The precursor gel preparation process was similar to that described in our preliminary study ^[20], except that yellow and green precursor gels were obtained by controlling the amount of H₂O₂ and the oxidation time. The precursor gels were transferred to dry Teflon-lined autoclaves, and then heated to 160 °C for different times. After natural cooling to room temperature, the gray-blue or light-blue mixture was collected and washed with distilled water and ethanol three times, and then dried in an oven at 60 °C for 3 h. The products obtained from the yellow gel are labeled as Y-t, while the products obtained from green gel are labeled as G-t, where t indicates the hydrothermal treatment time (hours). For comparison, pure TiO₂ (denoted p-TiO₂) was obtained by hydrothermal treatment at 160 °C for 24 h using TiCl₄ as the Ti source.

The phases of the final products were identified using XRD (Rigaku D/max-2500VPC) with graphite monochromated Cu K_a (λ = 0.15418 nm) radiation at a scanning rate of 0.02°/s. A Hitachi H-800 model TEM and JEOL JEM-2100 model HRTEM were used to determine the size and shape of the particles. TEM sample specimens were prepared by briefly ultrasonicating the sample powders in ethanol, followed by placing a drop of suspension onto a carbon-coated copper grid. The grids were dried before imaging. XPS measurements were carried out on an X-ray photoelectron spectrometer (ESCA 3000) using Al K_a (1486.6 eV) X-rays as the excitation source. C 1s (284.6 eV) was chosen as the reference. ESR measurements were carried out using a FA-200 spectrometer/X-bond. UV-vis DRS were recorded on a Shimadzu UV-2550 UV-vis spectrophotometer in the range 200-800 nm at room temperature, with a BaSO₄ standard used as the reference.

The photocatalytic activities of the as-prepared samples were evaluated using the photodecomposition of MB solution under visible light irradiation at room temperature. The visible light was provided by a 300 W Xe arc lamp (PLS-SXE300, Beijing Trusttech Co., Ltd.) equipped with an ultraviolet cutoff filter (UVCUT 400, Beijing Trusttech Co., Ltd.). The vertical distance between the surface of the sample suspension and the light source was set at 30 cm. In a typical photocatalytic experiment, 0.06 g of sample was dispersed into 100 mL MB solution (2 x 10⁻⁵ mol/L) under constant stirring. The suspension was stirred for 30 min in the dark prior to irradiation. Aliquots of about 3 mL were removed from the suspension at 10 min intervals during visible light illumination and centrifuged to remove the remaining particles. The absorbance of the centrifuged solutions was measured at 500-800 nm as a function of irradiation time by UV-vis spectroscopy (UV-7200, Unico, Shanghai).

XRD was used to investigate the structural changes and phase purity of the samples obtained from the different precursor gels after different hydrothermal treatment times at 160 °C, as shown in Fig. 1. When the hydrothermal treatment time was 20 h, the content of TiH₂ in the products obtained from the yellow precursor gel was lower than that of the products obtained from the green precursor gel. Except for two weak diffraction peaks of TiH₂ at 2θ = 34.9°, 40.6° (JCPDS 65-0934), all the observed diffraction peaks corresponded to those of anatase TiO₂ (JCPDS 21-1272). However, the content of TiH₂ was very high in the products obtained using the green gel as the precursor. The yellow gel precursor transformed to pure anatase TiO₂ after 24 h hydrothermal treatment, compared with 27 h for the the yellow green gel ^[22], while the green gel still contained a small amount of TiH₂ after 27 h, and only became pure anatase phase TiO₂ after the treatment time was extended to 32 h.

Fig. 1.XRD patterns of the samples obtained from different gels after hydrothermal treatment at 160 °C.

