F- Serve as Surface Trapping Sites to Promote the Charge Separation and Transfer of TiO2.

Finding an effective strategy to promote the charge transfer and separation of TiO2 is urgently needed. Herein, a surface fluorination (F-)-modified TiO2 (denoted as TO-xF, where x represents the volume of HF added in the solution) catalyst has been prepared by a mild and facile post-treatment method. The changes induced by surface F- on the morphological, structural, and surface electronic features and the charge separation and transfer efficiency of TiO2 were specifically examined. Compared with pristine TO, TO-0.4F exhibits enhanced photocatalytic degradation of methyl orange and phenol, production of hydroxyl radicals, and photocurrent response. The enhanced photocatalytic activities of TO-0.4F can be attributed to the role of surface F- as surface trapping sites in effectively boosting the charge transfer and separation processes, which is verified by the steady-state and time-resolved fluorescence spectroscopy, electrochemical impedance spectroscopy, Bode plot, transient photocurrent response, and open-circuit voltage measurements. This study emphasizes the role of surface F- in promoting the charge transfer and separation and improving the photocatalytic activity of TiO2.

Introduction

As a typical semiconductor, titanium dioxide (TiO2) has received extensive attention due to its high stability, nontoxicity, facile and low-cost synthesis, and tunable electronic band structure features. Particularly, TiO2-based semiconductor materials have been extensively used in a wide range of applications, such as H2 production,[1−3] CO2 reduction,[4−7] nitrogen fixation,[8−10] wastewater treatment,[11−13] gas sensors,[14,15] and solar cells.[16,17] However, the ineffective separation and transport of charge carriers as a consequence of a shorter carrier lifetime largely restricts its application of photogenerated carriers for photoelectrochemical and chemical reactions.[18−20] Thus, research studies were tremendously intensified to overcome these bottlenecks, and momentous advancements were attained over the past few decades. Generally, strategies proposed to boost the separation and transport efficiency of charge carriers were developed, such as morphology modulation,[21−23] crystal facet engineering,[24−26] element doping,[27−30] heterostructure construction,[18,31−34] co-catalyst loading,[35,36] plasma enhancement,[37,38] and combinations thereof.[39] Though these approaches were devoted to facilitate charge transfer to a certain degree, the vast majority of photoinduced electrons and holes still recombined during the charge transportation process, and the photocatalytic performance of TiO2 was still far from satisfactory. Hence, it is of great significance to develop highly efficient TiO2.

Recent research results show that the post-treatment procedure after sample preparation has emerged as an attractive strategy to effectively address the issue of separation and transport of charge carriers over semiconductor photocatalysts.[40,41] For example, a facile and efficient Cl– surface modification method has been demonstrated earlier to drastically enhance photocatalytic water oxidation over a BiVO4 catalyst.[40] Due to the acceleration of photogenerated holes by Cl–, as-prepared Cl–-modified BiVO4 achieved excellent oxygen evolution, whose apparent quantum efficiency could reach 34.6% at 420 nm. Moreover, Cl–-modified Bi2WO6 nanosheets were prepared to examine their photocatalytic oxidation performance.[42] It is surprising to find that the CH4 yield of Cl–-modified Bi2WO6 was almost 10 times of that obtained with the pristine Bi2WO6 one. These surface-modified catalysts with higher photoreactivity offer insights and provide more opportunity to promote the efficiency of separation and transport of carriers and enhance the photoreactivity of semiconductor photocatalysts.

