H₂-oxidation driven by its behavior of losing an electron over B-doped TiO₂ under UV irradiation

Abstract

In this work, TiO₂ was modified by doping the electron-deficient B element, and then the gas-sensing response of B-TiO₂ to H₂ under UV irradiation at room temperature in a N₂ atmosphere and the oxidation of H₂ over B-TiO₂ under corresponding conditions were tested. It was found that H₂ would accept an electron when adsorbed on the TiO₂ surface, while H₂ would donate an electron when adsorbed on the B-TiO₂ surface. Correspondingly, H₂ could not be oxidized over TiO₂, but could be oxidized over B-TiO₂. This indicated that the oxidation of H₂ was dependent on the electron-transfer behavior between H₂ and the surface of TiO₂ or B-TiO₂. Based on the relevant characterization results, it was proposed that H₂ could accept an electron from TiO₂ due to the higher Fermi level of TiO₂, while H₂ could donate an electron to B-TiO₂ due to the lower Fermi level of B-TiO₂ induced by doping B. This indicated that the electron-transfer behavior between H₂ and TiO₂ could be changed by adjusting the Fermi level of TiO₂, while the electron-transfer behavior would further affect the photocatalytic activity of oxidizing H₂. This result shows that the doable H₂ photocatalytic oxidation in thermodynamics can be controlled by a kinetics factor (H₂ losing-an-electron behavior). This work can be applied to provide an understanding of the photocatalytic oxidation behavior of other reactants over semiconductor materials.

---

1. Introduction

Heterogeneous catalytic reactions are always accompanied by the adsorption of reactants, wherein the adsorption of reactants sometimes plays a decisive role in the subsequent reaction process.¹ Therefore, studying the adsorption behavior of reactants is very essential to explore the heterogeneous catalytic reaction process. As the reaction gas is adsorbed on the surface of semiconductor catalysts (e.g., TiO₂, ZnO), a corresponding electron-transfer behavior will be apparent between the semiconductor and adsorbate, which can be explained by the boundary layer theory.² When a gas is adsorbed on the surface of an n-type semiconductor, if the adsorption energy level of the gas is higher than the Fermi level of the semiconductor, electrons will transfer from the adsorbed gas to the conduction band of the semiconductor, which makes the conductivity of the semiconductor increase. In contrast, if the adsorption energy level of the gas is lower than the Fermi level of the semiconductor, electrons will transfer from the conduction band of the semiconductor to the adsorbed gas, resulting in a decrease of the conductivity of the semiconductor. Gas-sensor detection of the target gas is mainly through testing the change in the sensor’ conductivity when the gas is adsorbed on the sensor,^3,4 so we can use such a gas sensor to characterize the conductivity change value when the semiconductor adsorbs the gas, which can reflect the chemisorption strength of the adsorbed gas.⁵ That is to say, it is beneficial to understand the chemisorption behavior (mainly electron-transfer behavior) of the adsorbed gas on a semiconductor surface by studying the gas-sensing property of the semiconductor.

Metal–oxide semiconductors, such as TiO₂, ZnO, SnO₂, and WO₃, usually exhibit good gas sensitivity.^6–11 Since semiconductors exhibit the characteristics of light absorption and photoconductivity, it is a novel idea to improve their gas sensitivity by light excitation. Kumar et al.¹² found that the introduction of UV light could remarkably enhance the H₂-response sensitivity of a nanostructured Au/ZnO thin film. Wang et al.¹³ reported that UV irradiation also could improve the gas response of TiO₂ at room temperature. These results show that illumination can improve the semiconductors’ gas-sensing performance. Considering that TiO₂ semiconductors usually exhibit photocatalytic behavior under UV irradiation, may be such photo-assisted gas sensitivity will also play a role in the photocatalytic performance.

In recent years, our group has made some progress in exploring the relation between the photo-assisted gas sensitivity and photocatalytic activity of catalysts. The study of the photocatalytic oxidation of CO over TiO₂ under H₂-rich conditions showed that only CO could be oxidized under UV irradiation at room temperature, while H₂ could not be oxidized.^14,15 However, according to the convention of semiconductor photocatalysis, the photocatalytic oxidation reaction that is feasible in thermodynamics can occur if the reduction potential of the reductant is higher than the valence band of the semiconductor.¹⁶ As the reduction potential of H⁺/H₂ is higher than the valence band of TiO₂,¹⁷ H₂ should undergo photocatalytic oxidation over TiO₂. To explain this unconventional phenomenon, We explored the electron-transfer behaviors of CO and H₂ over TiO₂ and their oxidation over TiO₂ under UV irradiation at room temperature, respectively.¹⁸ The results showed that CO could be oxidized because it would donate electrons to TiO₂ when CO was adsorbed on the surface of TiO₂, whereas H₂ could not be oxidized because it would accept electrons from TiO₂ when H₂ was adsorbed on the surface of TiO₂. This result indicated that the photocatalytic oxidation of a reactant over TiO₂ is indeed related to the electron-transfer behavior between the reactant and TiO₂. Furthermore, we also found that adjusting the Fermi level of TiO₂ was likely to change the electron-transfer behavior of the adsorbed H₂ on TiO₂, and then would change the photocatalytic oxidation behavior of H₂.¹⁹ As compared to the above TiO₂ sample, as-prepared TiO₂ (in situ) with surface defects prepared by an in situ method could accept electrons from the adsorbed H₂ and could catalytically oxidize H₂ under UV irradiation. This was because the TiO₂ (in situ) with surface defects had a lower Fermi level, resulting in electrons transferring from the adsorbed H₂ to TiO₂. Moreover, introducing impurity defects into TiO₂ by doping N could further lower the Fermi level of TiO₂, resulting in the adsorbed H₂ on N-TiO₂ losing electrons more easily and being easier oxidized. This study not only showed that H₂ photocatalytic oxidation over TiO₂ would be dependent on the electron-transfer behavior between the adsorbed H₂ and TiO₂, but also showed that the introduction of surface defects could adjust the Fermi level of TiO₂ and then change the electron-transfer behavior.

