower structure is assembled by the nanorods with a length of 1–2 μm and a thickness of 50–100 nm. As expected, more nano-pin like crystals arise and these nano-pins further grow into nanorods. From the present results, the possible formation mechanism of the flower-like TiO2 (T-5) can be elucidated in Figure d. In the process of the hydrothermal reaction, titanium(IV) sulfate is first hydrolyzed and then nucleates to form spherical nanoparticles. Next, within a fairly strong acidic aqueous medium, small particles gradually dissolve and oriented grow to form the nano-pin like crystals on the surfaces of microspheres.[31] The oriented nano-pins further grow, and the three-dimensional flower-like microstructures are finally formed.

Figure 3

SEM images of T-5 after hydrothermal reaction for (a) 10 min, (b) 30 min, and (c) 120 min; (d) schematic illustration of the flower-like morphological evolution.

Energy Band Structure

The UV–vis absorption spectra of T-2, T-5, and Ag/T-5 are measured in Figure S1. Although these sample show similar optical absorption in UV region, Ag/T-5 exhibits a slight red shift of the absorption band edge and elevated light absorption in visible area. The small optimization of light absorption on Ag/T-5 should be favorable photocatalytic reactions. Besides the optical absorption properties, the energy band structures of TiO2 samples exhibit subtle differences, which have been evaluated by (αhν)2 versus photo energy curves and Mott–Schottky plots in Figure . According to the obtained band gap and conduction band position, the valence band position of T-2, T-5, and Ag/T-5 can be calculated as 2.72, 2.73, and 2.74 V, respectively. [vs standard hydrogen electrode (SHE)]. These speculative valence band positions keep highly consistent with that of TiO2 (−7.25 eV versus the vacuum level, i.e., 2.75 V versus SHE) in the literature.[32] The specific band gaps and conduction/valence band positions derived from the graphs are listed in Table . Compared with T-2, T-5 has a narrower band gap of 3.09 eV, which means higher absorption ability of visible light. Loading with Ag can further narrow the band gap of Ag/T-5 to 3.08 eV. As the band gap narrows, the conduction bands of samples shift to positive direction. The obtained conduction band positions of these samples can completely satisfy the need of the reaction potential of photocatalytic H2 evolution (H+ → H2, 0 V vs SHE)[33] and O2•– radical generating (O2 → O2•–, −0.3 V vs SHE)[34] for organic pollutants degradation. Meanwhile, the valence bands have slightly positive shift as the band gap narrows, but the strong oxidation capacity of catalysts will not be greatly affected.

Figure 4

(a) (αhν)2 versus photo energy curves and (b) Mott–Schottky plots of T-2, T-5, and Ag/T-5.

Table 1

Band gaps and Conduction/Valence Band Positions of T-2, T-5, and Ag/T-5

sample	band gap (eV)	conduction band (V)	valence band (V)
T-2	3.12	–0.40	2.72
T-5	3.09	–0.36	2.73
Ag/T-5	3.08	–0.34	2.74

Photocatalytic Performance

After identifying the band positions of these hierarchical TiO2 samples, photocatalytic H2 evolution was first involved to evaluate the performance of the as-prepared photocatalysts. As shown in Figure a, the H2 evolution performance of Ag/T-5 with different Ag ratios is measured to affirm the appropriate Ag loading amount on T-5. 1.0-Ag/T-5 shows the highest H2 evolution rate, which makes 1.0% the best Ag loading amount. In this case, the Ag/T-5 sample in this work is subject to 1.0% Ag loading capacity. The H2 evolution activity of Ag/T-5 is compared with T-2 and T-5 in Figure b. Three samples exert low photocatalytic H2 evolution performance with little difference in the first 3 h. After the activation process in the early stage, the color of Ag/T-5 turned from light gray to dark (the inset of Figure ), and during the subsequent measurements, the performance of Ag/T-5 enhances and exerts the highest H2 evolution rate of 294 μmol g–1 h–1, which is about 4.5-fold enhancement in H2 evolution activity compared with that of T-5 before Ag modification under same experimental conditions. T-5 with more TiO2 rutile phase content, better crystallinity, and highly uniformity of flower-like morphology exhibits a higher photocatalytic hydrogen evolution activity than T-2. Besides the activities mentioned above, the photocatalytic H2 evolution stability of the best sample Ag/T-5 also should be taken into consideration. In Figure c, after six circle runs for 60 h in total, Ag/T-5 keeps the high H2 evolution rate as previous without significant change, indicating that this sample possesses high stability during the reaction. In addition, the wavelength-depended apparent quantum yield (AQY) of H2 evolution over Ag/T-5 in various wavelength ranges is calculated in Figure d, and the highest AQY reaches 3.2% under 365 nm monochromatic light irradiation.