In the presence of H₂O₂, Ti⁴⁺ is known to exhibit different colors at different pH in a system; orange in acidic conditions, yellow at pH ~8, and colorless in strong alkaline conditions. In the present gel preparation process the pH value remained neutral, so the different states of gel obtained were caused by the presence of different proportions of Ti-oxygen compounds produced by oxidation of H₂O₂ on TiH₂. The lower content of TiH₂ in the yellow gel indicated that the TiH₂ was oxidized to a high degree in the yellow gel, making it easier to completely transform to TiO₂ during the hydrothermal treatment. In the green gel, the degree of oxidation was low and the content of TiH₂ was high, so it took a long time for it to transform into TiO₂ during the hydrothermal treatment. In addition, no matter what kind of gel was used as the precursor, the TiO₂ diffraction peaks became increasingly sharp with increasing hydrothermal treatment time, indicating a corresponding better crystallinity and larger crystallite size of the obtained TiO₂. Therefore, when using TiH₂ as the Ti source and H₂O₂ as oxidant to prepare Ti³⁺ self-doped TiO₂, controlling the oxidation time and degree of oxidation to obtain the yellow gel precursor shortens the time required for the subsequent hydrothermal treatment.

Fig. 2 shows TEM and HRTEM images of the samples obtained after hydrothermal treatment at 160 °C for different times. Sample Y-20 was composed of small particles (Fig. 2(a)). Because the content of TiH₂ in this sample was very small, the lattice structure of TiH₂ could not be observed in the HRTEM image. When the hydrothermal treatment time was prolonged to 24 h, the morphology of the sample become more regular, and the dispersity was improved (Fig. 2(b)). The XRD results indicated that this sample was pure anatase phase TiO₂ (Fig. 1). Under the same conditions, the sample obtained using the green gel as the precursor was irregular in shape after hydrothermal treatment for 20 h, with a wide particle size distribution (Fig. 2(c)). HRTEM showed that TiH₂ was still present in this sample; the spacing d = 0.44 nm in Fig. 2(d) is TiH₂ ^[23]. The TEM and HRTEM results further indicated the products formed from the green gel precursor after shorter hydrothermal treatment times were composites of TiH₂ and TiO₂. The morphology of the samples became more regular with increasing hydrothermal treatment time. When the hydrothermal treatment time was 32 h, the sample obtained from the green gel precursor had a well-defined morphology and good dispersity (Fig. 2(e)). The HRTEM results show that this sample was pure anatase phase TiO₂ (Fig. 2(f)), in agreement with the XRD results. The particle sizes of the samples obtained from the yellow gel were the smallest, while those of the samples obtained from the green gel were the largest. In heterogeneous photocatalytic processes, small particles have good dispersity and strong adsorption ability, and can provide a greater number of active sites for the adsorption of pollutant molecules, which is conducive to photocatalytic reactions ^[24].

Fig. 2.TEM and HRTEM images of samples obtained from different gels after hydrothermal treatment at 160 °C. (a) Y-20; (b) Y-24; (c,d) G-20; (e,f) G-32.

XPS and EPR were used to determine the composition and the chemical state of Ti in the samples. Fig. 3(a) and (b) present high resolution Ti 2p and O1s XPS results for samples Y-20, Y-24, and G-32. Fig. 3(c) and (d) present the Ti 2p and O 1s fitting results of samples G-32 and Y-24, respectively. Pure TiO₂ had typical Ti 2p binding energies of ~464.5 and ~459.0 eV ^[25], while those of Y-20 and Y-24 were shifted to lower energy owing to the presence of Ti³⁺ in the samples ^[17]. This shift to lower energy was decreased with increasing hydrothermal treatment time, indicating that the concentration of Ti³⁺ in the samples decreased with increasing hydrothermal treatment time. The Ti 2p peaks were well de-convoluted into four peaks as Ti³⁺ 2p_3/2 at 457.7 eV, Ti⁴⁺ 2p_3/2 at 458.7 eV, Ti³⁺ 2p_1/2 at 463.0 eV, and Ti⁴⁺ 2p_1/2 at 464.5 eV ^{[25, 26]} for the sample obtained after treatment of the green gel precursor at 160 °C for 32 h (Fig. 3(c)), but the strength of the Ti³⁺ binding energy peak was weak, indicating that the content of Ti³⁺ was low.

Fig. 3.High-resolution XPS spectra of Ti 2p (a) and O 1s (b) for different samples. Fitting results of Ti 2p (c) for sample G-32 and O 1s (d) for sample Y-24.