Compared with Cl–, fluorine (F–) possesses the strongest electronegativity of 3.98, and it readily reacts with metal elements and oxidizes them to their highest valence states. Moreover, F– has been widely used as a morphology directing agent to stabilize the {001} facet of TiO2 since its initial use by Yang et al.[24] The effectiveness of F– in obtaining TiO2 crystals with a dominant {001} facet by forming the Ti–F bond to saturate surface Ti5c has been confirmed by numerous studies.[25,43−45] Though fluorine can be partially removed by either washing with a basic solution or a calcination process, the role of residual F– in influencing the electronic structure, charge transfer and separation behaviors and, ultimately, the photocatalytic activities of TiO2 cannot be neglected. Moreover, fluorine can be introduced into TiO2 to generate two forms of fluorine species: surface fluorine and lattice fluorine. The former is proved to influence the adsorption capacity, provide surface trapping sites, and regulate the charge transfer and storage.[46,47] The latter is commonly considered to form defective centers (Ti3+ or oxygen vacancy species) and introduce localized electronic states. F– existing on the surface could affect the electronic structure of TiO2 and exchange with the surface hydroxyl group (−OH). For instance, reported works showed that the displacement of the −OH group by F– could regulate the kinetics and mechanism over the photocatalytic degradation of phenol.[48] Moreover, F– may suffer from redox transformations after electron transfer with carriers and remarkably affect the photocatalytic activities depending on their affinity to the surface of the specified semiconductor. Consequently, the photoreactivity of surface F–-treated TiO2 has been reported either to increase[49,50] or to decrease[51] compared with those of the pristine TiO2 ones. Even so, evidence of the improved photocatalytic efficiency as well as its mechanism over surface-fluorinated TiO2 is obscure and remain to be explored thoroughly, to obtain a high solar energy conversion and particularly a high-performance TiO2 photocatalyst.

Herein, we report a facile F– surface modification method, which not only improves the efficiency of charge separation but also accelerates the charge-transfer process. Compared with pristine TiO2 (denoted as TO), F– surface-modified TiO2 (denoted as TO-xF, x represents the volume of HF added in the solution) of the TO-0.4F sample exhibits enhanced and optimal photocatalytic degradation of methyl orange (MO), photocatalytic production of the hydroxyl radical (HO•), and photocurrent response under UV light irradiation. Moreover, the improved charge carrier separation and transfer efficiency by surface F– was verified by steady and time-resolved photoluminecence (PL) spectroscopy, electrochemical impedance spectroscopy (EIS), Bode plot, transient photocurrent response, and open-circuit voltage (UOC) measurements.

Materials and Methods

The detailed information of preparation, characterizations, photocatalytic generation experiment of •OH radicals, photocatalytic degradation of methyl orange (MO) and phenol, and photoelectrochemical measurements over as-prepared F–-modified TiO2 can be found in the Supporting Information.

Results and Discussion

The crystal structures of TO and TO-xF catalysts were determined by powder X-ray diffraction (XRD), as shown in Figure a. Both anatase and rutile phases are observed over pristine TO2. More specifically, diffraction peaks at 2θ = 25.28, 36.95, 37.80, 38.58, 48.05, 53.89, 55.06, 62.69, 68.76, 70.31, 75.03, and 76.02° (JCPDS NO. 21-1272) are typically anatase TiO2, while 2θ = 27.45, 36.09, 41.23, 44.05, and 56.64° (JCPDS NO. 21-1276) are peaks of rutile TiO2. Moreover, it can be found that anatase is the dominant crystal phase in pristine TO. For TO-xF samples, peaks of anatase and rutile TiO2 are also detected at the corresponding positions, while no other new diffraction peaks can be observed, which implies that F– may have a negligible impact on the crystal structure of TiO2. Besides, no diffraction peak shift is observed in TO-xF, suggesting the maintained crystal structure during the F– post-treatment process. Figure b shows that all the TO and TO-xF catalysts exhibit typical Raman peaks at 143, 396, 516, and 638 cm–1, which are assigned to Eg(1), B1g(1), A1g + B1g(2), and Eg(2) vibration bands of anatase TiO2, respectively. No signals of rutile TiO2 could be observed, which can be explained by the relatively lower contents of rutile TiO2 which were concealed by the strong Raman signals of anatase TiO2.

Figure 1

(a) XRD patterns and (b) Raman spectra of TO and TO-xF.

