Formation mechanism of [111]-faceted anatase nanocrystals 75

Figure 2.7. XRD patterns of products synthesized by hydrothermal treatment of TTIP-TMA

suspension solution at 200 ^oC and different reaction times.

The FE-SEM images of aforementioned products are shown in Figure 2.8. Spherical and irregular polycrystal particles with micrometers size are obtained after drying the colloidal solution prepared by hydrothermal treatment at 200 ^oC for 17 and 25 min. The main phases of these samples are layered titanate. The triangular pyramidal anatase nanoparticles with a size of about 70 nm are obtained after reaction for 30 min and 1h, and finally cuboidal anatase nanocrystals are formed after the hydrothermal reaction over 2h.

Figure 2.8. FE-SEM images of the products synthesized by hydrothermal treatment TTIP-TMA suspension solution at 200 ^oC for (a) 17 min, (b) 30min, (c) 1h and (d) 2h, respectively.

TEM images reveal that the polycrystal particles with micrometers size obtained by hydrothermal reaction for 17-25 min are constructed from agglomerated nanofragments (Figure 2.9a and c). Lattice fringes with a d-value of 0.28 nm are observed in the HR-TEM image of the polycrystalline particles, which can be assigned to the layered titanate nanosheets because this sample is a layered titanate phase (Figure 2.7). Besides the lattice fringes with a d-value of 0.28 nm, some lattice fringes with a d-values of 0.351 nm are observed in the sample after reaction for 25 min, which can be assigned to the (101) plane of anatase. The main phase is the layered phase and a small amount of anatase phase is formed after the reaction for 25 min (Figure 2.9d). It has been reported that the titanate

nanosheet with a lepidocrocite-like structure can be formed by reacting TBAOH-HTO and PA-HTO solution.^{25, 53} The 0.28 nm d-value of lattice fringe is smaller than that of the 0.37 nm of layered titanate with the lepidocrocite-like structure, suggesting they have different structures.

The triangular pyramidal nanoparticles obtained after the reaction for 30 min at 200

oC are polycrystalline particles constructed from anatase nanofragments with a size of about 30 nm (Figure 2.9e). The triangular pyramidal particles show a Fast Fourier Transform (FFT) diffraction pattern of a polycrystal (Figure 2.9f). The polycrystalline triangular pyramidal anatase nanoparticles are transformed to single crystalline triangular pyramidal anatase nanoparticles after reaction for 1 h, which have a crystal size of about 40 nm and exhibit a FFT diffraction pattern of single crystal (Figures 2.9g and h). After the reaction for 2 h, [111]-faceted anatase single nanocrystals with a size of about 50 nm and the particle morphology close to cubic are obtained (Figure 2.9i). The [111]-faceted anatase single nanocrystals exhibit (101) and (011) planes in the HR-TEM. The uniform cuboidal [111]-faceted anatase single nanocrystals are formed after the reaction for 24 h (Figure 2.6d). Noteworthy, [111]-faceted and {010}-faceted anatase nanocrystals coexist in the polycrystalline particles obtained by hydrothermal treatment for 25 min and 30 min, seen in Figure 2.10. However, the {010}-facet dissolves due to its high-surface-energy with the hydrothermal reaction progress, finally remaining the [111]-facet.

Figure 2.9. TEM (a, c, e, g) and HR-TEM (b, d, i) images and FFT diffraction (f, h) patterns of products synthesized by hydrothermal treatment of TTIP-TMA suspension solution at 200 ^oC for (a, b) 17 min, (c, d) 25 min, (e, f) 30min, (g, h) 1 h and (i) 2 h.

Figure 2.10. HR-TEM images of products synthesized by hydrothermal treatment of TTIP-TMA suspension solution at 200 ^oC for (a, b) 25 min, (c, d) 30min.

Figure 2.11 shows the FT-IR spectral on the hydrothermally treated samples at different temperatures. In the FT-IR spectrum of TTIP (Figure 2.11a), the vibration bands at 2967, 2929 and 2864 cm^-1 can be assigned to the C-H stretching vibrations of -CH- and -CH3 group in -O-CH-(CH3)2 of TTIP. And the vibration bands at 1375, 1361 and 1463 cm^-1 can be assigned to the C-H bending vibrations of -CH- and -CH3 group in -O-CH-(CH3)2 of TTIP. In the FT-IR spectrum of TMAOH (Figure 2.11b), the broad vibration bands around 3430 cm^-1 can be assigned to the stretching vibration of O-H. The vibration bands at 1485 and 951 cm^-1 can be assigned to the C-H bending vibrations of -CH3 group in (CH3)4N⁺.