The binding energy of O in the Ti-O band is located at 529.5-530.5 eV ^[27]. Fig. 3(d) shows the high resolution O 1s XPS spectra of the samples. This binding energy appeared at 530 eV for all three samples, indicating that O existed mainly in the form of Ti-O, but the peaks were asymmetrical. Three peaks were obtained after fitting the peak of sample Y-24 (Fig. 3(d)). The binding energy at 530.1 eV corresponds to the Ti-O band, while the peak at 531.3 eV corresponds to the characteristic peak of surface oxygen defects (Ov) ^[28]. The weak peak at 532.3 eV corresponds to the characteristic peak of the -OH band ^{[29, 30]}, indicating that -OH existed on the surface of the particles.

ESR was used to study the uncoupled electrons in the samples and their interactions with the surrounding atoms in the material. Fig. 4 shows the ESR results of samples Y-20, Y-24, G-27, G-32, and pure TiO₂. Compared with pure TiO₂, a signal peak at about g = 1.94 appeared for all of the samples after hydrothermal treatment, corresponding to the Ti³⁺ signal peak. The intensity of this signal peak decreased with increasing hydrothermal treatment time, indicating that the Ti³⁺ concentration in the sample decreased ^[31], which is consistent with the XPS results. An electron trapped on an O_V would produce a signal at g = 2.003; this signal peak indicates the presence of surface Ti^{3+ [32]}, and shows that Ti³⁺ not only existed in the lattice, but also in the surface of the samples. The decreasing content of Ti³⁺ and Ov in the samples observed with increasing hydrothermal treatment time is consistent with the XPS results.

Fig. 4.ESR spectra for different samples.

The presence of Ti³⁺, Ov, and -OH groups on the surface of TiO₂ has important effects on the photoelectric conversion properties of the material. Surface -OH groups are able to capture holes to form oxidizing hydroxyl radicals (•OH), which play an important role in the photocatalytic degradation of organic pollutant molecules ^{[33, 34]}.

Fig. 5 shows the UV-vis DRS spectra of the Ti³⁺ self-doped TiO₂ nanoparticles obtained under different conditions and pure TiO₂. Unlike p-TiO₂, the spectra of the Ti³⁺ self-doped TiO₂ samples exhibited a broad absorption band between 400-800 nm, covering the entire visible range. Samples Y-20 and G-27 had strong absorptions in the visible light region, but weak in the UV region. On the one hand, this was caused by the presence of incompletely reacted TiH₂ in the samples, while on the other hand, because the hydrothermal treatment time was short, Y-20 and G-27 had poor crystallinity and more defects. The UV absorption of the samples was enhanced by prolonging the hydrothermal treatment time for both the yellow and green gel precursors. This is because the crystallinity of the samples continuously improved with treatment time, while the amount of defects decreased. Additionally, samples Y-24 and G-32 had strong absorption both in the visible and UV regions, indicating that these samples had good crystallinity, fewer defects, and more Ti³⁺. Sample Y-24 had a stronger absorption in the visible light range than sample G-32, indicating a higher concentration of Ti³⁺ in sample Y-24 than G-32, consistent with the XPS and ESR results.

Fig. 5.UV-Vis-DRS spectra for different samples.

According to the above results, we believe that the Ti²⁺ on the surface of the TiH₂ particles was gradually oxidized by H₂O₂, and a number of oxygenated Ti compounds and their hydrates were formed at the particle surface. When the hydrothermal treatment was carried out on the precursor gel, the oxygenated Ti compounds and their hydrates were dehydrated. At the same time, the incompletely oxidized TiH₂ and the oxygen compounds on the particle surface participated in ion diffusion-exchange reactions in the interface, resulting in the formation of Ti³⁺ and Ov ^[22]. Because the degree of surface oxidization of the yellow gel was higher than that of the green gel, its content of TiH₂ was lower, so pure anatase phase Ti³⁺ self-doped TiO₂ nanoparticles were obtained after only a short hydrothermal treatment time. Compared with the use of tetrabutyl titanate as the raw material to prepare TiO₂ containing surface Ti^{3+ [35, 36]}, our adoption of a low valent titanium compound as the raw material means that during the initial stage of the reaction, the Ti²⁺ ions located on the particle are converted to TiO₂ contains Ti³⁺ in both the lattice and surface through ion diffusion and exchange reactions, which ensures the stability of the material. Meanwhile, using H₂O₂ as an oxidant can guarantee the purity of the product, and controlling the degree of oxidation of TiH₂ and the subsequent hydrothermal treatment conditions allows the content of Ti³⁺ and Ov in the final products to be controlled to adjust their catalytic performance.