Figure 5

(a) Photocatalytic H2 evolution over Ag/T-5 with different Ag ratios; (b) photocatalytic H2 evolution over T-2, T-5, and Ag/T-5 during the reaction for 10 h; (c) stability of Ag/T-5 under the photocatalytic H2 evolution reaction for six cycles; (d) AQY of H2 evolution over Ag/T-5 photocatalyst in various light wavelength ranges.

Figure 7

ESR spectra of Ag/T-5 (a) without and (b) with irradiation; O 1s core level XPS spectra of Ag/T-5 (c) before and (d) after the photocatalytic reaction. The insert figures show the color variation of Ag/T-5 sample before and after irradiation. Photograph courtesy of Ying Wang. Copyright 2020.

The photocatalytic performance of these hierarchical TiO2 architectures was also evaluated with photocatalytic degradation of a representative organic pollutant RhB under simulated sun-light irradiation. The UV–vis absorption spectra of RhB aqueous solution at an interval of 20 min with Ag/T-5 in Figure S2 reveals that the main peak at ∼550 nm obviously decreases along with illumination time increasing. The relative concentrations of remained RhB in the solution in accordance with the irradiation time over different photocatalysts are shown in Figure a. After irradiation for 120 min, the photocatalytic degradation rate of Ag/T-5 reaches 92%, while those of T-5 and T-2 are only 81 and 62%, respectively. In addition, the fitting curve of ln(C0/C) versus time (Figure S3) illustrates the best photocatalytic degradation activity of Ag/T-5 as well. To investigate the photocatalytic stability of the as-prepared Ag/T-5, photocatalytic degradation of RhB is repeated for six cycles, as shown in Figure b. Similar to photocatalytic hydrogen evolution, an activation process occurs at the beginning of illumination, which leads to the slower degradation rate in the first circle. In the next five circles, Ag/T-5 exhibits enhanced degradation activity. After these cycles, photocatalytic degradation performance of Ag/T-5 basically remained unchanged, which illustrates that Ag/T-5 has both high photocatalytic stability and superior reusability. The outstanding photocatalytic stability in H2 evolution and RhB degradation reactions may result from the self-assembled flower-like hierarchical structure. The unique morphology is beneficial for restraining the unordered aggregation of the nanorods, consequently retaining the properties of the primary subunits.

Figure 6

(a) Photocatalytic RhB degradation over T-2, T-5, and Ag/T-5 during the reaction for 120 min; (b) stability of Ag/T-5 under the photocatalytic RhB degradation reaction for six cycles.

Usually, cocatalysts play vital roles in decreasing overpotential, facilitating charge transfer, inhibiting electron–hole recombination, providing catalytic active sites, and enhancing light harvesting and reactant adsorption in photocatalytic reactions. During initial period of the photocatalytic reaction, reduction cocatalyst Ag is used to trap electrons and serve as active sites for reduction reactions. Especially, the formation of a Schottky junction [metal–semiconductor (Ag–TiO2) junction] results in the electron transfer from TiO2 to metals. Consequently, metallic cocatalysts work as electron sinks and suppress the recombination of photogenerated charge carriers on the surface of TiO2 because of the Schottky barriers.[35] As the reaction goes on, the apparent color and the photocatalytic performance change drastically. It is speculated that a new electronic structure has been formed in Ag/T-5 after irradiation, which highly contributes to the photocatalytic activity enhancement of Ag/T-5.