The surface morphology and microstructures of as-prepared samples were observed by scanning electron microscopy (SEM) and transmission electron microscopy (TEM) analyses. As shown in Figure S1a, the TO sample displays an irregular morphology of nanoparticles, and the average particle size was about 20–25 nm. Moreover, the surface F– modification has a negligible influence on the morphology and particle size, as shown in Figure S1b–d. Furthermore, TEM images analysis further confirms that the TO catalyst consists of irregular nanoparticles with an average particle size of 25 nm, as shown in Figure a. It can be seen from Figure b that pristine TO exhibits three obvious lattice fringes with spacings of 0.35, 0.24, and 0.33 nm, which corresponds to the (101) and (001) facets of anatase TiO2 and (110) facet of rutile TiO2, respectively. After F– surface modification, the micromorphology of TO-0.4F in Figure c has no obvious changes compared with pristine TO. Similarly, the lattice fringes of both anatase and rutile TiO2 appear in Figure d. Moreover, no lattice fringes of bulk fluoride species can be observed, which indicates that HF post-treatment does not result in the bulk fluorination of the TO catalyst in our present work. To further verify the components of the F–-treated TO sample, the energy-dispersive X-ray (EDX) elemental mapping images of the TO-0.4F catalyst are provided. As can be seen from Figure a–d, Ti, O, and F elements are homogeneously distributed in the TO-0.4F catalyst. In addition, the EDX patterns of as-prepared TO-0.4F are also performed, as shown in Figure S2.

Figure 2

(a)TEM and (b) HRTEM images of TO. (c) TEM and (d) HRTEM images of TO-0.4F.

Figure 3

TEM image (a)and EDX mapping of Ti (b), O (c), and F (d) elemental distribution over TO-0.4F.

The effect of F– toward the surface element compositions and states over the prepared catalysts is examined by XPS characterization. As depicted in Figure a, the TO catalyst exhibits two typical XPS peaks located at 458.39 (Ti 2p3/2) eV and 464.11 eV (Ti 2p1/2), indicating the existing Ti4+ state in the crystal structure of TiO2.[34] Notably, the characteristic Ti4+ peaks shift to a higher and then lower binding energy positions after F– modification on the surface of TiO2. More specifically, the Ti 2p3/2 position of TO-0.1F, TO-0.4F, and TO-1.6F catalysts shifts to 458.48, 458.56, and 458.22 eV, respectively. This obvious shift may come from a charge imbalance induced by F– modification on the surface of TiO2.[52] Moreover, this suggests that the electron cloud density of Ti on the surface of TO-0.1F and TO-0.4F is slightly lower than that of Ti on the surface of pristine TO, which results from the electron transfer from Ti to F. However, too much F– on the surface of TiO2 is detrimental to the transfer of electron to F. The XPS results further confirm the existence of the electron interaction between F and Ti in the TO-xF catalysts. In addition, O 1s spectra of pure TO catalysts exhibit two XPS peaks at 531.51 and 529.63 eV, which are ascribed to the surface-adsorbed hydroxyl oxygen (OOH) and lattice oxygen(Olattice), respectively. Similarly, the XPS peaks of O 1s (Figure b) for TO-0.1F and TO-0.4F also shift positively and then negatively for TO-1.6F, further indicating that F– might have a vital influence on the surface chemical environment of TiO2. Considering the crucial role of the hydroxyl group in influencing the catalytic performance of the TiO2 catalyst, the proportion of the hydroxyl group in all oxygen atoms is estimated based on the corresponding XPS peak area, as summarized in Table S1. The percentage of hydroxyl oxygen atoms for TO is estimated to be 28.36%, while the value decreases to 21.26, 20.53, and 19.89% for TO-0.1F, TO-0.4F, and TO-1.6F, respectively. The reason for the decreased surface hydroxyl of TO-xF can be attributed the displacement of surface hydroxyl by F– ions. It was reported that the displacement of surface −OH by F– could change the adsorption and the surface interactions (e.g., the adsorption of oxygen followed by the production rate of superoxide would be strongly affected).[48,53] As shown in Figure c, compared with pure TO, TO-xF samples exhibit broad binding energy peaks centered ca. 684 eV, which are ascribed to the F species adsorbed on the TiO2 surface.[24,54,55] Moreover, no lattice F or substituted F[27,56] (the corresponding XPS peaks usually centered at ca. 688 eV) of oxygen in the crystal surface of TiO2 can be observed in the TO-xF catalyst. The band structure is very important for the photoreactivity because it largely influences the redox potential of the semiconductor photocatalyst. Therefore, the XPS valence band (VB) is examined, as shown in Figure d. The estimated VB positions of TO, TO-0.1F, TO-0.4F, and TO-1.6F are located at 2.42, 2.52, 2.59, and 2.29 eV, respectively. The shift of the VB energy level is possibly due to the F– substitution of −OH on the surface of TiO2.[47]

Figure 4

High-resolution XPS spectra of Ti 2p (a), O 1s (b), F 1s (c) and VB XPS spectra (d) for TO and TO-xF.

