**8. Crystal orientation and photoluminescence of rutile in thin film**

PL emission has been widely used to investigate the efficiency of charge carrier trapping, migration, and transfer, and to understand the fate of electron–hole pairs in semiconductor particles [86]. It is therefore helpful to examine the position and intensity of the PL bands of semiconductor particles to understand the photoreactivities of the particles [87]. In this section, we report the changes in the PL and photoreactive properties of the Vis-responsive rutile thin film fabricated by the MPM, which are effected by annealing in air at 700°C. The relationship between O deficiency and PL emission was examined to understand the incredibly high photoreactivity of the rutile thin film. Furthermore, the level of crystal orientation of the rutile thin film was quantitatively evaluated on the basis of data from XRD analyses. The amount of oxygen supplied during the annealing process was analyzed by XPS measurements. The growth of crystals and particles was also investigated by crystallitesize measurements and SEM observations. The heat-treatment of the fabricated O-deficient rutile **R** thin film was carried out in air at 700°C for 15, 30, and 60 min. The number in the notation of the post-annealed films indicates the annealing time (min). For example, **R-PA15** indicates that post-annealing treatment of the **R** thin film was carried out for 15 min. The extent orientation factor (*f*) for the (110) plane of the **R-PA***n* thin films increased with annealing time in air (Table 5).

310 Heat Treatment – Conventional and Novel Applications

Notation

standard deviations are presented in parentheses.

was 0.8 mW cm-2 [7].

photoreactivity of **A.** This is without precedent.

photoreactivity of **R** was extremely high under UV irradiation and higher than the

**R** 9(1) 23(1) **A** ─ 15(1) **Table 4.** The reaction rate *ν* of the decoloration reaction in an aqueous solution containing 0.01 mol L–1 of methylene blue under visible- and UV-light irradiation and under dark conditions [7]. Calculated

The photosensitivities of **R** and **A** were also examined by measuring the effects of Vis and UV irradiation on the water contact angle for the surfaces of the thin films. The results are shown in Figure 9 [7]. The rutile thin film **R** exhibited Vis-light-induced hydrophilicity with a fluorescent light, even though high-energy light below 400 nm was eliminated. In contrast, Vis light alone did not reduce the contact angle on **A** under the same conditions. Furthermore, a rapid decrease in the water contact angle for **R** was observed with weak UVlight irradiation. The super-hydrophilic property of **R** appeared after only 1 h. When fluorescent light with a UV component was employed, the contact angle on **R** rapidly reduced and the values reached 38° and 10° after irradiation for 1 and 24 h respectively.

**Figure 9.** Comparison of the contact angles of a 1.0-μL water droplet on the thin films **R** and **A**. Before the measurement, the thin films were exposed to UV irradiation of 1.3 mW cm–2 at 356 nm obtained by a black light (each on the left side), and to visible light without UV light was obtained from a fluorescent light by removing light of wavelengths shorter than 400 nm using a cut-off filter. The Vis light intensity after removing UV components from the fluorescent light, which was estimated by an illuminometer

It is noteworthy that the simple fabrication of a Vis-responsive rutile film with high photoreactivity could be attained. Thus, the O defects in titania are also effective at providing photoreactivity of rutile, which is usually insensitive to both UV and Vis light.

PL emission has been widely used to investigate the efficiency of charge carrier trapping, migration, and transfer, and to understand the fate of electron–hole pairs in semiconductor particles [86]. It is therefore helpful to examine the position and intensity of the PL bands of