In the FT-IR spectrum of TTIP-TMA suspension solution (Figure 2.11c), the C-H

bending vibrations of (CH3)4N⁺ at 1485 and 951 cm^-1, and also the C-H stretching vibrations of -O-CH-(CH3)2 group at 2967, 2929 and 2864 cm^-1 are observed, suggesting this sample contains (CH3)4N⁺ and -O-CH-(CH3)2 group. In the FT-IR spectrum of sample obtained after reaction at 60^oC, the C-H bending vibrations of (CH3)4N⁺ at 1485 and 951 cm^-1, and also the C-H stretching vibrations of -O-CH-(CH3)2 group at 2967, 2929 and 2864 cm^-1 are observed, suggesting this sample contains also (CH3)4N⁺ and -O-CH-(CH3)2 group. However, in the FT-IR spectrum of sample obtained after reaction at 80^oC, the C-H bending vibrations of (CH3)4N⁺ at 1485 and 951 cm^-1 are observed, but without the C-H stretching vibrations of -O-CH-(CH3)2 group at 2967, 2929 and 2864 cm^-1, suggesting this sample contains also (CH3)4N⁺ but without -O-CH-(CH3)2 group. These results suggest that the 1.54 nm-layered phase contains some -O-CH(CH3)2 (IP) groups, but without IP group in the 1.18 nm-layered phase. Furthermore, both of 1.54 nm- and 1.18 nm-layered phases contain TMA⁺ in the interlayer spaces.

Figure 2.12 shows the TG-DTA curves on the hydrothermally treated samples at different temperatures. In the TGA-DTA curves of TTIP-TMA suspension, the endothermic peak at 108 ^oC with weight loss corresponds desorption of the adsorbed water. The exothermic peaks at 326 and 429 ^oC correspond to the decomposition of TMA⁺ ions and IP groups, respectively. With increasing reaction temperature, the intensity of the exothermic peak at 429 ^oC reduces. This result suggests the IP group content in the hydrothermally treated sample decreases with increasing the reaction temperature, and almost no IP group in the 1.18 nm-layered phase.

Figure 2.11. FT-IR spectra of (a) TTIP, (b) TMAOH, (c) TTIP-TMA suspension, and products

obtained after hydrothermal treatment at (d) 60 ^oC (1.54 nm-layered phase) and (e) 80 ^oC (1.18 nm-layered phase).

Figure 2.12. TG-DTA curves of (a) TTIP-TMA suspension and the products obtained after hydrothermal treatment at (b) 60 ^oC (1.54 nm-layered phase) and (c) 80 ^oC (1.14 nm-layered phase).

The FTIR and TGDTA results suggest that the 1.54 nmlayered phase contains -OCH(CH3)2 (IP) groups, but IP group does not presence in the 1.18 nm-layered phase.

However, both of the two layered phases contain TMA⁺ ions. These results reveal the hydrolysis reaction of TTIP in the TMAOH solution is uncompleted at the stage of formation of the 1.54 nm-layered phase, where some IP groups combine with Ti(IV);

while the hydrolysis reaction is complete at the stage of formation of the 1.18 nm-layered phase.

On the basis of above results, a possible formation reaction mechanism for the [111]-faceted anatase single nanocrystal is proposed and shown in Scheme 2.1. Firstly, a hydrolysis reaction of TTIP in TMAOH solution causes formation a layered titanate with basal spacing of 1.54 nm and TMA⁺ in the interlayer spaces. In this layered structure, there are some IP groups combining on the titanate layers due to the uncompleted hydrolysis reaction. After completing the hydrolysis reaction, the basal spacing of the layered titanate reduces to 1.18 nm, in which without IP groups in the layered structure.