Electronic Structure

To figure out the mentioned new electronic structure leading to color variance of Ag/T-5, the electron spin-resonance spectroscopy (ESR) was measured without and with simulated solar light irradiation in Figure a,b. Before illumination, no obvious signal was observed for the Ag/T-5 sample. Under simulated solar light irradiation, two signals attributed to Vo and Ti3+ appear. The strong peak with a g factor of 2.003 can be assigned to Vo,[36] while the weak signal with a g factor of 1.999 can be attributed to Ti3+.[37] The emergence of Vo in quantity results in the color change of Ag/T-5.[17,18] Huge difference in signal strength between these two signals makes the Ti3+ peak un-conspicuous. The wholesale deletions of Ti3+ species may be related to a redox reaction of TiO2 and monovalent Ag(I) species, Ag+ + Ti3+ → Ag0 + Ti4+, as reported redox reaction of TiO2 and Cu2+ species.[21] Accordingly, Ag species is firmly bonded on the surface of TiO2 via the powerful chemical interaction, which brings considerable benefits to the migration of photogenerated carriers as well as suppresses the recombination of electron–hole pairs.

In addition to ESR, XPS was employed to verify the existence of oxygen vacancies in Ag/T-5 after irradiation. The C 1s peak at 284.6 eV is used as the internal reference value to revise the surface charging effects. The O 1s core level spectra of Ag/T-5 before and after the photocatalytic H2 evolution reaction are displayed in Figure c,d. For the pristine Ag/T-5 in Figure c, the main peak and a weak peak located at 530.1 and 531.6 eV can be attributed to O–Ti in TiO2 and O around the oxygen vacancy,[38] respectively. After the photocatalytic reaction, the peak at 531.6 eV belonging to O around the oxygen vacancy distinctly strengthens in Figure d, which demonstrates that a certain amount of oxygen vacancies has been generated in Ag/T-5 after irradiation.

Because Ag/T-5 has been changed after the photocatalytic reaction, the XRD pattern, DRS spectrum, and SEM image of used Ag/T-5 are compared with that of fresh Ag/T-5 in Figure S4–S6, respectively. As shown in Figure S4, the used Ag/T-5 has almost same crystal structure of fresh sample, but slight intensity enhancement can be observed in the used sample. The slightly increased crystallization of used Ag/T-5 may be caused by longtime irradiation[39] in the photocatalytic process. In Figure S5, the DRS spectrum of Ag/T-5 changes a lot after photocatalytic reaction. On account of the apparent color transformation caused by Vo appearance, the absorption edge of used Ag/T-5 shifts toward long wavelength as previous work reported.[40] Besides better light adsorption in the visible region, a noticeable peak is found at 450 nm, which can be attributed to the SPR of Ag nanoparticles. It has been reported that SPR frequency depends not only on the metal but also on the size and shape of the nanoparticle, the dielectric properties of the surrounding medium, and inter-nanoparticle coupling interaction, thus imparting a unique tunability to the nanoparticle optical properties.[41] In this case, the enhanced SPR of used Ag/T-5 may be affected by the electronic structure changes of composites caused by abundant oxygen vacancies. Both light absorption enhancement and hot electron injection benefited from SPR can promote photocatalytic reaction efficiency. Apart from XRD and DRS, the SEM images of the used Ag/T-5 and the fresh sample are also compared in Figure S6. Imperceptible change is found between two samples, which illustrates the stable morphology of Ag/T-5 during the photocatalytic reaction. In addition, the Ag cations leaching during the photocatalytic H2 evolution reaction are also determined in Figure S7. Few Ag+ ions have been leached in ppb range during the photocatalytic process, which further confirms the stability of Ag/T-5 photocatalyst.

To further investigate the effects of oxygen vacancies in the recombination process of photogenerated electron–hole pairs, a series of characterization techniques are used. PEC measurements utilized to determine the separation efficiency and transfer characteristic of photogenerated carriers were carried out under the irradiation of simulated sun light. In Figure a, the current responses of catalysts are all distinctly affected by illumination, leading to high light–dark current ratios. The photocurrent density of Ag/T-5 reaches 26.8 μA cm–2 without bias voltage, which is about 4 times higher than that of T-5 and 9 times higher than that of T-2.