The optical absorbance properties of TO and TO-xF were examined. Figure a shows the UV–vis diffuse reflectance spectroscopy (DRS) spectra of TO, TO-0.1F, TO-0.4F, and TO-1.6F. TO-xF catalysts exhibit a slight shift of the absorbance toward the UV range compared with pristine TiO2, which indicates the decreased utilization of the visible light region solar energy. The band gap energy (Eg) is estimated through an extrapolation method by fitting the linear part of the plot, as shown in Figure b. The calculated Eg are 2.99, 2.94, 2.92, and 2.91 eV for TO-1.6F, TO-0.4F, TO-0.1F, and TO, respectively.

Figure 5

(a) Plot of UV–vis DRS and (b) Kubelka–Munk function plotted against the band gap energy for TiO2.

Figure S3 displays the FTIR spectra of as-prepared TO and TO-xF catalysts. The region of the broad reflectance band at 3350 cm–1 for TO and TO-xF is generally ascribed to the characteristic band of the O–H stretching vibration related to the physical adsorbed water. Most importantly, compared with pure TO, the peak at 3350 cm–1(O–H stretching) for TO-xF is markedly strengthened and, in particular, the peak intensities of TO-0.1F and TO-1.6F are stronger, indicating that surface-modified F– could affect the concentration of the surface-adsorbed OH groups. This can be further verified by the enhanced vibration peak at 1632 cm–1 for TO-0.1F, TO-0.4F, and TO-1.6F, which is an indicator of the O–H bending mode of water.[57] The FTIR results are well in accordance with our XPS result, further highlighting the role of F– in affecting the surface OH groups of TiO2. The adsorption of water by TiO2 may be due to the presence of OH groups, or the stronger electron-acceptor centers, that is, the presence of coordinated unsaturated Ti4+ sites.[58] Additionally, the signals of the adsorbed CO2 and related carbonate on the surface of TiO2 at 2347, 1412, and 1339 cm–1, which belong to an asymmetric stretching vibration of C–O, stretching and bending vibrations of C–OH and C–O bonds, respectively. Moreover, a sharp and strong peak centered near 673 cm–1 is observed for TO and TO-xF, which can be attributed to the Ti–O stretching band for the characteristic peaks of TiO2 vibration.[59,60]

The specific surface areas and pore sizes of the as-prepared catalysts were determined from the nitrogen adsorption–desorption isotherms, as shown in Figure S4. All the TO and TO-xF samples show a typical IV adsorption–desorption isotherm with H2-type hysteresis loops, suggesting the presence of mesoporous structures. The specific surface areas calculated from the BET method are 54.00, 54.02, 50.86, and 55.24 m2·g–1 over TO, TO-0.1F, TO-0.4F, and TO-1.6F, respectively. In addition, by using the Barrett–Joyner–Halenda equation, the calculated average pore size is about 30 nm for TO and TO-xF samples. The specific surface area, pore volume, and average pore size of TO and as-prepared TO-xF catalysts are summarized in Table S2. Because the calculated porosity is equivalent to the size of the basic particles, it can be considered that the obtained mesopores come from substantial voids between the TiO2 nanoparticles. Based on the above results, we confirm that surface F– does not induce obvious textural changes of pristine TiO2, which is well consistent with the SEM and TEM analyses.

