**6. Formation of O-deficient anatase thin film**

To clarify the factors for designing an anatase thin film with a higher photoreactivity under UV irradiation, the relationship between the photoreactivity and O deficiency of anatase thin films fabricated with the heat-treated precursor films under regulated conditions was examined. Thin films were formed by heat-treating precursor films spin-coated onto FTO glass substrates with **SED** and **SSG** (sol–gel solution) under Ar or air.

The spin-coating method at ambient temperature was used for forming precursor films using a double-step method. The first step used 500 rpm for 5 s and the next step was 2000 rpm for 30 s, in all cases. The precursor films were pre-heated in a drying oven at 70 °C for 10 min and then heat-treated at 500°C for 30 min in a 0.1 L min–1 Ar gas flow. A tubular furnace made of quartz was employed for the heat-treatment. Thin films, **ED** and **SG**, were formed by applying the precursor solutions **SED** and **SSG,** respectively, before annealing in air. The film **EDair** was fabricated by firing the precursor film spin-coated with **SED** in air at 500°C for 30 min.

When the concentration of titanium was 0.4 mmol g–1 for **SED**, the film thickness was 100 nm. An **SSG** of 0.5 mmol g–1 was stirred for 3 days at ambient temperature to fabricate an anatase film of thickness 100 nm. The post-annealing treatment for the **ED**, **EDair,** and **SG** thin films was carried out in air at 500°C for 5, 10, 15, 20, and 30 min. The number in the notation used for post-annealed films indicates the annealing time (min). For example, **ED-PA5** indicates an **ED** film post-annealed for 5 min. The photoreactivities of the thin films are presented in Table 2. Each value was calculated as the difference between the decoloration rate under UV-light irradiation and the corresponding value measured for each thin film in the dark.


The maximum photoreactivity of **ED-PA15** produced by the MPM is twice that of **SG-PA10** prepared by a conventional sol–gel procedure.

306 Heat Treatment – Conventional and Novel Applications

The Vis-responsive property of the **OX–ED–OX** film was mainly due to the colored materials that were formed spontaneously during heat-treatment by chemical reactions between the reductant derived from the precursor complex containing OX in the upper layer and the organic residues derived from EDTA ligands in the lower one. Thus, the thermal reactions between the residues derived from the ligand of the precursor complex can afford novel functions such as the Vis-responsive nature of the resultant thin films through heat-treatment of the thin films fabricated by the MPM. The design of metal

**OX-OX-OX** 16(1) ─ **OX-ED-OX** 21(1) 6(1)

photoreaction with both thin films under UV and VIS irradiation [5]. The rate was measured by the decrease of absorption value at 664 nm of each test solution. Those obtained from the data measured

To clarify the factors for designing an anatase thin film with a higher photoreactivity under UV irradiation, the relationship between the photoreactivity and O deficiency of anatase thin films fabricated with the heat-treated precursor films under regulated conditions was examined. Thin films were formed by heat-treating precursor films spin-coated onto FTO

The spin-coating method at ambient temperature was used for forming precursor films using a double-step method. The first step used 500 rpm for 5 s and the next step was 2000 rpm for 30 s, in all cases. The precursor films were pre-heated in a drying oven at 70 °C for 10 min and then heat-treated at 500°C for 30 min in a 0.1 L min–1 Ar gas flow. A tubular furnace made of quartz was employed for the heat-treatment. Thin films, **ED** and **SG**, were formed by applying the precursor solutions **SED** and **SSG,** respectively, before annealing in air. The film **EDair** was fabricated by firing the precursor film spin-coated with **SED** in air at

When the concentration of titanium was 0.4 mmol g–1 for **SED**, the film thickness was 100 nm. An **SSG** of 0.5 mmol g–1 was stirred for 3 days at ambient temperature to fabricate an anatase film of thickness 100 nm. The post-annealing treatment for the **ED**, **EDair,** and **SG** thin films was carried out in air at 500°C for 5, 10, 15, 20, and 30 min. The number in the notation used for post-annealed films indicates the annealing time (min). For example, **ED-PA5** indicates an **ED** film post-annealed for 5 min. The photoreactivities of the thin films are presented in Table 2. Each value was calculated as the difference between the decoloration rate under UV-light irradiation and the corresponding value measured for each thin film in the dark.