Figure 8

(a) Periodic on/off photocurrent response, (b) EIS, (c) LSV, and (d) PL spectra of T-2, T-5 and Ag/T-5.

The EIS curves in Figure b correspond with the photocurrent results, which further certifies that the photocurrent of samples are in sequence of Ag/T-5 > T-5 > T-2. It is widely accepted that the higher photocurrent signifies better separation efficiency of photogenerated carriers.[42,43] In consequence, T-5 shows highly efficient separation and speedy transfer of photogenerated electrons and holes than T-2, which can be attributed to the well-organized hierarchical flower structure and more rutile content for T-5. Ag/T-5 exhibits superior separation efficiency and transfer rate of carriers, demonstrating the synergistic facilitation effect of oxygen vacancies and SPR of cocatalyst Ag. The order of photocurrent highly in accordance with photocatalytic performance illustrates that the separation efficiency and transfer rate of carriers is a main factor affecting photocatalytic reaction.

Because the photoreduction ability has been intuitively affirmed by photocatalytic reactions, the photo-oxidation capacity can also be simply tested by linear sweep voltammetry (LSV) under the simulated sun light irradiation. As shown in Figure c, the LSV curve scans from −0.2 to 1.0 V. During −0.2 to 0.8 V district, the current density grows slowly as the voltage increases. When the voltage is over 0.8 V, the current density of catalysts rapidly rises. Especially for Ag/T-5, the current density reaches 2.3 mA cm–2 at a voltage of 1.0 V, which is about 2 times higher than that of T-5 and 3 times higher than that of T-2. In addition, the enhanced the current density of Ag/T-5 is inseparable from light irradiation, as shown in Figure S8. The current density of Ag/T-5 varies little with time and keeps a low value without illumination. Therefore, the photooxidation ability of Ag/T-5 is detected as the best one, which can also be ascribed to the superior separation efficiency and transfer rate of carriers caused by synergistic facilitation effect of oxygen vacancies and SPR of cocatalyst Ag, in line with the photoreduction property.

It is widely accepted that radiative recombination of photogenerated electrons and holes can give rise to fluorescence.[44] Hence, the PL spectra of catalysts were recorded with an excitation wavelength of 325 nm to examine the recombination of photogenerated carriers (Figure d). The PL spectra of these samples present similar peak positions with different intensities. T-5 with a higher crystallinity rutile phase and regular hierarchical flower-like morphology show lower PL peak intensity than that of T-2, signifying less electrons and holes recombining in T-5. Moreover, Ag/T-5 exerts the lowest PL intensity, which illustrates higher separation efficiency and less recombination of electron–hole pairs in Ag/T-5, corresponding to the above electrochemical results and photocatalytic performance.

The highly efficient electron transport and carrier separation are speculated to have relationship with the oriented migration of photogenerated carriers in Ag/T-5 hybrid. In order to investigate the directional transport of carriers under illumination, the XPS spectra of Ag/T-5 before and after the photocatalytic H2 evolution reaction are compared in Figure . For the initial Ag/T-5 sample, two obvious peaks can be found in the Ti 2p core level spectra of Ag/T-5 in Figure a. The peaks located at 458.8 and 464.3 eV can be attributed to Ti 2p3/2 and Ti 2p1/2, respectively. These binding energies are identical to that of bulk TiO2, as reported previously.[45]Figure b shows the Ag 3d XPS spectra, and more than one valence state are perceived for Ag. The binding energies of Ag 3d located at 368.2 and 374.2 eV are attributed to Ag0 3d5/2 and Ag0 3d3/2, while the other two peaks at 367.3 and 373.3 eV can be attributed to Ag+ 3d5/2 and Ag+ 3d3/2, respectively.[46] Most of loaded Ag is monovalence silver, and the ratio of Ag+ and metal Ag is evaluated as 3:1 according to their corresponding peak areas. After the photocatalytic reaction, the two peaks in Ti 2p core level spectra of Ag/T-5 shift 0.1 eV to higher binding energy as shown in Figure c, which illustrates that electron density of titanium atom decreases after irradiation, whereas the ratio of Ag+ gets higher and reaches 94% in used Ag/T-5 as Figure d exhibited.