The doping of non-metallic elements (e.g., N, C, B) is an effective method to adjust a semiconductor's band gap structure and Fermi level, and it is also conducive to promoting the catalytic reaction efficiency of semiconductors.^20–22 In non-metal-doped TiO₂ catalysts, B-doping is very efficient in narrowing the band gap and reducing the recombination of charges.²³ Geng et al.²⁴ obtained the geometry and electronic structure of B-doped TiO₂ by theoretical calculation, and found that the photocatalytic performance of B-doped TiO₂ was better than that of simple TiO₂. Moon et al.²⁵ found that TiO₂ modified by boron oxide had better activity in degrading water under ultraviolet light. Grey and co-workers reported that B-modified rutile TiO₂ could convert part of Ti⁴⁺ to Ti³⁺, and as the capture sites of photogenerated electrons, Ti³⁺ could improve the separation efficiency and consequently the photocatalytic activity.²⁶ A large number of studies have shown that B modification can effectively improve the photocatalytic efficiency of catalysts. Considering that B is an electron-deficient element, doping it into TiO₂ may also adjust the Femi level of TiO₂, which would cause a change in the electron-transfer behavior of the adsorbed H₂ and consequently the photocatalytic efficiency.

Therefore, in this study, the non-metallic element B was doped into TiO₂ to prepare B-TiO₂, while naked TiO₂ acted as the contrast sample. The electron-transfer behaviors of H₂ over the two samples were characterized by assessing the photo-assisted gas sensitivity to H₂. After comparing the electron-transfer behaviors and photocatalytic H₂ oxidation over the two samples, respectively, the correlation between the electron-transfer behavior of H₂ and its photocatalytic oxidation over B-TiO₂ was explored.

2. Experimental

2.1 Synthesis of TiO₂ and B-TiO₂

B-TiO₂ catalyst was prepared according to the method in the literature.²⁷ First, 1.24 g boric acid powder was dissolved in 50 mL deionized water and then 6.90 mL tetrabutyl titanate was added dropwise to the H₃BO₃ solution under vigorous stirring at room temperature. After stirring for 12 h, the sample was dried at 100 °C and calcined in air at 500 °C for 1 h. The atomic ratio of B to Ti of the obtained B-TiO₂ catalyst was 1. The simple TiO₂ catalyst was prepared with the same procedure without H₃BO₃.

2.2. Preparation of the gas sensors

TiO₂ and B-TiO₂ gas sensors were fabricated by the drop-coating method reported in our previous work.²⁸ First, 30 mg of the as-prepared TiO₂ was uniformly dispersed in 1 mL ethanediol solvent by sonicating to obtain a TiO₂ suspension. Second, an interdigitated Au electrode (15 mm × 10 mm) was used as the substrate for the gas sensor. Before using it, the electrode was pretreated with acetone, ethanol, and deionized water successively and then dried at 80 °C in an oven. Finally, a 50 μL TiO₂ suspension was dropped onto the clean surface of the Au electrode. The coated electrode was then dried at 100 °C and calcined in air at 500 °C for 1 h (heating rate 2 °C min⁻¹). After cooling down to room temperature, a TiO₂ gas-sensor device was obtained. Using the same process to replace TiO₂ with B-TiO₂, a B-TiO₂ gas-sensor device was obtained.

2.3 Characteristics

The phase and crystal structure of the TiO₂ and B-TiO₂ powder samples were characterized by X-ray diffraction (XRD, D8 Advance, Brucker, Germany) with Cu Kα radiation (2θ: 10°–80°, data analyzed by X’Pert HighScore). The lattice fringes and selected area diffraction of the samples were tested by TEM (Tecnai G2 F20 S-TWIN). The Raman spectra of the samples were analyzed by using a Renishaw 2000 inVia confocal Raman spectrometer, with the excitation wavelength from a 532 nm Ar⁺ laser and a scanning range of 100–1000 nm. UV-vis diffuse reflectance spectra (UV-vis DRS) of the samples were recorded with a UV-vis spectrophotometer (Cary-5000, Varian Co.). X-ray photoelectron spectrometry (XPS) was performed with Mg Kα radiation after monochromatic treatment on the ESCALB mk-ii X-ray photoelectron spectrometer. The Fourier-transform infrared (FT-IR) spectra were recorded on a Thermo Scientific Nicolet iS50 FT-IR spectrometer under the transmission scheme using the KBr pellet technique.

Electrochemical testing was performed on the Metrohm-Autolab AUT302N Electrochemical workstation, and was conducted in a 3-electrode glass cell setup using the TiO₂ film as the working electrode, Ag/AgCl (1 M KCl) as the reference electrode, and Pt wire as the counter electrode. The frequency and amplitude were set to 500, 1000, and 1500 Hz and 5 mV, respectively. Mott–Schottky curves were measured in 0.02 mol L⁻¹ Na₂SO₄ solution and tested in the dark and in UV light, respectively. Photocurrent tests were also carried out in 0.02 mol L⁻¹ Na₂SO₄ solution and electrochemical impedance spectroscopy (EIS) was executed in 0.5 M KCl electrolyte containing 0.01 M K₃[Fe(CN)₆]/K₄[Fe(CN)₆] (1 : 1) under open-circuit potential conditions. The UV light was supplied by a 500 W Xe lamp (also used as the light source in the other experiments in this work; its spectrogram can be seen in Fig. S1 in the ESI†).