**8. Crystal orientation and photoluminescence of rutile in thin film** 

*ν* [nmol L-1 min-1] under visible light under UV light


**Table 5.** The crystallite size, orientation factor and O/Ti ratio of rutile crystals in the **R** thin film and in the post-annealed **R-PAn** thin films [8]. The crystallite size of rutile was measured with a typical Scherrer-Hall method by employing the peaks assignable to (1 1 0), (1 0 1) and (2 1 1) of rutile. The extent of the orientation was estimated in terms of Logtering orientation factor, *f*, from the XRD peak intensities (*I*). For calculating the orientation factor, the intensity data of non-oriented rutile was cited from the JCPDS card 21-1276. The O/Ti ratios determined by the XPS peak areas of O 1s and Ti 2p3/2 peaks observed from the surfaces of **R** and post-annealed thin films. The XPS peaks of the thin film surface were measured without bombarding Ar+ ion beam. The peak area of O 1s and Ti 2p was calculated by FWHM and peak height at the positions 531.0 and 459.0 eV, respectively. The averaged O/Ti ratios determined by the XPS peak areas of O 1s and Ti 2p3/2 peaks of **R** and post-annealed thin films. The XPS peaks of thin films were measured after bombarding Ar+ ion beam with 2 kV and 18 μA cm–2 for 3 min, in order to remove surface oxides. The peak area of O 1s and Ti 2p was calculated by FWHM and peak height at the positions 531.0 and 459.0 eV, respectively, obtained from each depth profile in Ar+ ion etching mode. a) The estimated standard deviations are presented in parentheses

The extent orientation could be estimated from the XRD peak intensity by using the Lotgering method [8]. The terms *I*(*hkl*)ideal and Σ*I*(*hkl*)ideal are defined as the intensity of the peak attributable to the specific plane (*hkl*) and the sum of each intensity obtained for the

non-oriented rutile crystals; thus, *P*0 can then be expressed as

$$P\_0 = I(hkl)\_{\text{ideal}} / \sum I(hkl)\_{\text{ideal}} \tag{1}$$

In this study, each *I*(*hkl*)ideal value was cited from the standard data of the corresponding rutile phase.

The definitions of the terms *I*(*hkl*)obsd and Σ*I*(*hkl*)obsd for the thin films **R** and **R-PA***n* are identical to those for Σ*I*(*hkl*)ideal and *I*(*hkl*)ideal in equation (1), respectively, and the *Pn* value can be calculated as

$$P\_n = I(\text{lvl}I)\_{\text{cbsd}} / \Sigma I(\text{lvl}I)\_{\text{cbsd}} \tag{2}$$

Heat Treatment in Molecular Precursor Method for Fabricating Metal Oxide Thin Films 313

**Figure 10.** The photoluminescence emission spectra of the thin films, O-deficient rutile **R** and post annealed **R-PA***n* (*n* = 15, 30, 60). The spectra were measured in the wavelength range 190-850 nm at

**Wavelength/nm**

**200 400 600 800**

**Surface**

**(110) plane**

**Substrate**

**<sup>a</sup> a R-PA60**

**a a**

**a = 4.56 nm c = 2.93 nm**

**c**

**Figure 11.** Proposed schematic diagrams for the (110) plane orientated O-deficient rutile **R** (left) and post annealed **R-PA60** (right) thin films are shown. Cell parameters were refined by a least square

**; Titanium ; Oxygen ; Oxygen deficiency**

room temperature [8].

**PL intensity/a.u.**

**c** 

**R**

**R-PA15**

**R-PA30**

**R-PA60**

method [8].

**R**

**a = 4.58 nm c = 2.97 nm**

The Lotgering orientation factor *f* is defined as

$$f = (P\_n - Pv)/(1 - Po)\tag{3}$$

The factor *f* is defined for *P* or *P*0 values over a certain range of 2θ. As the level of orientation increases, the *f* value will increase from 0 to 1. The factor *f* is therefore a measure of the crystal phase orientation. Consequently, the *f* value of the **R** thin film was the smallest among the thin films fabricated by the MPM. However, the present *f* value of the **R** thin film was larger than those reported for rutile thin films fabricated by a sol–gel method on several substrates such as quartz glass, alumina, and single-crystals of quartz or silicon [88]. This result may be related to the different mechanisms governing the formation of the rutile lattice. In the sol–gel method, the Ti4+ and O2– ions that are tightly linked in titanoxane polymers formed by the condensation of partially hydrolyzed alkoxide might be rearranged to form the rutile lattice at a higher temperature. In the MPM, however, small units composed of Ti4+ and O2– ions with higher mobility could be formed when the organics were decomposed and removed by heat-treating the precursor complex. Therefore, a rutile thin film with a higher level of crystal orientation could be formed at these lower temperatures.