MB is a brightly colored blue cationic thiazine dye often used as a test model pollutant in semiconductor photocatalysis. Fig. 6(a) illustrates the ability of the Ti³⁺ self-doped TiO₂ nanoparticles samples to photodegrade MB solutions in comparison with that of p-TiO₂. Before illumination, all samples exhibited a certain MB absorption, among which the sample obtained from the yellow gel precursor after 20 h hydrothermal treatment had the best adsorbability. This is because the shorter hydrothermal treatment time meant that the sample size was smaller and the specific surface area was larger. The adsorption properties of the samples decreased gradually with increasing hydrothermal treatment time, owing to larger particle sizes and smaller specific surface area, consistent with the TEM results. Sample Y-24 showed the best photocatalytic performance, followed by sample G-32. Although samples Y-20 and G-27 exhibited strong absorption in visible light region, their catalytic performance was not very good compared with that of Y-24 and G-32. First, incompletely reacted TiH₂ in these samples caused their absorption in the visible light. Second, the poor crystallinity of Y-20 and G-27 meant that there were more internal defects in the particles, which acted as recombination centers for electrons and holes, decreasing the photocatalytic performance. Comparatively, the hydrothermal treatment time of sample Y-24 was short, the particle size was small, and the specific surface area was large, making it good for photocatalytic reaction. Moreover, the XPS and ESR results showed that the content of Ti³⁺ and Ov was higher when the hydrothermal treatment time was shorter; high concentrations of Ti³⁺ and Ov enhanced the visible light response, so the photocatalytic activity of sample Y-24 was the best among these samples.

Fig. 6.(a) Photodegradation of MB solutions using p-TiO2 and obtained samples under visible light irradiation; (b) Kinetic study of photocatalytic oxidation of MB monitored using ln(C0/C) versus irradiation time; (c) Recycling test results for samples Y-24 and G-32; (d) Schematic illustration of electron transition caused by visible light irradiation and photocatalytic process of Ti3+ self-doped TiO2.

The photocatalytic activity of the powders can be quantitatively evaluated by comparing their apparent reaction rate constants using the Langmuir-Hinshelwood equation K_app = ln(C₀/C)/t to calculate the photocatalytic reaction rate, where K_app is the first order reaction rate constant. K_app was obtained from the linear time dependences of ln(C/C₀), as shown in Fig. 6(b). Under visible light irradiation, the rate constants for MB degradation followed the sequence Y-24 > G-32 > Y-20 > G-27 > p-TiO₂. Sample Y-24 exhibited a first-order rate constant of K_app = 0.0439 min⁻¹, about 18.3 times greater than that of p-TiO₂ under visible light irradiation.

The stability of a photocatalyst is highly important to its application. The cyclic stability of samples Y-24 and G-32 were examined, and the results showed the maintenance of an excellent photocatalytic activity for degrading MB solution, as shown in Fig. 6(c). After 8 reuses, the photocatalytic degradation efficiency of the samples was over 95%, which illustrated that the samples had good stability and good potential for practical application.