Electron paramagnetic resonance (EPR) spectra were registered at a microwave frequency of 9.8 GHz and power of 6.36 Mw by using a Bruker EPR A300 spectrometer and all the EPR spectra were obtained at room temperature.

2.4. Gas-sensing testing

The gas-sensing properties of the films were investigated in a 100 mL chamber, which was made of stainless steel. During the tests, four UV lamps with a wavelength centered at 365 nm (4 W, Philips TL/05) were used as the irradiation source and the total light intensity on the surface of the sensing film was 7.3 mW cm⁻², with high-purity N₂ introduced into the chamber as the background atmosphere and H₂ (balanced against the high-purity N₂) used as the probe gas, and the total flow rate was kept at 250 mL min⁻¹. The film-sensor sample was maintained at 200 °C for 1 h in the high-purity N₂ before testing (to remove the water and the other gas adsorbates), and was then cooled down to room temperature. In addition, the applied voltage was controlled at 8.5 V in the testing process. The response of the film sensor to the gas was described by the variation of its impedance, and the resistance of the film sensor was measured by using a JF02E gas-sensing test system (Kunming Gui Yan Jin Feng Tech. Corp. Ltd.). The relative gas sensitivity (S) of the sensor sample was defined as

S = |R₀ − R_gas|/R₀ (1)

where R_gas and R₀ are the impedances measured in the testing gas and in the background atmosphere, respectively.

2.5. Photocatalytic performances

The catalytic oxidation of H₂ over the catalyst sample was conducted in a 50 mL quartz reaction tube. First, 0.1 g catalyst with a grain size of 0.2–0.3 mm was loaded on to the bottom of the batch. The reaction gas (including 5 mL high-purity O₂, 2.5 mL high-purity H₂, and 42.5 mL He) was injected after first vacuuming the tube (to let the reaction gases spread more easily). The tube was then placed in the dark for 10 min to reach a balance, and then 1 mL gas in the tube was extracted for analyzing the contents. Then, the tube was kept in the dark for 50 min, and the H₂ and O₂ contents were tested. Finally, the reaction tube was illuminated with UV light for 1 h, and the reactants contents were tested again. The reacted gas contents were analyzed using an offline gas chromatograph system equipped with a thermal conductivity detector and a flame ionization detector (Agilent 7890A, Porapark R). Here, since the formed H₂O by the H₂ oxidation product may be adsorbed at the sample surface, the catalytic activity was evaluated by the decrease in the H₂ content.

3. Results and discussion

3.1. Characteristics of TiO₂ and B-TiO₂ samples

Fig. 1 shows the XRD patterns of the TiO₂ and B-TiO₂ samples. X’Pert HighScore software was used to analyze the XRD patterns. The diffraction peaks of TiO₂ and B-TiO₂ were mainly located at 2θ = 25.31°, 37.79°, 48.04°, 53.89°, 55.07°, 62.67°, 68.76°, 70.30°, and 75.05°, which corresponded to the (101), (004), (200), (105), (211), (204), (116), (220), and (215) crystal planes of anatase TiO₂. For TiO₂, a characteristic diffraction peak appeared at 30.81°, which was the (121) crystal plane of brookite TiO₂. Regarding B-TiO₂, it showed two characteristic diffraction peaks of B₂O₃ at 14.56° and 27.77° (310) while the diffraction peak of brookite was reduced. These results demonstrated that the crystal phase of TiO₂ was consistent with anatase and a small amount of brookite. After doping B, the crystal phase of B₂O₃ appeared and the main crystal phase of TiO₂ did not change significantly.

Fig. 1 XRD patterns of the TiO₂ (a) and B-TiO₂ (b) samples.

The TEM and SEAD images of B-TiO₂ in Fig. 2 showed that B-TiO₂ grain was mainly divided between 4–15 nm by particle-size analysis. The lattice fringes corresponding to each crystal plane were obtained according to the results from the XRD analysis. The (101) crystal plane and (310) crystal plane of B₂O₃ are shown in the figure, and the main crystal plane was analyzed in the SEAD image.

Fig. 2 TEM images of B-TiO₂ (a–c) and the SEAD image of B-TiO₂ (d).

The Raman spectra of TiO₂ and B-TiO₂ are shown in Fig. 3. TiO₂ and B-TiO₂ exhibited strong adsorption peaks at 144, 399, 519, and 639 cm⁻¹. These peaks could be assigned to the characteristic vibration peaks of the different positions of anatase TiO₂, which were Eg, B1g, B1g, and Eg, respectively.²⁹ A weak characteristic vibration peak of anatase at 197 cm⁻¹ (Eg) and a weak peak of brookite at 368 cm⁻¹ (B2g) were also exhibited, indicating that the sample still contained a small amount of brookite after B modification, which was consistent with the XRD results. For simple TiO₂, the vibration peak of brookite appeared at 244 cm⁻¹ (A1g) and 325 cm⁻¹ (B2g). For B-TiO₂, the peak at 639 cm⁻¹ was shifted to 643 cm⁻¹, while the characteristic vibration peak that appeared at 882 cm⁻¹ belonged to B₂O₃,³⁰ indicating that after B-doping, part of the B entered the lattice as ions, and another part remained on the TiO₂ surface as B₂O₃.

Fig. 3 Raman spectra of the TiO₂ (a) and B-TiO₂ (b) samples.

Fig. 4 shows the UV-vis diffuse reflectance spectra of the TiO₂ and B-TiO₂ samples. As can be seen, both TiO₂ and B-TiO₂ were white powders, but the UV adsorption of B-TiO₂ was stronger than that of TiO₂ at 300–370 nm. It could be seen from Fig. 4 that B-doping enhanced the absorption intensity of ultraviolet light, but the absorption band edge did not shift significantly. The values in the absorption band edges of TiO₂ and B-TiO₂ were 379 and 380 nm, while the corresponding values of the band gap widths of the two samples were calculated to be 3.27 and 3.26 eV, respectively.