The photocatalytic degradation of methyl orange (MO) under UV light is carried out, which is shown in Figure a. The pristine TO sample shows a high degradation efficiency of 79% toward MO after UV light irradiation for 40 min. In contrast, surface modification with a small amount of F– on TO leads to an enhanced degradation efficiency of 89% for TO-0.1F. Furthermore, TO-0.4F shows the highest degradation efficiency of 95% among the F–-modified TiO2 samples. However, the presence of too much F– on the surface of TiO2 brought about a noticeable decline of the photocatalytic degradation efficiency of 51% for the TO-1.6F catalyst. The activity of one catalyst is generally the result of interactions of various factors. In the current case, because the textural and morphological features of TiO2 are almost retained before and after surface fluorination, the higher or lower photocatalytic efficiency of TO-xF can be related to the surface and optoelectronic features. The fitted curves of −ln(C/C0) versus reaction time were plotted to determine the reaction rate constant (k), as shown in Figure b. Specifically, the k values are 3.94 × 10–2, 6.05 × 10–2, 6.60 × 10–2, and 1.79 × 10–2 min–1 for TO, TO-0.1F, 0.4F, and 1.6F, respectively. The optimal k for TO-0.4F is about 1.67 times that of pristine TO, demonstrating that surface modification by F– is beneficial to enhance the photocatalytic activity of TiO2. In addition, the colorless phenol was also used as a model pollutant to evaluate the photocatalytic activity of the as-prepared TO and TO-xF samples, as shown in Figure S5. Similarly, the TO-0.4F catalyst exhibits an excellent and higher photocatalytic degradation rate of phenol, compared with that of pristine TO, TO-0.1F, and TO-1.6F.

Figure 6

(a) Degradation performance of MO; (b) fitted curves of −ln(C/C0) vs reaction time; (c) comparison of •OH radicals recorded from fluorescence spectra; and (d) temporal profiles of •OH radicals over TO, TO-0.1F, TO-0.4F, and TO-1.6F.

The photocatalytic redox activities were further investigated by the production of hydroxyl radicals (•OH) under UV light irradiation. The relative concentration of •OH is determined by examining the PL intensity centered at 426 nm, as shown in Figure c. TO-0.4F and TO-0.1F exhibit higher •OH generation than pure TO, and TO-0.4F shows the highest •OH production performance among the series samples, as shown in Figure d. The above results demonstrate that the photoinduced oxidation activity is enhanced in TO-xF as compared with pure TO. The photocatalytic degradation reactions generally proceed either directly through photogenerated holes or indirectly through oxidation mediated by the •OH radical. Due to the much higher redox potential of the F•/F– couple (3.6 V), the holes in the VB cannot oxidize the fluorine groups on the TiO2 surface. Therefore, more •OH radicals are generated in the TO-xF catalyst to participate in the photocatalytic reactions. Thus, the hole-induced oxidation of water-producing •OH radicals is highly favored. When the surface of TiO2 is hydroxylated, the reaction precedes either through the surface trapping of holes or direct electron transfer. In contrast, when the surface is covered by F–, the kinetic path of the reaction is dominated by the sub-surface holes and free •OH.[48] Therefore, most of the reaction proceeds on the fluorinated TiO2 through the free •OH radical pathway. However, a too high concentration of surface F– becomes a new recombination center for photogenerated charge carriers, which leads to lower photocatalytic activity and production rates of •OH radicals. To examine the stability features, three cycles of degradation of MO were tested for TO and TO-0.4F, as shown in Figure S6. It can be seen that the MO degradation efficiency over TO-0.4F after cycling three times is still retained about 95%, which is much higher than that of pristine TO (ca. 80%).