The introduction of Ti³⁺ or O_V in TiO₂ can form a local state at the bottom of the conduction band (CB) in the range of 0.75-1.18 eV, so that the TiO₂ has visible and infrared light absorptions. When the content of Ti³⁺ and Ov reaches a certain concentration, the local state can form a similar doping energy level, thereby reducing the width of the forbidden band of TiO₂ ^[17]. During visible light excitation, electrons can transit from the valence band (VB) of TiO₂ to the local state formed by Ti³⁺ and Ov, and then transit to the CB, so that TiO₂ can be excitated by visible light and even infrared light ^{[10, 20, 36]}. It is reported that defects, especially lattice defects, increase the probability of the recombination of photogenerated electron and holes, thereby reducing thephotocatalytic activity ^[37]. As a kind of defect, Ti³⁺, on the one hand, acting as an electron capture agent, can increase the electrical conductivity of the photocatalyst and accelerate the transfer of electrons and holes ^[38]. On the other hand, the synergistic effect of Ti³⁺ and Ov can prolong the life time of electrons and holes, reducing the recombination rate ^{[11, 28, 39, 40]}. Figure 6(d) shows a schematic diagram of the visible light photocatalytic reaction mechanism of the Ti³⁺ self-doped TiO₂ nanoparticles. Upon excitation with visible light, electrons transit from the VB of TiO₂ to the local state, leaving a hole in the VB. The electrons in the local state can continue to the CB of TiO₂. The hole in the VB is captured by -OH and H₂O to form •OH, and the electrons in the local state and CB react with the O₂ molecules dissolved in the water to form superoxide radical anions (O₂•⁻). These •OH and O₂•⁻ species are able to degrade the MB molecules in solution. Meanwhile, the photogenerated holes also have strong oxidation ability and can directly degrade MB, thereby contributing to the high visible light photocatalytic activity.

前驱体凝胶的制备流程与文献^[20]类似,不同的是本文通过控制H₂O₂的加入量和氧化时间,分别得到黄色凝胶和绿色凝胶前驱体,然后转移到内衬聚四氟乙烯的反应釜中,在160°C下水热处理不同时间.反应完成后,自然冷却到室温,过滤、洗涤、烘干后即得到蓝灰色到浅蓝色产物.其中黄色凝胶所得产物标记为Y-t,绿色凝胶所得产物标记为G-t,t表示水热处理时间(h).为了比较,本文还以TiCl₄为Ti源,采用水热法,在160°C下反应24h制备了纯的TiO₂(标记为p-TiO₂).

物相组成表征在Rigaku DMax-2500型X-射线衍射仪上进行,CuK_a(λ=0.15418nm)辐射,石墨单色器,扫描速度0.02°/min.通过HitachiH-800和JEOLJEM-2100型TEM观测材料的形貌和结构,测试时加速电压150kV,测试前样品在无水乙醇中超声分散5-10min,然后取悬浮液滴在附有碳膜的铜网上,干燥后进行观察.用ESCA 3000型X射线光电子能谱仪进行样品的表面成分和价态分析,AlK_a(hν=1486.6eV)为辐射源.采用FA-200型电子顺磁共振波谱仪在室温下进行ESR测试.在岛津UV-2550型紫外-可见吸收光谱仪上测量UV-Vis-DRS光谱,以标准BaSO₄为参比.

采用可见光下对MB溶液的降解来评价材料的性能,光源采用带有400nm滤波片的PLS-SXE300型氙灯(北京泊菲莱科技有限公司).将60mg催化剂加入到100mLMB溶液(2×10^-5mol/L)中,首先在暗处搅拌30min,使MB在催化剂表面达到吸-脱平衡,然后保持催化剂悬浮状态,开启光源进行光催化降解实验,光源与溶液的距离为30cm.每间隔10min,吸取约3mL上层溶液进行高速离心分离,用UV-7200型分光光度计在MB的最大吸收波长处测定上层清液的吸光度.

采用XRD表征产物的物相和结晶性,图1为不同状态前驱体凝胶在160°C下水热处理不同时间所得产物的XRD谱.可以看出,在水热处理时间为20h时,由黄色凝胶经水热处理后所得产物中TiH₂的含量很低,除了在2q=34.9°和40.6°处出现两个微弱的TiH₂的衍射峰之外(图1,JCPDS 65-0934),只观察到锐钛矿相TiO₂的衍射峰(JCPDS21-1272).而以绿色凝胶为前驱体所得产物中TiH₂的含量很高.进一步延长水热处理时间时,黄色凝胶前驱体经24h水热处理即得到纯的锐钛矿相TiO₂,黄绿色凝胶经27h完全转化为TiO₂^[22],而绿色凝胶经水热处理27h后仍含有少量TiH₂,只有延长至32h后才得到纯的锐钛矿相TiO₂.