Fig. 4 UV-vis diffuse reflectance spectra (DRS) of the TiO₂ (a) and B-TiO₂ (b) samples.

Fig. 5 shows the EPR spectra of the TiO₂ and B-TiO₂ samples at room temperature in the dark (the EPR signal intensity of each sample was not enhanced when exposed to UV light, so the spectra are not shown here). As can be seen, the EPR signal peak intensity of the material was weakened after B-doping. Here, the peak at g = 2.004 could be assigned to the paramagnetic oxygen species O⁻ or O²⁻, while the peak at g = 1.989 could be assigned to the surface defects (Ti³⁺) of anatase.³¹ Due to B being an electron-deficient element, the 2s orbit of B³⁺ would overlap with the O²⁻ energy band in TiO₂ after B-doping, causing the density of holes to increase, which could capture electrons, so the signal peak of EPR was weakened.

Fig. 5 EPR spectra of TiO₂ (a) and B-TiO₂ (b) at room temperature in the dark.

3.2. Surface structures of TiO₂ and B-TiO₂ characterized by XPS and FT-IR

XPS analyses were conducted to elucidate the surface compositions and chemical states of the TiO₂ and B-TiO₂ samples. As shown in Fig. 6-(1), the two peaks at 529.79 and 531.78 eV in the O1s spectra of TiO₂ could be assigned to the lattice oxygen (Ti–O) and the oxygen of the surface hydroxyl (Ti–O–H), respectively. For the O1s spectra of B-TiO₂, the peak at 530.45 eV was assigned to the lattice oxygen, while the peak at 533.05 eV was assigned to the oxygen in the B–O bond and surface hydroxyl.^32,33 Compared with TiO₂, after B-doping, the lattice oxygen exhibited an increased electron-binding energy (from 529.79 to 530.45 eV) but a decreased proportion, while the surface hydroxyl exhibited an increased electron-binding energy (from 531.78 to 533.05 eV) and an increased proportion, indicating that the presence of the B–O bond could cause some lattice oxygen to change to surface hydroxyl. Here, boron doped into the TiO₂ lattice as B³⁺ ions to form Ti–O–B bonds, and as B is an electron-deficient element, doping B could enhance the Lewis acidity on the B-TiO₂ surface, which would lead to an increase in the surface hydroxyl groups, so that the amount of adsorbed species would increase.^34–36 For the Ti 2p spectra shown in Fig. 6-(2), the peaks at 458.65 and 464.32 eV, respectively, were related to Ti 2p_3/2 and Ti 2p_1/2 of TiO₂, while the corresponding peaks of B-TiO₂ were shifted to 459.16 and 464.84 eV, respectively, indicating that doping B reduced the electron density of Ti 2p. For the B 1s spectra in Fig. 6-(3), the peaks at 192.84 and 193.78 eV could be assigned to the Ti–O–B bond and the B–O–B bond, respectively.³⁷

Fig. 6 XPS high-resolution spectra of the TiO₂ and B-TiO₂ samples: (1) O1s, (2) Ti2p, (3) B1s.

The FT-IR spectra of TiO₂ and B-TiO₂ were obtained to further explain the existing forms of the various bonds in B-TiO₂. As shown in Fig. 7, both TiO₂ and B-TiO₂ showed a vibrational band at 1643 cm⁻¹, which may be assigned to the surface-adsorbed water and hydroxyl groups,²⁷ while the band at 1378 cm⁻¹ corresponded to the C–H bending vibrations, and the broad peak at 656 cm⁻¹ corresponded to the Ti–O bond in TiO₂.³⁸ After doping B, the characteristic peaks related to B appeared. The peak at 1431 cm⁻¹ may be attributed to the Ti–O–B bond, whereas the peak at 1334 cm⁻¹ may be ascribed to the stretching vibrations of the B–O bond of BO₃ units,^27,39 and the peaks observed at 1047 and 927 cm⁻¹ may be attributed to the valence vibrations of the B–O bonds in BO₄ units.⁴⁰ These results were consistent with the XPS analysis results.

Fig. 7 FT-IR spectra of the TiO₂ and B-TiO₂ samples: (a) TiO₂, (b) B-TiO₂.

3.3. Mott–Schottky plots of the TiO₂ and B-TiO₂ samples

The conduction band position of an n-type semiconductor is very close to the position of its Fermi level,^41,42 and so the flat-band potential could approximately be regarded as the Fermi level. Here, a series of Mott–Schottky plots of the TiO₂ and B-TiO₂ samples were performed to further investigate the influence of B-doping on the Fermi level of TiO₂.

Fig. 8 displays the Mott–Schottky plots of the TiO₂ and B-TiO₂ samples in the dark or under UV irradiation, respectively. The flat-band potential (corresponding to E_F) with respect to the Ag/AgCl (1M KCl) electrode could be measured at different frequencies (500, 1000, and 1500 Hz) of the Mott–Schottky plots. In the dark condition, the flat-band potentials of TiO₂ and B-TiO₂ were −1.18 and −0.88 V, respectively. Under UV irradiation, the flat-band potentials of TiO₂ and B-TiO₂ were −1.31 and −0.92 V, respectively. This phenomenon indicated that doping B could reduce the Fermi level of TiO₂. However, the Fermi level of B-TiO₂ only had a slight change under UV irradiation, which was due to the electron accumulation in the semiconductor determining the position of the Fermi level, and as more electrons accumulated, the more negative the Fermi level became.^43,44 After doping an electron-deficient element, the density of holes increased, which could capture the photogenerated electrons; this phenomenon could also be observed in the transient photocurrent response results and the electrochemical impedance spectra (see Fig. S2 and S3 in the ESI†). The positions of the conduction band (vs. RHE) of TiO₂ and B-TiO₂ were −0.53 and −0.23 V in the dark, and −0.66 and −0.27 V under UV irradiation, respectively (see Table 1).