After 15 min of heat-treatment in air, the crystallite size of the **R** thin film increased, while those of the **R-PA***n* thin films remained almost constant (Table 5). These results indicate that crystallite growth at a temperature of 700°C was completed by heat-treatment over a period of 30–45 min. The additional thermal energy was consumed mainly for the process of grain growth after crystallite growth because the grain size gradually increased with annealing time.

Previously, the peak position of the PL emission band obtained for rutile crystals was observed at ca. 450 nm [89]. However, PL emission bands of the **R** and **R-PA***n* thin films formed by the MPM were not detected at 450 nm (Figure 10) [8].

Nakamura *et al*. reported that in the case of a rutile single-crystal, the PL emission band attributable to the (110) plane can be observed at 810 nm [90]. Taking into account the high levels of crystal orientation with reference to the (110) plane in the **R** and **R-PA***n* thin films, the PL emission bands observed at around 800 nm in the spectra of the **R-PA***n* thin films can be attributed to rutile crystals oriented along the (110) plane in these thin films (Figure 11) [8].

For the O-deficient rutile thin film **R** with high photoreactivity, no PL emission was observed in the range 190–850 nm. Thus, as previously suggested, the O-defect sites on the rutile thin film may suppress recombination of the photoinduced electron–hole pairs by electron trapping. In contrast, the PL emission from rutile thin films after annealing in air may be due to the oxide ions that are supplied to the O-defect sites on the film surface, because they function as recombination centers. As a result, the lattice oxygens of titania, especially in rutile thin films, function as recombination centers for the photoinduced electron–hole pairs.

312 Heat Treatment – Conventional and Novel Applications

The Lotgering orientation factor *f* is defined as

formed by the MPM were not detected at 450 nm (Figure 10) [8].

Nakamura *et al*. reported that in the case of a rutile single-crystal, the PL emission band attributable to the (110) plane can be observed at 810 nm [90]. Taking into account the high levels of crystal orientation with reference to the (110) plane in the **R** and **R-PA***n* thin films, the PL emission bands observed at around 800 nm in the spectra of the **R-PA***n* thin films can be attributed to rutile crystals oriented along the (110) plane in these thin films (Figure 11) [8]. For the O-deficient rutile thin film **R** with high photoreactivity, no PL emission was observed in the range 190–850 nm. Thus, as previously suggested, the O-defect sites on the rutile thin film may suppress recombination of the photoinduced electron–hole pairs by electron trapping. In contrast, the PL emission from rutile thin films after annealing in air may be due to the oxide ions that are supplied to the O-defect sites on the film surface, because they function as recombination centers. As a result, the lattice oxygens of titania, especially in rutile thin films, function as recombination centers for the photoinduced electron–hole pairs.

can be calculated as

The definitions of the terms *I*(*hkl*)obsd and Σ*I*(*hkl*)obsd for the thin films **R** and **R-PA***n* are identical to those for Σ*I*(*hkl*)ideal and *I*(*hkl*)ideal in equation (1), respectively, and the *Pn* value

 *Pn* = *I*(*hkl*)obsd/Σ*I*(*hkl*)obsd (2)

 *f* = (*P*n – *P*0)/(1 – *P*0) (3) The factor *f* is defined for *P* or *P*0 values over a certain range of 2θ. As the level of orientation increases, the *f* value will increase from 0 to 1. The factor *f* is therefore a measure of the crystal phase orientation. Consequently, the *f* value of the **R** thin film was the smallest among the thin films fabricated by the MPM. However, the present *f* value of the **R** thin film was larger than those reported for rutile thin films fabricated by a sol–gel method on several substrates such as quartz glass, alumina, and single-crystals of quartz or silicon [88]. This result may be related to the different mechanisms governing the formation of the rutile lattice. In the sol–gel method, the Ti4+ and O2– ions that are tightly linked in titanoxane polymers formed by the condensation of partially hydrolyzed alkoxide might be rearranged to form the rutile lattice at a higher temperature. In the MPM, however, small units composed of Ti4+ and O2– ions with higher mobility could be formed when the organics were decomposed and removed by heat-treating the precursor complex. Therefore, a rutile thin film with a higher level of crystal orientation could be formed at these lower temperatures. After 15 min of heat-treatment in air, the crystallite size of the **R** thin film increased, while those of the **R-PA***n* thin films remained almost constant (Table 5). These results indicate that crystallite growth at a temperature of 700°C was completed by heat-treatment over a period of 30–45 min. The additional thermal energy was consumed mainly for the process of grain growth after crystallite growth because the grain size gradually increased with annealing time. Previously, the peak position of the PL emission band obtained for rutile crystals was observed at ca. 450 nm [89]. However, PL emission bands of the **R** and **R-PA***n* thin films