在H₂O₂存在下,随体系的酸碱条件不同,Ti⁴⁺显现不同的颜色,在酸性条件时为橙色,pH~8时为黄色,强碱性条件时为无色.而本文在凝胶的制备过程中始终保持反应体系的pH为中性,因此所得不同状态的凝胶是由于H₂O₂对TiH₂的氧化程度不同、凝胶中Ti的含氧化合物和TiH₂的比例不同造成的.在相同的水热处理条件下,由黄色凝胶得到的产物中TiH₂的含量低,说明它被氧化的程度大,同时在水热处理过程中更易完全转化为TiO₂;而绿色凝胶中TiH₂被氧化的程度小,TiH₂含量高,需要较长的水热处理时间才能转化为TiO₂.另外,无论以哪种状态的凝胶为前驱体,随水热处理时间的延长,TiO₂的衍射峰越来越尖锐,说明TiO₂的结晶性越来越好,晶粒度越来越大.因此,在以TiH₂为Ti源,以H₂O₂为氧化剂制备凝胶前驱体并采用水热处理制备Ti³⁺自掺杂的纳米TiO₂时,可以通过控制氧化时间和氧化程度,得到黄色凝胶,从而缩短水热处理时间.

图2为不同状态前驱体凝胶在160°C下水热处理不同时间所得产物的TEM和HRTEM照片.可以看出,黄色凝胶前驱体在160°C下水热处理20h所得产物为分散较好的小颗粒组成(图1(a)),但由于样品中TiH₂的含量非常少,因此在其HRTEM照片中没能发现TiH₂的晶格结构.当水热处理时间延长到24h后,样品形貌变得更加规则,而且分散性得到改善(图1(b)),XRD结果表明,此时样品为纯的锐钛矿相TiO₂(图1,Y-24).而在相同的条件下,由绿色凝胶经水热处理20h所得产物形貌不规则,粒度分布范围较宽(图1(c)),HRTEM结果表明,样品中仍有TiH₂,图2(d)中d=0.44nm为TiH₂^[23].这进一步说明采用绿色凝胶作为前驱体进行水热处理时,在较短的时间内得到的产物为未完全转化的TiH₂和TiO₂的复合物.随水热处理时间的延长,样品的形貌越来越规则,至32h时,由绿色凝胶所得产物的分散较好、具有一定形貌的颗粒(图1(e)),且为纯的锐钛矿相TiO₂(图2(f)),与XRD结果一致.比较由不同状态的前驱体凝胶全部转化为TiO₂后的结果可知,采用黄色凝胶所得产物的粒度最小,而采用绿色凝胶所得产物的粒度最大(图2(b),(e)).对非均相光催化反应来说,催化剂的粒度越小,则比表面积越大,样品的分散性好,吸附性强,催化剂更容易分散并与溶液中的有机分子接触,从而有利于光催化反应的进行^[24].

为了确定样品的组成和Ti的状态,对样品进行了XPS和ESR分析.图3(a)和(b)分别给出了样品Y-20,Y-24和G-32的Ti2p和O1s高分辨XPS谱.图3c,d分别是样品G-32和Y-24的Ti2p与O1s的拟合结果.Ti⁴⁺的2p结合能位于~464.5和~459.0eV处^[25],而样品Y-20和Y-24的Ti⁴⁺结合能都向低能方向移动,这是因为样品中存在Ti³⁺所致^[17].随水热处理时间的延长,向低能方向移动的程度减小,说明样品中Ti³⁺的浓度降低.对样品G-32的Ti2p结合能进行拟合后得到四个峰(图3(c)),分别为在457.7eV处的Ti³⁺2p_3/2,458.7eV处的Ti⁴⁺2p_3/2,463.0eV处的Ti³⁺2p_1/2和464.5eV处的Ti⁴⁺2p_1/2^{[25, 26]},但是Ti³⁺的结合能峰强度很弱,说明在160°C下水热处理32h后样品中Ti³⁺的含量低.