E_RHE = E_Ag/AgCl + 0.059 pH + E^θ_Ag/AgCl (1 MKCl) (2)

where E_RHE is the potential of the reversible hydrogen electrode, E^θ_Ag/AgCl (1M KCl) = 0.236 V at 25 °C.⁴⁵

Fig. 8 Mott–Schottky plots (the graph is from the point of taking the same number dots of straight lines.): (a)-1 TiO₂ (in dark), (a)-2 TiO₂ (UV); (b)-1 B-TiO₂ (in dark), (b)-2 B-TiO₂ (UV).

Table 1 Fermi level values of the TiO₂ and B-TiO₂ samples in the dark and under UV irradiation from Fig. 8

Samples	Fermi level value in the dark (EF), V		Fermi level value under UV irradiation (EF*), V
	vs. Ag/AgCl	vs. RHE	vs. Ag/AgCl	vs. RHE
TiO2	−1.18	−0.53	−1.31	−0.66
B-TiO2	−0.88	−0.23	−0.92	−0.27

---

3.4. Gas-sensing responses of the sensor samples

Fig. 9 shows the gas-sensing responses of the TiO₂ and B-TiO₂ sensor samples to H₂ under UV light irradiation at room temperature in a N₂ atmosphere. It was found that the impedances of the TiO₂ and B-TiO₂ sensor samples decreased rapidly when the UV light was introduced, which could be attributed to the production of photogenerated electrons. With the recombination of the photogenerated electrons and holes, the impedances picked up gradually. For the TiO₂ sample, when H₂ was just introduced, there was a sudden change in the impedance because the surface state of TiO₂ changed after H₂ adsorption. Afterwards, the impedance presented an upward trend, indicating that the adsorbed H₂ would accept electrons from TiO₂. In the cycle test, it was found that the impedance value tended to reach a stable state with the extension of the H₂ introduction time, indicating that the adsorption of H₂ on the surface of TiO₂ gradually reached saturation. For the B-TiO₂ sample, the impedance decreased continuously after passing through H₂, which indicated an increase in electron density on the B-TiO₂ surface, i.e., the adsorbed H₂ would donate an electron to B-TiO₂. Moreover, as H₂ gas was continuously introduced, the impedance decreased and then slowly rose to reach equilibrium, indicating that H₂ adsorption on the B-TiO₂ surface reached saturation. That is, the B-doping into TiO₂ changed the electron-transfer direction between H₂ and TiO₂. Here, B-TiO₂ was conducive to obtaining an electron from H₂, which may be ascribed to the change in the Fermi level of TiO₂ induced by doping B.

Fig. 9 Gas-sensing responses to H₂ under UV light irradiation at room temperature in N₂ atmosphere over the TiO₂ (a) and B-TiO₂ (b) sensor samples.

The above gas–sensitivity values (S) of the TiO₂ and B-TiO₂ sensor samples (the selected data were from the second cycle experimental data) from Fig. 9 are also shown in Table 2. According to formula (1), the gas-sensitivity value of TiO₂ was calculated as 1.55 and that of B-TiO₂ was 2.91. These results suggested that doping B could not only facilitate H₂ to provide electrons to the B-TiO₂ surface, but also improved the sensitivity of the H₂ gas response.

Table 2 The photo-assisted gas-sensitivity values of TiO₂ and B-TiO₂ to H₂ under UV light at room temperature from the results in Fig. 9

Sensor samples	R 0 (KΩ)	R gas (KΩ)	S	Electron-transfer direction
R 0 was assigned to the last impedance value prior to introducing H2; Rgas was assigned to the lowest value after introducing H2; S was assigned to the gas–sensitive response value.
TiO2	73786.50	114992.97	1.55	From TiO2 to H2
B-TiO2	22308.08	7661.13	2.91	From H2 to B-TiO2

---

3.5. Photocatalytic activities of the TiO₂ and B-TiO₂ samples

The results from the gas-sensitivity tests showed that H₂ could accept electrons when adsorbed on TiO₂, but it would donate electrons when adsorbed on B-TiO₂. According to our previous work,¹⁸ the adsorbed H₂ over B-TiO₂ (H₂ losing electrons) under UV irradiation should be oxidized over B-TiO₂, but that of H₂ (accepting electrons) could be not oxidized over TiO₂. To further prove the relationship between the electron-transfer behavior and photocatalytic activity, the photocatalytic performances of oxidizing H₂ over TiO₂ and B-TiO₂ were tested under UV irradiation at room temperature. Table 3 shows the experimental results for the TiO₂ and B-TiO₂ photocatalytic oxidation of H₂. After reacting for 1 h in the dark, it was found that the concentration of H₂ over both TiO₂ and B-TiO₂ was reduced slightly because of the adsorption of H₂. However, after reacting for a further 1 h under UV irradiation, the concentration of H₂ over TiO₂ only had a tiny decrease while the concentration of H₂ over B-TiO₂ was decreased obviously, indicating that H₂ adsorption on TiO₂ involved only physical adsorption, but involved chemical adsorption on B-TiO₂, and thus further reacted with O₂. Hence, B-TiO₂ could catalytically oxidize H₂ under UV irradiation, but TiO₂ hardly did. These results confirmed that there indeed was a certain relationship between the electron-transfer behavior (photo-assisted gas sensitivity) and photocatalytic activity.

