A. Crystal structure and morphology

Figure 1(a) represents XRD patterns of prepared black TiO₂ samples by Mg reduction with the Mg and TiO₂ ratio of 9:1 in nitrogen atmosphere at the calcining temperatures of 400, 450, 500, and 600 °C. As shown in Fig. 1(a), for the samples at temperatures 400, 450, and 500 °C, the diffraction peaks at 2θ = 25.4, 37.8, 47.5, 54.3, and 62.4°, respectively, are the TiO₂ anatase crystal planes of (101), (004), (200), (105), and (204); severally, implying the three samples are all anatase TiO₂,^{Reference Kumar, Rajadurai and Sundararajan47,Reference Xia, Li, Zhang, Kruger, Murowchick, Selloni and Chen48} and there were no other phases, illustrating that the samples are kept in the same phase as below 500 °C, and as the temperature increases, the crystallinity of the sample increases, which will be beneficial to the photocatalytic reaction.

FIG. 1.

XRD curves of TiO₂ samples prepared by Mg reduction at different calcination temperatures (a) and the photos of different color prepared TiO₂ samples (b).

Comparing the XRD curves of the former three samples, the XRD curves of the fourth sample show strong diffraction peaks at 2θ = 27°, indicating that the anatase phase begins to transform into rutile phase, and the content of rutile phase in the product is higher. This indicates that the anatase phase begins to transform into rutile phase at 600 °C, which is different from that of many literatures. Most of the literatures reported that the phase transition temperature of bulk TiO₂ is higher. The different results in our tests are mainly attributed to two reasons. One is due to the quantum effect of nanoparticles. The phase transition temperature of different nanocrystals is not the same. When the grain size is reduced to nanoscale, with the further increase of the temperature, the grain begins to grow large, but the grain size is not uniform and presents a normal distribution. Thus, smaller-grain-sized nanoparticles exhibit lower phase transition temperatures. Another reason may be due to the addition of the Mg reductant, which is active in high temperature and the chemical reaction between the Mg and the relatively stable crystal structure of TiO₂ occurred in the calcination process, which affects the crystal structural changes, further causing changes in its phase transition temperature. Based on the high crystallinity anatase phase TiO₂, the photocatalytic performance is even more superior. In the next comparative experiment, the calcination temperature of 500 °C will be used as the reference condition.

In Fig. 1(b), we can see the color deepening as the temperature increases, and the color changes from white to gray and black. Therefore, Mg reduction of TiO₂ nanoparticles leads to surface modification, which is reflected in enhanced light absorption and the color change.^{Reference Li, Xu, Yan, Liu, Chen, Luo, He and Shen45}

Based on XRD data, the corresponding crystallite sizes were estimated using the Scherrer formula from the full-width at half-maximum of the (101) diffraction peak (2θ = 25.4°) of TiO₂. The statistics of crystallite sizes are shown in the Table I. From 400 to 500 °C, when the temperature rises, the crystallite size increases slowly. However, the crystallite size of the sample 600 °C is higher than others, which may indicate that the presence of Mg at high temperature also facilitates the phase transformation. Similar results were also found in other reports.^{Reference Sinhamahapatra, Jeon and Yu46,Reference Wang, Fu, Yin, Wu, Guo, Xu and Zhong49}

TABLE I.

The crystallite sizes of samples using the Scherrer equation from XRD data.

Figure 2(a) represents the XRD patterns of black nano-TiO₂ samples with different ratios of Mg and TiO₂. In Fig. 2(a), the XRD curves represent the characterization results of samples: pure TiO₂ and Mg and TiO₂ ratios of 7:1, 9:1, and 11:1, prepared at the calcining temperature of 500 °C. The crystal phase composition of three samples was anatase TiO₂ except the sample 11:1, according to Joint Committee on Powder Diffraction Standards (JCPDS) card no. 21-1272, with no other crystalline impurities detected. The characteristic peaks of nano-TiO₂ crystals synthesized by the sol–gel method were sharp, suggesting high crystallinity. When the ratios of TiO₂ and Mg increase to 11:1, the TiO₂ is completely reduced because of excessive reductant, resulting in a lower valence titanium oxide (Ti₂O₃). Figure 2(b) shows corresponding samples. It can be seen from Fig. 2(b) that the color of TiO₂ gradually deepens with increasing ratio of Mg and TiO₂.

FIG. 2.

XRD curves of samples with different ratios of Mg and TiO₂ (a) and the photos of the different samples (b).