Ti-O键中O的结合能位于529.5-530.5eV处^[27].图3b是样品的O1s高分辨XPS谱,三个样品的O结合能都处在530eV左右,说明样品中O主要以Ti-O键的形式存在,但是O结合能峰不对称.对样品Y-24的O结合能进行拟合后得到三个峰(图3(d)),其中位于530.1eV处的谱峰对应于Ti-O,而531.3eV处的峰对应表面氧缺陷即Ov^[28],另外在532.3eV处还存在一个微弱的峰,这对应于-OH中O的结合能特征峰^{[29, 30]},说明颗粒表面存在-OH.

ESR被用来研究物质中未耦合电子及它们与周围原子的相互作用,而且从峰的强度还可进行定量分析.图4为样品Y-20,Y-24,G-27和G-32以及纯TiO₂的ESR谱.与纯TiO₂相比,采用不同凝胶前驱体进行水热处理所得样品中,都在g=1.94左右都出现一个信号峰,这对应于Ti³⁺的信号峰.随水热处理时间的延长,该峰强度降低,说明样品中Ti³⁺的浓度随水热处理时间的延长而下降^[31],和XPS结果一致.而作为电子陷肼的Ov在g=2.003左右产生信号峰,表明样品中存在表面Ti^3+[32],从而说明样品中的Ti³⁺既存在于晶格内部,也存在于表面.样品中Ti³⁺和Ov的量都随水热处理时间的延长而逐渐降低,与XPS 结果一致.

XPS和ESR结果表明,不同氧化程度的凝胶前驱体经水热处理后所得样品中含有晶格Ti³⁺、表面Ti³⁺和Ov,这些Ti³⁺和Ov的存在对其光电转换性能有重要影响,而样品表面的-OH基团能够捕获空穴形成重要的氧化性羟基自由基(•OH),这对有机污染物分子的光降解具有重要作用^{[33, 34]}.

图5为p-TiO₂与不同条件下所得样品的UV-Vis DRS谱.与p-TiO₂相比,不同条件下所得样品在可见光区都有一定的吸收.其中Y-20和G-27样品虽然在可见光区的吸收较强,但是其在紫外光区的吸收弱.这是由于样品中有未完全反应的TiH₂引起的;另一方面,由于水热处理时间较短,样品结晶性差、缺陷较多导致的.无论是采用黄色凝胶还是绿色凝胶作为前驱体,所得样品在紫外光区的吸收都随水热处理时间的延长而增强.这是因为样品的结晶性越来越好,缺陷越来越少所致.另外,无论是在可见光区还是紫外光区,样品Y-24和G-32都有强的吸收,这说明样品结晶性好,缺陷少,同时样品中存在Ti³⁺.但是在可见光区,样品Y-24的吸收强于样品G-32,这说明Y-24样品中Ti³⁺的浓度比G-32中的高,这也与XPS和 ESR结果一致.

结合以上结果我们认为,在H₂O₂的氧化作用下,TiH₂颗粒表面的Ti²⁺逐渐被氧化,并且在颗粒表面形成Ti的含氧化合物及其水合物.当对前驱体凝胶进行水热处理时,表面的Ti的含氧化合物及其水合物发生脱水反应,同时颗粒内部未完全被氧化的TiH₂和表面Ti的含氧化合物在界面处发生离子扩散与交换反应,从而形成Ti³⁺和Ov^[22].由于黄色凝胶比绿色凝胶表面被氧化的程度大,TiH₂含量低,所以在较短的水热处理时间下即可得到纯锐钛矿相Ti³⁺自掺杂纳米TiO₂.与通常采用的以四价钛醇盐为原料制备含有表面Ti³⁺的TiO₂相比^{[35, 36]},我们采用的原料为低价钛的化合物TiH₂,在制备过程中,在反应初期Ti²⁺离子位于颗粒的内核,然后通过离子扩散与交换反应得到TiO₂晶格内部含有Ti³⁺的TiO₂,这保证了材料的稳定性;同时采用H₂O₂为氧化剂,可以保证所得产物的纯净性,而且可以通过控制H₂O₂对Ti H₂的氧化程度及后续的水热处理条件,调控最终产物中Ti³⁺和Ov的量,进而调节产物的性能.