This layered titanate can be exfoliated to titanate nanosheets, which causes the formation of the colloidal solution, similar to other layered metal oxides with TMA⁺ ions in the interlayer space.^27,29 The titanate nanosheets have a small size of about 30 nm in wide (Figure 2.9a), corresponding to their low crystallinity. The small-sized titanate nanosheets are assembled into the triangular pyramidal particles by stacking together, and then are transformed into anatase structure by the topochemical reaction under hydrothermal conditions, which causes formation of anatase polycrystalline particles with triangular pyramidal morphology (Figure 2.9e). The triangular pyramidal polycrystalline anatase nanoparticles change to the triangular pyramidal anatase single nanocrystals (Figure 2.9g), and finally, to the [111]-faceted cubic anatase single nanocrystals (Figure 2.9i) under the

hydrothermal reaction conditions by Ostwald ripening crystal growth mechanism of dissolution-deposition reaction.

Scheme 2.1. Formation reaction mechanism of [111]-faceted anatase single nanocrystal in TTIP-TMAOH hydrothermal reaction system

2.3.3 Effect of crystal facet on photocatalytic activity and electronic band structure

To study the effect of the crystal-facet on the photocatalytic activity, we measured photocatalytic activities of 0.3-200-5, 0.3-200-9, 0.3-200-11, and 0.3-200-13 samples by degradation of methylene blue (MB) under UV-light irradiating condition and compared with the standard TiO2 nanocrystal samples of P25 and ST-20. P25 is a commercially

available TiO2 nanoparticle sample containing 80% anatase phase and 20% rutile phase, in which half of anatase is [111]-faceted cubic nanocrystals and the other half of anatase is spherical nanocrystals without a specific facet on the surface.²⁶ ST-20 is an anatase nanocrystal sample with spherical nanocrystal morphology without a specific facet on the surface.²⁶ The photodegradation percentage of MB increases in an order of 0.3-200-13 <

ST-20 < 0.3-200-11 < P25 < 0.3-200-9 < 0.3-200-5 (Figure 2.13(a)). The increasing order of photocatalytic activity corresponds to the increasing order of specific surface area (Table 2.1), namely 0.3-200-13 < 0.3-200-11 < P25 < 0.3-200-9 < 0.3-200-5, except ST-20 without specific facet on the surface. It has been reported that the main influencing factors on the photocatalytic activity are surface area, crystallinity, and facet on the surface. In the [111]-faceted nanocrystal samples, the increasing order of the photocatalytic activity, 0.3-200-11 < P25 < 0.3-200-9 < 0.3-200-5, is consistent with the increasing order of surface area. Specifically, low photocatalytic activity of ST-20 can be contributed to that non-faceted surface.

Figure 2.13. Time-dependent photocatalytic degradation of MB under UV-light irradiating condition by 0.3-200-5, 0.3-200-9, 0.3-200-11, 0.3-200-13, P25 and ST-20 samples. (a) Degradation percentage per 20 mg of TiO2 sample and (b) degradation percentage per surface area of TiO2 sample.

Table 2.1. Exposed facet, crystal phase, bandgap, crystal size, and specific surface area for different TiO2 samples and cell performance parameters of DSSCs fabricated by different TiO2 samples

To normalize the specific surface area effect on the photocatalytic activity, we evaluated the MB degradation percentage per surface area of the TiO2 sample (MB degradation (%)/ (TiO2 specific surface area)), and the result is shown in Figure 2.13(b).

The surface photocatalytic activity increases in an order of ST-20 < 5 < 0.3-200-13 < P25 < 0.3-200-9 < 0.3-200-11. For the samples having similar crystal sizes of about 20 nm, the surface photocatalytic activity enhances in an order of ST-20 < P25 < 0.3-200-9, corresponding to increasing order of the fraction of [111]-facet on the surface. This result reveals the higher photocatalytic activity of [111]-faceted surface than that of the non-faceted surface.

For the samples synthesized by the hydrothermal treatment of TTIP-TMAOH suspension solution at different pH conditions, the surface photocatalytic activity increases in an order of 0.3-200-5 < 0.3-200-13 < 0.3-200-9 < 0.3-200-11. This order corresponds to the increasing order of crystal size, 10 nm 5) < 20 nm (0.3-200-9) < 50 nm (0.3-200-11), except non-faceted 0.3-200-13 (400 nm). The enhancement of surface photocatalytic activity is due to the enhanced crystallinity of anatase nanocrystal

and the fraction of [111]-faceted nanocrystals. Non-faceted 0.3-200-13 exhibits a higher surface photocatalytic activity than that of the [111]-faceted 0.3-200-5 which can be contributed from much higher crystallinity of 0.3-200-13 than that of 0.3-200-5, and also the low fraction of [111]-faceted nanocrystals in 0.3-200-5. Namely, higher crystallinity exhibits higher photocatalytic activity because the crystal defects can act as the charge recombination center of photogenerated electron-hole pair in the photocatalytic reaction.³