As shown in Table II, the crystallite sizes of pure TiO₂ and black TiO₂ samples with reducing agent ratio of 7:1, 9:1, and 11:1 are 12.3, 12.5, 13.6, and 15.7 nm, respectively, based on the XRD results calculated by Scherrer equation. When the ratios of Mg and TiO₂ are 7:1 and 9:1, the increase in average crystal size is not evident. The grain size of the sample with a reducing agent ratio of 11:1 is obviously larger than those of other samples, which may be attributed to excessive magnesium, for which titanium dioxide reacts dramatically with magnesium, forming the new compound Ti₂O₃, as shown in Fig. 2(a), so that the crystallite size of the sample increases more than others.

TABLE II.

The crystallite sizes of samples prepared at different ratios of Mg and TiO₂.

B. Raman spectroscopy

Figure 3 shows the Raman spectroscopy of the prepared black nano-TiO₂ with different ratios of Mg reduction. The Raman spectrum of three curves shows characteristic Raman bands of the typical anatase TiO₂ phase with the strongest E _g band at around 144 cm⁻¹. The results show that the three samples are mainly anatase TiO₂. It is observed that with the increase of Mg reductant proportion, the intensity of each characteristic peak gradually decreases and the peaks tend to be broadened. Several small peaks of sample of ratio 9:1 with the most amount of Mg reductant tend to be flattened. In particular, the characteristic peak at 516 cm⁻¹ completely disappears, indicating that the crystallinity of the sample gradually decreased. It shows that because of the addition of metal reductant, the original crystalline surface of TiO₂ began to appear disorderly layer structure, and with the increasing amount of reducing agent, the disordered layer thickness increases, and the core–shell structures become more obvious.^{Reference Sinhamahapatra, Jeon and Yu46} As reported in the previous studies, the shift and broadening of the peaks of TiO₂ were ascribed to the crystal domain size and nonstoichiometry.^{Reference Wang, Yang, Lin and Yin38,Reference Sinhamahapatra, Jeon and Yu46}

FIG. 3.

Raman curves of white and black nano-TiO₂ samples.

C. TEM analysis

Figure 4 is the test results of TEM and high resolution transmission electron microscopy (HRTEM) of the black nano-TiO₂ sample at the calcining temperature of 500 °C with Mg and TiO₂ ratio of 9:1. Figure 4(a) reflects the aggregation of black nano-TiO₂ particles, the grain size is about 15 nm, and the particle size distribution is relatively uniform. As is seen in Fig. 4(c), the average interplanar space of the sample is 0.35 nm by measuring lattice fringes. The sample’s Debye ring has more bright spots in the picture, and the Debye ring is more obvious, indicating that the sample has higher crystallinity [Fig. 4(b)]. Figs. 4(c) and 4(d) together indicate that the sample has a good crystal structure. It can be seen from Fig. 4(d) that the reduced nanocrystal displays a unique crystalline core–disordered shell structure and the thickness of the disordered layer is about 2–3 nm. This is because after the Mg reduction reaction, a disordered layer microstructure is formed on the surface of the black nano-TiO₂ particles. This is consistent with the results of Raman spectroscopy and has similar microstructure to black nano-TiO₂ particles reported in other studies.^{Reference Chen, Liu and Huang30,Reference Sinhamahapatra, Jeon and Yu46}

FIG. 4.

TEM and HRTEM images of black nano-TiO₂ samples at the calcining temperature of 500 °C with the ratio of Mg and TiO₂ (9:1): the aggregation TiO₂ particles (a), the Debye ring image of the black nano-TiO₂ (b), the average interplanar space (c), the crystalline core–disordered shell structure (d).

D. Specific surface area and pore size distribution

Figure 5(a) represents N₂ adsorption–desorption isotherms for the as-prepared samples. It can be observed that the adsorption–desorption isotherms in the low pressure region (P/P ₀ < 0.4) are coincident, which shows that there is no microporous structure (<2 nm) in the sample. In the region of medium pressure (0.4 < P/P ₀ < 0.8), a hysteresis loop appeared in the pure TiO₂ sample. The adsorption–desorption isotherms for the pure TiO₂ showed a typical type IV behavior, indicating the formation of mesopores. For the black TiO₂ samples, the hysteresis loop broadened to high pressure (P/P ₀ > 0.8), indicating mesoporous–macroporous structure, which is attributed to the Mg reduction.

FIG. 5.

The N₂ adsorption–desorption isotherms for different samples with the ratio of Mg and TiO₂ (a), and figure of pore size distribution (b).