In order to trace the origin of the enhanced photocatalytic activity over TO-xF, the PL spectra of TO and TO-xF was first examined to evaluate the charge separation behavior. Figure a shows the PL spectra of pristine TO and TO-xF catalysts. It is worth noting that the TO and TO-xF samples exhibit a strong and wide PL signal in the range from 350 to 550 nm, having two obvious PL peaks at about 420 and 480 nm, which are attributed to band-edge-free excitons and bound excitons, respectively.[61] These PL signals are generally attributed to excitonic PL signals, which mainly result from surface oxygen vacancies and defects of TiO2 nanoparticles.[62] In addition, it can be found that an appropriate amount of surface F– modification (such as TO-0.1F and TO-0.4F) can efficiently promote the separation of photogenerated carriers and thus reduce the excitonic PL signals. However, too high a surface fluoride ion will form more surface defect sites, which acts as new recombination centers of carriers and leads to enhanced excitonic PL signals (such as TO-1.6F). The above discussions demonstrate that the weaker the excitonic PL spectrum, the higher the separation rate of photoinduced electrons and holes, during the F– surface-modification of TiO2. The optimal F– on the TO surface can boost the transport and separation of carriers, thereby improving its photocatalytic performance. To further ascertain the key role of F– toward the separation of photoinduced charge carriers, the transient photocurrent response was established under UV light irradiation. As depicted in Figure b, the transient photocurrent densities of TO-0.1F and TO-0.4F increase up to 55 and 69 μA cm–2, which are 1.1 and 1.4 times that of pristine TO (48 μA cm–2), respectively. These excellent photocurrents further verify that F– surface modification of TiO2 can remarkably boost the separation of photogenerated charge carriers. However, compared with that of TO, the TO-1.6F sample exhibits a decreased photocurrent density of only 47 μA cm–2, which is consistent with the PL results. In addition, EIS measurements were further conducted to examine the charge-transfer process of as-prepared TO and TO-xF samples. As depicted in Figure c, the arc radius of TO-0.4F, which is attributed to the whole charge-transfer resistance (Rct), is the smallest, further confirming that the optimal amount of F– is favorable to promote the transfer of photoinduced charge carriers. The fitted values of Rct (Table S3) for TO, TO-0.1F, TO-0.4F, and TO-1.6F are 913.1, 294.9, 289.9, and 1053 Ω, respectively. This conclusion can be further validated by the Mott–Schottky measurement, as shown in Figure S7. More importantly, the calculated flat band potentials (vs NHE) are in the order of TO (−0.12 V) < TO-0.1F (−0.10 V) = TO-1.6F (−0.10 V) < TO-0.4F (−0.08 V). Further analysis shows that a positive shift in flat band potentials of TO-xF compared with that of the TO catalyst can be found, which indicates the decrease in the bending of the band edge and acceleration of the separation and transportation of photogenerated carriers. Moreover, the Mott–Schottky plots for TO and TO-xF depict a positive slope, indicating that they are all n-type semiconductors. Based on the slope of Mott–Schottky curves, the calculated carrier density of TO-0.4F is 14.65 × 1019 cm–3, which is remarkably higher than that of pure TO (11.71 × 1019 cm–3), as shown in Figure S8. The higher carrier density implies that the carrier separation and transport for TO-0.4F is faster than that of pure TO. Moreover, the Bode plot (Figure d) is used to reveal the charge-transfer behaviors over bare TO and TO-xF catalysts. Compared with bare TO, the frequency peak of TO-0.4F shifts to a lower frequency position, revealing a faster electron transport process. According to the correlation between the frequency (f) and the lifetime of injected electrons: τe = 1/(2πf),[63,64] the calculated τe for TO-0.4F is 114.6 μs, which is much longer than those of pure TO (75.6 μs),TO-0.1F(98.5 μs), and TO-1.6F(71.1 μs). Based on the above analysis, we confirm that the presence of an optimal amount of F– on the surface could facilitate the charge separation and transfer of TiO2, which is indispensable to enhancing its photocatalytic activities. Moreover, due to the strong electronegativity, F– could serve as a surface trapping site to extract the photogenerated holes from the bulk to the surface of TiO2, which facilitates the separation of charge carriers and increase the possibilities of the photocatalytic reaction. However, too much F– could serve as a new recombination center for photogenerated charge carriers and lead to a decreased lifetime of the photogenerated electrons.

Figure 7

(a) Steady PL spectra; (b) i–t curves; (c) EIS Nyquist plots; and (d) Bode plot of as-prepared TO-xF.

UOC was measured to investigate the surface charge trapping of TO and TO-xF catalysts, as shown in Figure . When one electrode is immersed in the electrolyte, the increase of photoinduced UOC largely depended on the photoinduced charge carrier density, thereby generating a new quasi-Fermi level. Generally, the photovoltage difference under light and dark conditions can directly reflect the potential loss of reaction kinetics in the Helmholtz layer on the surface of the catalyst.[65] The photovoltage for TO-0.4F is detected to be 0.29 V, which is 0.25 V higher than that of pure TO. The increased UOC over TO-0.4F implies that F– surface modification can boost surface charge trapping, thereby leading to both the higher separation efficiency and transfer efficiency of charge carriers.

Figure 8

Open circuit potentials (UOC) of TO and TO-xF tested under dark (solid, black) and illumination (hollow, red); inset: transient photovoltage response to immediate illumination.