Table 3 Results from catalytically oxidizing H₂ over TiO₂ and B-TiO₂ under UV irradiation at room temperature

Samples	Initial concentration of H2 (ppm)	Concentration of H2 in the dark for 1 h (ppm)	Concentration of H2 under UV light for 1 h (ppm)	Characteristic
a The concentration decrease value of H2 in the dark or under UV light for 1 h (ppm). b O2 contents (not shown here) also decreased, indicating that H2 was oxidized by O2.
TiO2	67670.95	60491.94(−7179.01)^a	56961.28(−3530.66)^a	H2 not oxidized
B-TiO2	58185.24	52830.41(−5354.83)^a	41073.32(−11757.09)^a	H2 oxidized

---

3.6. Proposed process of electron transfer over TiO₂ and B-TiO₂ absorbing H₂

Based on the above series of experimental studies and our previous work,¹⁹ we summarized the correlation between the H₂ sensing response and the oxidation of H₂ over TiO₂ and B-TiO₂ under UV irradiation (see Table 4), and then proposed the process of electron transfer over TiO₂ and B-TiO₂ absorbing H₂, as shown in Fig. 10. For TiO₂ (Fig. 10-(A)), from a thermodynamic point of view, H₂ could be oxidized by TiO₂, but the photocatalytic experiment results showed that H₂ could not be oxidized by TiO₂, which was related to the surface adsorption energy level of H₂ (H₂ (ad)) and the position of the Fermi level of TiO₂ (i.e., E_F). Generally, the H₂ (ad) was higher than the Fermi level of TiO₂, so H₂ could provide electrons to TiO₂. Under UV irradiation, the Fermi level of TiO₂ (i.e., E_F*) rose, even higher than the energy level of H₂ (ad), resulting in H₂ accepting electrons from TiO₂ and then H₂ could not undergo photocatalytic oxidation. For B-TiO₂ (Fig. 10-(B)), doping B introduced an impurity energy level B2p, which effectively reduced the position of the Fermi level of TiO₂ (i.e., E_F). Under UV irradiation, the Fermi level of B-TiO₂ (i.e., E_F*) was lower than the energy level of H₂ (ad), and thus H₂ could donate electrons to B-TiO₂ and could undergo photocatalytic oxidation over B-TiO₂.

Table 4 Correlation between H₂ sensing response and its oxidation over TiO₂ and B-TiO₂ under UV irradiation

Samples	Fermi level value (V) under UV irradiation	Electron-transfer direction	H2 oxidation under UV irradiation
TiO2	−1.31	From TiO2 to H2	No
B-TiO2	−0.92	From H2 to B-TiO2	Yes

---

Fig. 10 The proposed processes of electron transfer over (A) TiO₂ and (B) B-TiO₂ adsorbing H₂. Here, electrons transfer from TiO₂ to the adsorbed H₂ (H₂ (ad)) due to the higher E_F*, while electrons transfer from H₂ (ad) to B-TiO₂ due to the lower E_F*.

The above results further showed that the oxidation of H₂ would be actually dependent on the electron-transfer behaviors of the adsorbed H₂ induced by the Fermi level of TiO₂ and B-TiO₂, suggesting that the photocatalytic oxidation of H₂ over a semiconductor was controlled by both the thermodynamics and kinetics. Here, the behavior of H2 of losing an electron, as a kinetics factor, would play a key role in its photocatalytic oxidation. Moreover, doping some elements into TiO₂ could adjust the Fermi level of TiO₂-based semiconductors. This viewpoint may provide a greater applicable understanding for studying the photocatalytic oxidation behaviors of other reactants over other semiconductor materials.

4. Conclusions

In this study, TiO₂ was modified by doping the electron-deficient B element to obtain a B-TiO₂ sample. After investigating the photo-assisted gas sensitivity (electron-transfer behavior), the photocatalytic activity, and performing other characterizations of B-TiO₂, the following conclusions could be obtained:

(I) Compared with the simple TiO₂, the B-doped TiO₂ formed a Ti-O-B bond, which reduced the Fermi level of B-TiO₂.

(II) The photo-assisted gas sensitivity to H₂ in N₂ atmosphere indicated that H₂ would act as an electron acceptor when adsorbed on the TiO₂ surface, while it acted as an electron donor when adsorbed on the B-TiO₂ surface.

(III) H₂ could not be oxidized over TiO₂ by O₂, but it could be oxidized over B-TiO₂, indicating a losing-an-electron behavior of the adsorbed H₂ to drive H₂ oxidation over B-TiO₂.

(IV) The change in the Fermi level altered the electron-transfer behavior of H₂ over TiO₂, which further affected the photocatalytic performances.

Overall, H₂ photocatalytic oxidation over a semiconductor would be controlled not only by the thermodynamics but also by the kinetics, and the losing-an-electron behavior of H₂, as a kinetics factor, could play a key role in its photocatalytic oxidation. This conclusion may be applicable for other semiconductors.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was financially supported by the National Key R & D Program of China (No. 2018YFE0208500) and the National Natural Science Foundation of China (No. 21872030).