Specific surface areas were determined as shown in Table III. The obtained surface areas are 70.67, 38.76, 32.15, and 17.00 m²/g for pure TiO₂ and the black TiO₂ samples prepared with Mg and TiO₂ ratios of 7:1, 9:1, and 11:1, respectively. The specific surface area of pure TiO₂ is larger than those of all the black TiO₂ samples. The data showed that the specific surface areas decrease obviously with increase of Mg content, indicating that this decrement is mainly due to Mg reduction. The drastic decrement of the specific surface areas for 11:1 can be attributed to the increase of crystallite size, for which there is excessive Mg in the reduction reaction and the Ti₂O₃ was formed as shown in XRD results.

TABLE III.

The specific surface areas of different samples.

Figure 5(b) shows pore size distribution for different samples with Mg and TiO₂ ratios of 7:1, 9:1, and 11:1. The data showed that the pore size increases obviously with increase of Mg content. The average mesopore sizes of the samples estimated using the Barrett–Joyner–Halenda (BJH) approach from the desorption branch are 6.4, 18.5, 30.7, 35.1 nm for the pure TiO₂, the samples with different ratios (7:1, 9:1, and 11:1), respectively. From the crystallite size of the samples as discussed in the XRD results, the pores were probably formed by the agglomeration and connection of adjacent nanoparticles in the sample, which were seen in the TEM results [Fig. 4(a)].

E. PL emission spectra

The PL emission spectra are effective to track the behavior of light-generated electrons and holes in the samples because PL emission results from the recombination of free carriers. The PL spectra of pure TiO₂ and the Mg-reduced black TiO₂ samples are shown in Fig. 6. The shapes of the emission spectra of two samples are similar profiles. Four main emission peaks are observed at 402 nm (3.12 eV), 426 nm (2.88 eV), 450 nm (2.75 eV), and 472 nm (2.65 eV), respectively. The first one is attributed to the emission of band gap transition of anatase TiO₂. The peaks at 450 and 472 nm are attributed to the free excitons at the band edge. Moreover, several small PL peaks observed in the wavelength range are mainly due to the surface defects of the TiO₂ samples. Compared with pure nano TiO₂, the peak intensity of black nano-TiO₂ is weaker. The intensity of the PL line is related to light-generated electrons and holes recombination. This shows that the black nano-TiO₂ has a relatively low light-generated electrons and holes recombination rate, which benefits enhancement of the photocatalytic activity.

FIG. 6.

Photoluminescent spectra of pure TiO₂ and the black TiO₂ sample calcined at the temperature of 500 °C with the ratio of Mg and TiO₂ (9:1).

F. UV-vis diffuse reflectance spectra

Figure 7 shows the UV-vis diffuse reflectance curves of pure nano-TiO₂ and black nano-TiO₂ samples. The UV-vis curve is used to analyze the light intensity response of the different samples. It can be seen from the figure that all the TiO₂ samples have strong absorption in the UV region of 200–400 nm with little differences, indicating that the modified black nano-TiO₂ sample basically remains unchanged in response to UV light. The absorption edge of pure TiO₂ is about 410 nm. Compared with the pure TiO₂, the absorption edges (about 440, 460, and 510 nm, respectively) of the black TiO₂ samples (7:1, 9:1, and 11:1) shift into visible region, and their light absorption in the visible region obviously intensifies. In the visible region of 400–800 nm, the absorption intensity of all samples showed a decreasing trend, but the degree of decline is different. The curve of white nano-TiO₂ sample drops sharply at 390 nm, and the absorption intensity in the visible region is almost invisible. In contrast, the light absorption intensity in the visible region of black nano-TiO₂ samples is relatively stronger, and with the increase of the ratio of Mg and TiO₂, the light absorption intensity increases. This shows that as the amount of reducing agent increases, the sample color deepens and the visible light absorption intensity increases. As shown in other reports, the surface disorder layer is one of the reasons for the darkening of nano-TiO₂.^{Reference Zhang, Hu, Zhang, Wang, Sun, Li, Qiao, Wang and Zhou23,Reference Sinhamahapatra, Jeon and Yu46,Reference Wang, Fu, Yin, Wu, Guo, Xu and Zhong49}

FIG. 7.

UV-vis diffuse reflectance spectra of normal nano-TiO₂ and black nano-TiO₂ samples.