The PL decay curves shown in Figure could provide information about the recombination of free charges located near the edge of the band, rather than the recombination caused by deep trapped charges.[66,67] Compared with pristine TO, all the TO-xF samples give two radiative lifetimes over the deconvolution of the fluorescence decay spectra, as listed in Table S3. As for TO-0.1F, TO-0.4F, and TO-1.6F catalysts, the population of the short-lived charge carriers (1.02, 1.08, and 0.98 ns for TO-0.1F, TO-0.4F, and TO-1.6F) is greatly reduced (31.2, 32.2, and 25.2% for TO-0.1F, TO-0.4F, and TO-1.6F, respectively) compared with that of pure TO (100%), which results from the rapid recombination of photoinduced electrons and holes. However, compared with pure TO, the percentage of charge carriers with a lifetime longer than 2.5 ns for TO-0.1F and TO-0.4F tremendously increases, respectively, to 68.8 and 67.8%, which is beneficial to increase the possibility of charge carriers to participate in the photocatalytic reaction. The change in the percentage of short/long-lived charge carriers might be related to the localized defective sites associated with F– in the energy band of TiO2. In addition, the high concentration of long-lived trapped electrons and holes increases the probability of contact with the reactants, thereby improving the photocatalytic activity of TiO2. In addition, the calculated average carrier lifetime(τ) order is as follows: TO-0.4F (2.55 ns) > TO-0.1F (2.08 ns) > TO-0.4F (1.45 ns) > TO-0.4F (1.43 ns).

Figure 9

Time-resolved PL decay spectra over TO and TO-xF.

To compare the energy band structure differences between pristine TO and TO-xF, the energy levels of both conduction band (CB) and valence band (VB) potentials are provided, as shown in Figure S9. According to the formula Eg = Ev – Ec (Eg: band gap, Ev: VB potential, and Ec: CB potential), the CB potentials of TO and TO-0.4F were calculated to be −0.49 and 0.35 eV, respectively. Moreover, it can be found that the VB potential of TO-0.4F downward shifts to 2.59 eV compared with that of pure TO (VB maximum at 2.42 eV). The downward shift of the VB implies that the oxidizing ability of the photogenerated holes in TO-0.4F may become stronger, which is beneficial to the photocatalytic oxidation reactions. Based on the above-mentioned characterization analysis and band structure, a possible photocatalytic mechanism over the TO-xF catalyst is proposed, which is illustrated in Figure . As can be seen from Figure a, by coupling F– species with the strong electronegativity on the surface of TO, the photogenerated holes could be trapped, and the migration and separation of photogenerated electrons to reactive sites to participate in the redox reactions are highly promoted. As for the photoelectrochemical reaction, it is reasonably inferred that F– on the surface of TO inhibits the recombination rates of charge carriers and facilitates electron–hole pairs of separation and hole transfer, which finally leads to the enhancement of the photoelectrochemical performance. The F–-enhanced surface trapping sites and hole transfer through the solid/liquid interface is shown in Figure b,c.

Figure 10

(a) Proposed mechanism for F–-promoted surface charge separation of TO-F. Schematic diagram of F–-modified TO-promoted carrier charge separation and accelerated photoelectrochemical performance for (b) TO and (c) TO-F photoanodes.

Conclusions

In summary, a series of surface F–-modified TiO2 (P25) catalysts were prepared by a facile post-treatment at room temperature. The effects of F– toward the geometrical morphology, chemical state, optical absorption, surface electronic structure, and energy band alignment of TiO2 are systematically analyzed. Compared with pure TO, the TO-0.4F sample exhibits enhanced photocatalytic degradation of MO and phenol, production of OH radicals, and photocurrent response irradiated under ultraviolet light. The improved photocatalytic performance of TO-0.4F can be attributed to the fact that F– acts as surface trapping sites to promote the transfer and separation of photogenerated charge carriers, which is validated by steady PL spectra, EIS plot, Bode plot, open circuit potential measurement, and time-resolved PL decay spectra. This work highlights the role of surface F– modification in boosting the separation and transfer of charge carriers and ultimately enhancing the photocatalytic activity of TiO2.