References

1. B. Yang, R. Burch, C. Hardacre, G. Headdock and P. Hu, Understanding the optimal adsorption energies for catalyst screening in heterogeneous catalysis, ACS Catal., 2014, 4, 182–186.
2. S. R. Morrison, The Chemical Physics of Surface, Plenum Press, New York, 1977.
3. A. Mirzaei, S. S. Kim and H. W. Kim, Resistance-based H₂S gas sensors using metal oxide nanostructures: a review of recent advances, J. Hazard. Mater., 2018, 357, 314–331.
4. E. Lee, A. VahidMohammadi, B. C. Prorok, Y. S. Yoon, M. Beidaghi and D. J. Kim, Room temperature gas-sensing of two-dimensional titanium carbide (MXene), ACS Appl. Mater. Interfaces, 2017, 9, 37184–37190.
5. W. Göpel and G. Rocker, Phys. Rev. B: Solid State, 1983, 28, 3427–3438.
6. I. Cho, Y. C. Sim, M. Cho, Y. H. Cho and I. Park, Monolithic micro light-emitting diode/metal oxide nanowire gas sensor with microwatt-level power consumption, ACS Sens., 2020, 5, 563–570.
7. A. V. Raghu, K. K. Karuppanan, J. Nampoothiri and B. Pullithadathil, Wearable, flexible ethanol gas sensor based on TiO₂ nanoparticles-grafted 2D-titanium carbide nanosheets, ACS Appl. Nano Mater., 2019, 2, 1152–1163.
8. R. K. Sonker, S. R. Sabhajeet, S. Singh and B. C. Yadav, Synthesis of ZnO nanopetals and its application as NO₂ gas sensor, Mater. Lett., 2015, 152, 189–191.
9. X. Y. Kou, F. Q. Meng, K. Chen, T. S. Wang, P. Sun, F. M. Liu, X. Yan, Y. F. Sun, F. M. Liu, K. Shimanoe and G. Lu, High-performance acetone gas sensor based on Ru-doped SnO₂ nanofibers, Sens. Actuators, B, 2020, 320, 128292–128299.
10. G. Jung, Y. Jeong, Y. Hong, M. Wu, S. Hong, W. Shin, J. Park, D. Jang and J. H. Lee, SO₂ gas sensing characteristics of FET- and resistor-type gas sensors having WO₃ as sensing material, Solid State Electronics, 2019, 165, 107747–107753.
11. A. Nikfarjam, S. Hosseini and N. Salehifar, Fabrication of a highly sensitive single aligned TiO₂ and gold nanoparticle embedded TiO₂ nano-fiber gas sensor, ACS Appl. Mater. Interfaces, 2017, 9, 15662–15671.
12. G. S. Kumar, X. J. Li, Y. Du, Y. F. Geng and X. M. Hong, UV-light enhanced high sensitive hydrogen (H₂) sensor based on spherical Au nanoparticles on ZnO nano-structured thin films, J. Alloys Compd., 2019, 798, 467–477.
13. P. Wang, T. F. Xie, L. Peng, H. Y. Li, T. S. Wu, S. Pang and D. G. Wang, Water-assisted synthesis of anatase TiO₂ nanocrystals:  Mechanism and sensing properties to oxygen at room temperature, J. Phys. Chem. C, 2008, 112, 6648–6652.
14. W. X. Dai, X. Chen, X. X. Wang, P. Liu, D. Z. Li, G. S. Li and X. Z. Fu, CO preferential oxidation promoted by UV irradiation in the presence of H₂ over Au/TiO₂, Phys. Chem. Chem. Phys., 2008, 10, 3256–3262.
15. W. X. Dai, X. Chen, X. P. Zheng, Z. X. Ding, X. X. Wang, P. Liu and X. Z. Fu, Photocatalytic oxidation of CO on TiO₂: Chemisorption of O₂, CO, and H₂, ChemPhysChem, 2009, 10, 411–419.
16. S. Y. Chen and L. W. Wang, Thermodynamic oxidation and reduction potentials of photocatalytic semiconductors in aqueous solution, Chem. Mater., 2012, 24, 3659–3666.
17. J. G. Yu, Y. Hai and B. Cheng, Enhanced photocatalytic H₂-production activity of TiO₂ by Ni(OH)₂ cluster modification, J. Phys. Chem. C, 2011, 115, 4953–4958.
18. X. Y. Peng, Z. J. He, K. Yang, X. Chen, X. X. Wang, W. X. Dai and X. Z. Fu, Correlation between donating or accepting electron behavior of the adsorbed CO or H₂ and its oxidation over TiO₂ under ultraviolet light irradiation, Appl. Surf. Sci., 2016, 360, 698–706.
19. K. Wang, T. Peng, Z. M. Wang, H. Wang, X. Chen, W. X. Dai and X. Z. Fu, Correlation between the H₂ response and its oxidation over TiO₂ and N doped TiO₂ under UV irradiation induced by Fermi level, Appl. Catal., B, 2019, 250, 89–98.
20. W. Qian, P. A. Greaney, S. Fowler, S. K. Chiu, A. M. Goforth and J. Jiao, Low-temperature nitrogen doping in ammonia solution for production of N-doped TiO₂-hybridized graphene as a highly efficient photocatalyst for water treatment, ACS Sustainable Chem. Eng., 2014, 2, 1802–1810.
21. R. Borah, R. Borah and S. Paria, Carbon-doped mesoporous anatase TiO₂ multi-tubes nanostructures for highly improved visible light photocatalytic activity, Inorg. Chem., 2017, 56, 10107–10116.
22. N. D. Feng, A. M. Zheng, Q. Wang, P. P. Ren, X. Z. Gao, S. B. Liu, Z. R. Shen, T. H. Chen and F. Deng, Boron environments in B-doped and (B, N)-codoped TiO₂ photocatalysts: A combined solid-State NMR and theoretical calculation Study, J. Phys. Chem. C, 2011, 115, 2709–2719.
23. N. Patel, A. Dashora, R. Jaiswal, R. Fernandes, M. Yadav, D. C. Kothari, B. L. Ahuja and A. Miotello, Experimental and theoretical investigations on the activity and stability of substitutional and interstitial boron in TiO₂ photocatalyst, J. Phys. Chem. C, 2015, 119, 18581–18590.