G. Photocatalytic hydrogen generation

The photocatalytic hydrogen generation from water under the full solar wavelength light irradiation was used to evaluate the photocatalytic activity of black nano-TiO₂ samples. Pt (1 wt%) was used as a cocatalyst by a UV-vis light–induced reduction procedure. Hydrogen production profile under the full solar wavelength range of light for samples in different calcining temperatures is shown in Fig. 8(a). After 3 h, the H₂ production rate of pure TiO₂ (prepared by sol–gel method) is about 0.81 mmol/(h g). It is obvious in Fig. 8(a) that the photocatalytic activity of the black nano-TiO₂ has been improved with the Mg reduction temperature. The 500 °C Mg–reduced black TiO₂ sample steadily produced hydrogen gas at about 7.44 mmol/(h g). Compared with pure TiO₂, the rates of hydrogen generation for the black TiO₂ samples have been enhanced obviously. The sample below 500 °C is still an anatase crystal, and the crystallinity becomes higher as the temperature rises, so the photocatalytic rate will increase when the calcining temperature rises.^{Reference Green, Xu, Liu, Zhao, Li, Liu, Qin, Zhu, Shen and Chen50,Reference Xia, Zhang, Oyler and Chen51} However, for the 600 °C Mg–reduced black TiO₂ sample, the rate of hydrogen generation is lower than that of 500 °C Mg–reduced black TiO₂ sample, which may be attributed to the phase transform and increase of crystal size for the 600 °C Mg–reduced black TiO₂ sample, as shown in above discussion.^{Reference Riaz, Qi, Fang, Xu, Zhou, Jin, Hong, Zhi and Liu52,Reference Yan, Chen, Zou and Wang53} These results have illustrated that within 400–500 °C, increasing calcining temperature can enhance the photocatalytic activity of hydrogen generation.

FIG. 8.

The black TiO₂ photocatalytic water splitting for hydrogen generation under the full solar wavelength range of light at different calcining temperatures (a), with different ratios of Mg and TiO₂ (b), and recycling measure of hydrogen generation through photocatalytic water splitting with black TiO₂ (500 °C, 9:1) under the full-spectrum solar light source (c).

The continuous hydrogen production profile and the rate of hydrogen production under the full solar wavelength range of light for different samples in different magnesium and TiO₂ ratios are shown in Fig. 8(b). Compared with white TiO₂, the rates of hydrogen production for the black TiO₂ samples have been enhanced significantly. The hydrogen generation rate increases as the ratio of Mg and titanium dioxide increases to 9:1. However, the hydrogen generation rate starts to decrease at the ratio of 11:1 although it is observed that with the increase of the Mg amount, the sample turns darker and the absorption of visible light also increases. This result can be ascribed to the generation of new recombination sites because of over-reduction in the presence of a high amount of Mg. Because of its excellent photocatalytic activity, the stability of the sample (9:1) as a photocatalyst was further studied for hydrogen production under full solar light. As shown in Fig. 8(c), the recycle experiments reveal that the sample does not exhibit evident decline of photocatalytic activity for H₂ generation after 6 photocatalytic cycles. The cycling tests reveal that the black TiO₂ sample was very stable in 6 photocatalytic cycles.

As shown in Fig. 9, the light is radiated on the semiconductor. When the energy of the radiation is greater than or equal to the forbidden band width of the semiconductor, the electrons in the semiconductor are excited to transition from the valence band to the conduction band. The holes remain in the valence band, separating the electrons and holes, and then reducing the water to hydrogen or oxidizing the water to oxygen at different locations in the semiconductor. The water splitting reaction comprises two half reactions, the H₂-evolution reaction [Eq. (1)] and O₂-evolution reaction [Eq. (2)]. The standard redox potential ΔE ⁰ of the multielectron water splitting (WS) is given in the equation.

(1)$$2{{\rm{H}}^ + } + 2{{\rm{e}}^ - } \to {{\rm{H}}_2}\left( {\rm{g}} \right)\quad \Delta {E^0} = - 0.41\;{\rm{V}}\quad ,$$

(2)$$2{{\rm{H}}_2}{\rm{O}}\left( {\rm{l}} \right) \to {{\rm{O}}_2}\left( {\rm{g}} \right) + 4{{\rm{H}}^ + } + 4{{\rm{e}}^ - }\quad \Delta {E^0} = + 0.82\;{\rm{V}}\quad .$$

FIG. 9.

Schematic diagram of the black TiO₂ photocatalytic water splitting for hydrogen generation.

For water to be decomposed into hydrogen and oxygen, the bottoms of the conduction band must be located at a more negative potential than the reduction potential of H⁺ to H₂ (−0.41 V, at pH 7), whereas the top of the valence band must exceed the oxidation potential of H₂O to O₂ (0.82 V).^{Reference Li, Yu, Low, Fang, Xiao and Chen26} Thus, the minimum band gap energy of the photocatalyst should be 1.23 eV to convert the energy into H₂ and O₂.^{Reference Li, Yu, Low, Fang, Xiao and Chen26} For the black TiO₂ samples, the valence band maximum energy blue-shifts toward visible region. Combined with the results from UV-vis diffuse reflectance spectra suggests a narrowed band gap, which is due to the Mg reduction. This is the reason that the black samples can be used as efficient photocatalysts to decompose water under the full solar wavelength range of light.