To study the electronic band structures, we measured the UV-visible spectra of TiO2

nanocrystals, seen in 2.14A. The bandgap energy of TiO2 can be estimated from the transformed Kubelka−Munk function (A･hν/B)^1/2= hν – Eg as shown in Figure 2.14B, where A, B, and Eg are the absorption coefficient, absorption constant, and bandgap energy, respectively.⁵⁴ The bandgap energy results are listed in Table 1. The bandgap energy increases in the order of 0.3-200-13 < ST-20 < 0.3-200-11 < P25 < 0.3-200-9 <

0.3-200-5. It is well known that the bandgap energy is dependent on crystal size for the nanocrystals, that is, the quantum size effect or blue shift. To give a relationship between the bandgap energy and crystal size for different faceted anatase nanocrystals, we plot the bandgap against the BET surface area for the anatase nanocrystal samples in Figure 2.15.

The bandgap energies of {010}-faceted anatase nanocrystals reported in our previous study are also presented in Figure 2.15.²⁶

Figure 2.14. (A) UV−visible absorption spectra and (B) Corresponding plots of transformed Kubelka−Munk function versus the energy of photon for 5, 9, 11, 0.3-200-13, P25, and ST-20 samples

The result reveals that the [111]-faceted nanocrystals exhibit the higher bandgap energy and a larger crystal size-dependent bandgap energy or larger blue shift than that of non-faceted nanocrystals of ST-20 and 0.3-200-13. The bandgap energy of P25 is close to the [111]-faceted nanocrystals because it contains 40% of the [111]-faceted anatase nanocrystals. The {010}-faceted anatase nanocrystals exhibit a further higher bandgap energy and a larger blue shift than that of the [111]-faceted nanocrystals. Our previous studies have revealed that the photocatalytic activity of anatase nanocrystals enhances in an order of non-facet < [111]-facet < {010}-facet.^28-29 The above results reveal that the crystal-facet can affect the bandgap energy and blue shift of the anatase nanocrystals. We think higher bandgap energy of the [111]-faceted nanocrystals is a reason why it shows higher surface photocatalytic activity than that of the non-faceted nanocrystals. The higher bandgap energy causes higher energy levels of the lowest conduction band that can generate more strongly reductive electrons for the photocatalytic reduction reaction.²¹ The higher bandgap energy of the {010}-faceted anatase nanocrystals also gives a higher surface photocatalytic activity than that of the [111]-faceted anatase nanocrystals.²⁶

Therefore, we can conclude that the surface photocatalytic activity enhances in an order of non-facet < [111]-facet < {010}-facet, corresponding to the increasing order of the bandgap energy.

Figure 2.15. Relationships between bandgap energy and BET surface area for TiO2 nanocrystal samples

It has been documented that the quantum size effect^55-60 and crystal structure^60-61 can affect the bandgap blue shift of TiO2 nanocrystals. A clear bandgap blue shift is observed in a crystal size range of < 2 nm by quantum size effect for anatase nanocrystals, for example anatase nanoparticles produce an obvious blue shift of approximately 1 eV with variation of crystal size from 2 nm to 1nm.⁶⁰ However, no obvious quantum size effect was found for anatase TiO2 nanocrystals when the crystal size surpasses 2 nm.^55-59 In

Figure 2.15, we can see no obvious bandgap blue shift for the non-faceted anatase TiO2

nanocrystals when the crystal size changes from 20 nm (3.01 eV) to 400 nm (3.00 eV).

However, both of the {010}-faceted and [111]-faceted nanocrystals show the larger bandgap blue shifts, where the bandgap energies enhance linearly with increasing SBET

by slops of 2.68 × 10^-3 and 2.17 × 10^-3, respectively, in the crystal size range of 10 to 150 nm, seen in Table 2.2. Therefore, we can introduce a new factor, namely, the facet effect on the blue shift for TiO2 nanocrystals on the basis of the above results.

Table 2.2. Facet on nanocrystal surface, bandgap, and crystal size for different anatase TiO2 nanocrystal samples.