24. H. Geng, S. W. Yin, X. Yang, Z. G. Shuai and B. G. Liu, Geometric and electronic structures of the boron-doped photocatalyst TiO₂, J. Phys.: Condens. Matter., 2006, 18, 87–96.
25. S. C. Moon, H. Mametsuka, S. Tabata and E. Suzuki, Photocatalytic production of hydrogen fromwater using TiO₂ and B/TiO₂, Catal. Today, 2000, 58, 125–132.
26. I. E. Grey, C. Li, C. M. MacRae and L. A. Bursill, Boron incorporation into rutile. Phase equilibria and structure considerations, J. Solid State Chem., 1996, 127, 240–247.
27. D. M. Chen, D. Yang, Q. Wang and Z. Y. Jiang, Effects of boron doping on photocatalytic activity and microstructure of titanium dioxide nanoparticles, Ind. Eng. Chem. Res., 2006, 45, 4110–4116.
28. X. Y. Peng, Z. M. Wang, P. Huang, X. Chen, X. Z. Fu and W. X. Dai, Comparative study of two different TiO₂ film sensors on response to H₂ under UV Light and room temperature, Sensors, 2016, 16, 1249.
29. C. P. Kumar, N. O. Gopal, T. C. Wang, M. S. Wong and S. C. Ke, EPR investigation of TiO₂ nanoparticles with temperature dependent properties, J. Phys. Chem. B, 2006, 110, 5223–5229.
30. Z. Wang, Q. F. Shu and K. C. Chou, Study on structure characteristics of B₂O₃ and TiO₂-bearing F-free mold flux by Raman spectroscopy, High Temp. Mater. Processes, 2013, 32, 265–273.
31. K. Y. Jung, S. B. Park and S. K. Ihm, Local structure and photocatalytic activity of B₂O₃-SiO₂/TiO₂ ternary mixed oxides prepared by sol-gel method, Appl. Catal., B, 2004, 51, 239–245.
32. A. Zaleska, J. W. Sobczak, E. Grabowska and J. Hupka, Preparation and photocatalytic activity of boron-modified TiO₂ under UV and visible light, Appl. Catal., B, 2008, 78, 92–100.
33. H. Yang, Y. Z. Wang and X. X. Xue, Influences of glycerol as an efficient doping agent on crystal structure and antibacterial activity of B-TiO₂ nano-materials, Colloids Surf., B, 2014, 122, 701–708.
34. A. Ansón-Casaos, M. J. Sampaio, C. Jarauta-Córdoba, M. T. Martínez, C. G. Silva, J. L. Faria and A. M. T. Silva, Evaluation of sol-gel TiO₂ photocatalysts modified with carbon or boron compounds and crystallized in nitrogen or air atmospheres, Chem. Eng. J., 2015, 277, 11–20.
35. Y. Wang, K. Jia, Q. Pan, Y. D. Xu, Q. Liu, G. W. Cui, X. D. Guo and X. P. Sun, Boron-doped TiO₂ for efficient electrocatalytic N₂ fixation to NH₃ at ambient conditions, ACS Sustainable Chem. Eng., 2019, 7, 117–122.
36. M. Quesada-González, N. D. Boscher, C. J. Carmalt and I. P. Parkin, Interstitial boron-doped TiO₂ thin films: the significant effect of boron on TiO₂ coatings grown by atmospheric pressure chemical vapor deposition, ACS Appl. Mater. Interfaces, 2016, 8, 25024–25029.
37. Y. N. Huo, X. Y. Zhang, Y. Jin, J. Zhu and H. X. Li, Highly active La₂O₃/Ti_1−xB_xO₂ visible light photocatalysts prepared under supercritical conditions, Appl. Catal., B, 2008, 83, 78–84.
38. Y. Zhang, A. Barber, J. Maxted, C. Lowe, R. Smith and T. Z. Li, The depth profiling of TiO₂ pigmented coil coatings using step scan phase modulation photoacoustic FTIR, Prog. Org. Coat., 2013, 76, 131–136.
39. N. D. Feng, A. M. Zheng, Q. Wang, P. P. Ren, X. Z. Gao, S. B. Liu, Z. R. Shen, T. H. Chen and F. Deng, Boron environments in B-doped and (B, N)-codoped TiO₂ photocatalysts: A combined solid-state NMR and theoretical calculation study, J. Phys. Chem. C, 2011, 115, 2709–2719.
40. S. Kabi and A. Ghosh, Correlation of structure and electrical conductivity of CdI₂ doped silver borophosphate glass and nanocomposite, J. Phys. Chem. C, 2011, 115, 9760–9766.
41. J. Yu, Y. L. Yang, R. Q. Fan, D. Q. Liu, L. G. Wei, S. Chen, L. Li, B. Yang and W. W. Cao, Enhanced near-infrared to visible upconversion nanoparticles of Ho³⁺-Yb³⁺-F^– tri-doped TiO₂ and its application in dye-sensitized solar cells with 37% improvement in power conversion efficiency, Inorg. Chem., 2014, 53, 8045–8053.
42. S. Burnside, J. E. Moser, K. Brooks, M. Grätzel and D. Cahen, Nanocrystalline mesoporous strontium titanate as photoelectrode material for photosensitized solar devices: increasing photovoltage through flatband potential engineering, J. Phys. Chem. B, 1999, 103, 9328–9332.
43. M. Jakob, H. Levanon and P. V. Kamat, Charge distribution between UV-irradiated TiO₂ and gold nanoparticles: determination of shift in the Fermi Level, Nano Lett., 2003, 3, 353–358.
44. V. Subramanian, E. E. Wolf and P. V. Kamat, Catalysis with TiO₂/Gold nanocomposites. Effect of metal particle size on the Fermi Level equilibration, J. Am. Chem. Soc., 2004, 126, 4943–4950.
45. A. J. E. Rettie, K. C. Klavetter, J. F. Lin, A. Dolocan, H. Celio, A. Ishiekwene, H. L. Bolton, K. N. Pearson, N. T. Hahn and C. B. Mullins, Improved visible light harvesting of WO₃ by incorporation of sulfuror iodine: a tale of two impurities, Chem. Mater., 2014, 26, 1670–1677.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/d0cp04039h

---