Both the increase ratio of Ag+ in Ag/T-5 and the positive shift of binding energy in the Ti XPS spectrum indicate the electron decrease, demonstrating that the hot electrons migrate from Ag to Vo instead of TiO2. In this case, the photogenerated electrons of TiO2 and the hot electrons from Ag nanoparticles are simultaneously enriched in Vo for highly efficient photocatalytic H2O reduction.

Because the compositions of the Ag/T-5 has been changed, the photocatalytic H2 evolution activity of Ag+/T-5 is also explored as comparison in Figure S9. Ag+/T-5 photocatalyst was prepared by dropping NaOH (4 M) into the mixture of T-5 and AgNO3 solution to load Ag2O nanoparticles on T-5. Ag+/T-5 shows similar photocatalytic H2 evolution activity with unmodified T-5, indicating that Ag+ is not the active component that affects the photocatalytic performance. Moreover, the obvious color transformation of Ag/T-5 during the photocatalytic process is not observed on Ag+/T-5. According to this result, it can be confirmed that Ag0 acts as the most active species in producing Vo and optimizing the photocatalytic performance. In this work, the thermal reduction method is used to synthesize Ag/T-5 with more Ag0. Unfortunately, Ag0 nanoparticles are easilyto be oxidized in the air,[47] so high-content Ag+ is found in as prepared Ag/T-5. However, fortunately, the oxidized Ag contact with air can protect the inner Ag0 that contact with TiO2 to keep its chemical state.

Reaction Mechanism

To confirm the main active species during the photodegradation process, DMPO spin-trapping ESR was used to detect the active radical of T-5 and Ag/T-5. As shown in Figure a, the typical quadruplet with a signal-to-intensity ratio of 1:1:1:1 indicates that superoxide radicals (•O2–) are generated in the both samples under UV light illumination.[48] Ag/T-5 with more photogenerated •O2– radicals exhibits superior photocatalytic performance of RhB photodegradation than unmodified T-5. No signal of hydroxyl radicals (•OH) can be found in DMPO spin-trapping ESR spectra for DMPO–•OH, which indicates that •O2– acts as the main active radical in the photodegradation process.

Figure 10

(a) DMPO spin-trapping ESR spectra for DMPO–•O2– of T-5 and Ag/T-5 samples under UV light irradiation; (b) ESR spectra of Ag/T-5 under simulated sunlight irradiation for 1 and 3 h.

To investigate the relationship of Vo concentration and photocatalytic activity, the ESR spectra of Ag/T-5 under irradiation for 1 and 3 h are compared in Figure b. Ag/T-5 under irradiation for 3 h shows enhanced signal intensity of Vo than Ag/T-5 under irradiation for 1 h, corresponding to the higher photocatalytic H2 evolution rate at 3 h than that at 1 h in Figure b. At the initial period of photocatalytic reaction, a higher concentration of Vo is beneficial for the photocatalytic performance.[49,50] However, the formation of Vo would reach a dynamic equilibrium as the experiment goes on, and a relatively stable concentration of Vo is also conducive to the stable photocatalytic reactivity.

Combined with the previous characterizations, the reaction mechanism of photocatalytic H2 evolution and RhB degradation over Ag/T-5 are illustrated in Figure . On account of huge alterations in Ag/T-5, there is significant change in the photocatalytic mechanisms of fresh Ag/T-5 and Ag/T-5 with Vo. The photocatalytic mechanisms for H2 evolution reaction are illustrated in Figure a. At initial period of the photocatalytic reaction, photogenerated electrons directionally migrate from TiO2 to Ag driven by the Schottky junction (Ag–TiO2).[35] The low recombination chance of carrier profits from Schottky barrier improve the carrier separation efficiency, thus increasing the photocatalytic reactivity. Ag nanoparticles enriched with electrons serve as active sites for H2O reduction reaction to produce H2. The photogenerated holes move from bulk to the surface of TiO2 and oxidize sacrificial agent CH3OH to the products such as HCHO and HCOOH.