**Figure 10.** The photoluminescence emission spectra of the thin films, O-deficient rutile **R** and post annealed **R-PA***n* (*n* = 15, 30, 60). The spectra were measured in the wavelength range 190-850 nm at room temperature [8].

**Figure 11.** Proposed schematic diagrams for the (110) plane orientated O-deficient rutile **R** (left) and post annealed **R-PA60** (right) thin films are shown. Cell parameters were refined by a least square method [8].

The reduction of rutile surfaces by heated hydrogen activates the photoreactivities of these surfaces [36]. As a result, the formation of an oxygen deficiency provides photoreactivity. The present study revealed that the unprecedentedly high photoreactivity of the **R** thin film is suppressed by oxygen supply during the annealing process. However, the high levels of photoreactivity of these films could be maintained even after oxygen supply to the surface by the post-annealing treatment (Table 5). These results indicate that the enhanced photoreactivity is related not only to the surface but also to the inner part of the thin films as a result of an interparticle electron transfer (IPET) effect, as proposed in our previous paper.

Heat Treatment in Molecular Precursor Method for Fabricating Metal Oxide Thin Films 315

even higher volumetric fraction of Ag nanoparticles can be homogeneously mixed with the titania precursor solution. The electrical conductivity of the resultant films is largely dependent on the volumetric fraction, size, and connectivity of the Ag nanoparticles, and

Using the MPM, mixed precursor solutions for fabricating Ag-NP/TiO2 composite thin films could be easily prepared. As a result, Ag-NP/TiO2 composite thin films of Ag volumetric fractions from 0.03 to 0.68 were fabricated with heat-treatment of the mixed precursor films at 600°C in air. To obtain quantitative information about the effects of Ag nanoparticles on the electrical properties, the nanostructures of the films were examined by TEM. The TEM images films with *φ*Ag of 0.26, 0.30, and 0.55 are shown in Figures 12 (A), (B), and (C) respectively [9]. The presence and distribution of Ag nanoparticles (black dots) inside the TiO2 film can be clearly seen. The percolation threshold of Ag nanoparticles in the titania thin film was found to be *φ*Ag 0.30. It is near the percolation threshold, when Ag particles are still not totally connected, that increasing the *φ*Ag by adding a small amount of Ag nanoparticles helps to build the conductive network and reduce the resistivity of the composite. Therefore, the decrease in resistivity was attributed to a change in the Ag nanoparticles` size, shape, and center-to-center distance between the Ag nanoparticles. As the Ag volumetric fraction increased further from 0.27 to 0.55, the electrical resistivity decreased from 10−2 to 10−<sup>5</sup> Ω cm, respectively. At *φ*Ag 0.61 to 0.68, the resistivity increased from 10−5 to 10−<sup>3</sup> Ω cm due to the inevitable increase in resistivity caused by agglomeration of the Ag particles. This study shows that the MPM, which offers excellent miscibility of the silver and titania precursor solutions, is effective at overcoming the miscibility limitations of the conventional sol–gel method and is necessary for fabricating composite thin films with

**Figure 12.** TEM images of the Ag-NP/TiO2 composite thin films at Ag volumetric fractions, φAg, of (A)

The excellent miscibilities of the precursor complexes in the MPM overcame the limitations of the extremely low Ag volumetric fraction in the previous sol–gel process. Therefore, the percolation threshold for the electrical resistivity of the composite film could be examined for a wide range of Ag fractions. Heat-treatment plays an important role in the production of Ag nanoparticles by reducing Ag+ ions in the precursor film and

the homogeneity of the dispersed silver in the dielectric titania matrix [92-95].

large *φ*Ag values.

0.26, (B) 0.30, and (C) 0.55, respectively [9].