作为一种偶氮类染料,MB溶液常用来作为光催化降解的模拟废水.本文在可见光下对MB溶液的光催化降解性能和循环使用性能,结果见图6.可以看出,在打开光源前,样品都具有一定的吸附性能,其中以黄色凝胶为前驱体经水热处理20h所得样品的吸附性能最好.这是因为水热处理时间短,样品的粒度小、比表面积较大引起的,随水热处理时间的延长,样品的吸附性能逐渐下降,这是因为样品的粒度越来越大,比表面积越来越小造成的.当打开光源进行光照时,采用黄色凝胶、在160°C下水热处理24h所得样品的催化性能最好,其次是样品G-32.虽然样品Y-20和G-27在可见光区的吸收较强,但其催化性能不是很好.这是因为样品中含有未完全反应的TiH₂导致的在可见光区的吸收;另一方面,样品的结晶性较差,颗粒内部含有较多的缺陷,这些缺陷会成为光生电子和空穴的复合中心,因此其光催化性能相对较差.与G-32相比,样品Y-24水热处理时间短,粒径小,比表面积大,有利于光催化反应的进行;同时从XPS和ESR结果可知,在较短的水热处理时间下样品中Ti³⁺和Ov的含量较高,对可见光的响应强,因此样品Y-24的光催化性能最好.

进一步采用准一级动力学反应方程式K_app=ln(C₀/C)/t计算光催化反应速率,K_app是表观一级反应速率常数,图6(b)给出了不同条件下所得样品及p-TiO₂在达到吸附平衡后光催化降解MB的反应动力学曲线.可以看出,样品的催化活性从高到低的次序为Y-24>G-32>Y-20>G-27>p-TiO₂,其中样品Y-24的光光催化速率常数为0.0439min^-1,是p-TiO₂的18.3倍.

对催化剂来说,材料的耐久性是一个重要指标,对材料的重复使用可证明材料具有耐久性.图6(c)是样品Y-24和G-32的循环使用性能.可以看出,随使用次数的增加,样品的催化性能有所下降,但使用8次后其催化性能仍能达到95%左右,说明样品有较好的实际应用前景.

已有研究表明,Ti³⁺和Ov能够在TiO₂导带(CB)下方的0.75-1.18eV范围内形成局域态,当Ti³⁺和Ov达到一定浓度时,这种局域态类似形成掺杂能级,从而降低TiO₂的禁带宽度^[17].在可见光激发下,电子可以从TiO₂的价带跃迁至由Ti³⁺和Ov形成的局域态中,然后从局域态再跃迁至导带,从而使TiO₂能够被可见光甚至红外光激发^{[10, 20, 36]}.虽然Ti³⁺自掺杂的纳米TiO₂有很多Ov和Ti³⁺缺陷,而有研究表明,TiO₂中的缺陷特别是体相缺陷的增加会导致光生电子和空穴的复合几率增加,从而降低光催化活性^[37].但是,由于Ti³⁺的掺杂能够降低TiO₂的电阻,加速电子的传输^[38];同时由于Ov和Ti³⁺的协同作用,能够延长光生载流子的寿命,降低复合几率^{[11, 28, 39, 40]}.图6(d)给出了Ti³⁺自掺杂的纳米TiO₂可见光催化反应机理示意图.在可见光激发下,电子从TiO₂的价带(VB)跃迁至局域态中,在VB中留下空穴,处于局域态中的电子可以继续跃迁至TiO₂的CB中.处于VB中的空穴被催化剂表面的-OH和溶液中的H₂O分子捕获形成•OH,而处于局域态和CB中的光生电子和水中溶解的O₂分子形成超氧自由基负离子(O₂•^-),•OH和O₂•^-都能将溶液中的MB分子降解;同时,光生空穴也具有强的氧化能力,能够直接将MB分子催化降解,从而使表现出良好的可见光催化活性.