Figure 11

Reaction mechanism of (a) photocatalytic H2 evolution and (b) photocatalytic RhB degradation over Ag/T-5.

After a period of irradiation, oxygen vacancy (Vo) has been formed in Ag/T-5. Besides the changes in sample color and light absorption properties, the enhanced SPR of Ag nanoparticles makes the photocatalytic mechanism different and highly improves catalytic efficiency. Generally, Vo is considered to be a favorable electron captor and an active reduction site in photocatalytic process.[51] The hot electrons arise from SPR can inject from Ag to Vo to participate in the H2O reduction reaction.[52] The increasing ratio of Ag+ in Ag/T-5 indicates the electron decrease of Ag. In addition, the binding energy of Ti shifts to higher energy, which demonstrates the electron cloud density decrease of TiO2 (Figure c). These results further certify that the hot electrons are trapped by Vo instead of staying on TiO2. The photogenerated electrons of TiO2 and the hot electrons from Ag nanoparticles are simultaneously enriched in Vo for photocatalytic H2O reduction. The oxidation of sacrificial agent CH3OH remain unchanged in the new mechanism.

The photocatalytic mechanisms of RhB degradation reaction are exhibited in Figure b. The migration paths of electrons and holes are similar to H2 evolution process. Therefore, the point of the discussion is the photocatalytic mechanism of RhB degradation. It has been confirmed that RhB is stable under simulated sunlight irradiation without photocatalyst.[53] However, in the presence of the photocatalyst, the main reactions follow the equations below.

RhB molecules absorb photons and get excited to generate RhB*. Simultaneously, the electrons on Vo react with adsorbed oxygen molecules to produce •O2– as reactive oxygen species (ROS). Then, ROS •O2– attack dye molecules and destroy auxochromic groups to form the N-de-ethylation/de-ethylation of the alkyl amine group,[54] and the following degradation leads to the ring structure destruction.

Overall, we can conclude that the hierarchical flower-like Ag/T-5 assembled by nanorods was obtained successfully. As reported, the large aspect of nanorods is expected to enhance the separation of photogenerated carriers,[55] which can improve the effective carrier concentrations and be favorable for photocatalytic reactions. The formation of hierarchical flower-like structures is supposed to prevent the disorderly arrangement of nanorods, contributing to the rapid mass transfer in photocatalytic processes. Besides the advantages of morphology, abundant oxygen vacancies are generated under simulated solar irradiation. It has been reported that oxygen vacancies can give rise to distortion of crystal lattice and fabricate appropriate defects on the surface of catalyst, which on the contrary enhances the structural stability.[42,56,57] In addition, Vo can also optimize the light absorption and enhance the SPR of cocatalyst Ag. The hot electron flow from Ag to Vo, which is benefited from the enhanced SPR, changes the carrier transfer mechanism and supplies abundant electrons for reactive sites. The synergistic effect of Vo and the SPR of Ag can further promote the concentration, migration, and separation efficiency of photogenerated carriers, thus achieving a high and stable photocatalytic performance of Ag/T-5.

Conclusions

In summary, hierarchical flower-like TiO2 was synthesized through acidity regulation of hydrothermal reaction. For T-5, the regular morphology together with high degree of crystallinity and rutile phase content directly affects its light absorption properties and electronic band structure, leading enhanced separation efficiency and migration rate of photogenerated carriers for superior photocatalytic performance. With the help of cocatalyst Ag, high concentration of oxygen vacancy in Ag/T-5 under illumination brings higher absorption of visible light and enhanced SPR of Ag. Abundant electrons on Vo, including the photogenerated electrons from TiO2 and the hot electrons from the SPR of Ag, make Vo an effective active site for photocatalytic reduction. This work proposes some perspectives for the intensive comprehension of the hierarchical flower-like structure and oxygen vacancy effects on the photocatalytic performance. Hence, it exploits a new strategy for acidity-controlled synthesis of a hierarchical structure and precious metal-induced formation of oxygen vacancies, obtaining efficient materials for photocatalytic processes. It is expected that the convenient methods for constructing the hierarchical morphology and a special electronic structure of catalysts have enormous potential to be widely applied in the energy and environmental-related fields.