**Table 1.** The rate *ν* [nmol L–1 min–1] of decoloration rate of 0.01 mol L–1 MB solution by the

under dark are also indicated. Calculated standard deviations are presented in parentheses.

under UV irradiation under VIS irradiation

complexes for the precursor and of the heating program are crucial.

**6. Formation of O-deficient anatase thin film** 

500°C for 30 min.

glass substrates with **SED** and **SSG** (sol–gel solution) under Ar or air.

Multilayer film *ν* [nmol L-1 min-1]

**Table 2.** The rate *ν* [nmol L–1 min–1] of decoloration rate of 0.01 mol L–1 MB solution by the photoreaction with each thin film under UV-light irradiation [6]. The rate was measured by the decrease of absorption value at 664 nm of each test solution. Those obtained from the data measured under dark are also indicated. Calculated standard deviations are presented in parentheses.

It is generally accepted that the main factors to consider when designing enhanced photoreactivity of anatase are (1) higher crystallinity, (2) larger surface area, and (3) decreased impurities. The crystallite size is an indicator of crystallinity [78, 79]. Among the crystallite sizes of the three anatase thin films, **ED**, **EDair** and **SG**, the **SG** thin film had the largest value and the **ED** film had the smallest (Table 3). These values for the anatase crystallites in **EDair** and **SG** thin films were not affected by post-annealing treatment in air. The thin film **ED-PA15** (whose crystallite size was the smallest) showed the highest photoreactivity in the decoloration of an MB aqueous solution among the various thin films formed in this study. The specific surface areas of the thin films were not measured quantitatively because of the difficulties involved. However, the degrees of adsorption of MB molecules in aqueous solution were nearly equal among the thin films, including those formed by the sol–gel method. Therefore, the differences in the photoreactivity among these thin films should be due to other factors than the specific surface area. The XPS spectra suggested that the thin films **SG** and **SG-PA***n* have higher purities than the other thin films. Therefore, the highest photoreactivity, of **ED-PA15** thin film, cannot be due to its purity. The O/Ti peak area ratio determined from the XPS of the anatase film **ED-PA15** with the highest photoreactivity was extremely small, 1.5. The refractive indices of the thin films **EDair** and **SG** increased gradually, depending on the post-annealing time. On the other hand, the refractive index of the **ED** thin film decreased with post-annealing treatment time up to 15 min and then increased with further annealing. The largest index (2.17; **ED** thin film) may be related to the strong and wide absorbance by the above-mentioned impurities. Furthermore, the smallest value (1.99; **ED-PA15)** could be affected by the largest O deficiency in the anatase thin film after purification. The decrease in permittivity of the thin film arose from the lower charge density derived from the O deficiency because the structure of the anatase lattice is rigid. Thus, the O deficiency formed by this method was one of the most important factors for fabricating highly UV-sensitive anatase. This O deficiency was formed during heat-treatment of the precursor metal complexes.

A coordination skeleton of (TiO4N2) or (TiO5N2) can be assumed in the EDTA complex as a precursor molecule from the structural study of a Ti complex [Ti(H2O)(EDTA)]·1.5H2O reported by Fackler *et al.* [80]. In the precursor films, two N and at least four O atoms link to one Ti ion. As a result of heat-treating the precursor complex in an Ar gas flow, neighboring complexes reacted with each other. In this process, several O atoms linked to one Ti ion could be covalently bonded by other Ti ions, and the anatase lattice was gradually created. By eliminating large amounts of C, H, and N atoms with O atoms, oxide ion sites of the anatase lattice were partially occupied by a rather stable nitride ion derived from the coordinated N atom originally belonging to the ligand. As a result, the total negative charge of the N-substituted anatase in the **ED** thin film is ca. 3.6 toward one Ti ion. This value is the summation of 2.8 from the oxide ions and 0.8 from the nitride ions. This charge toward one Ti ion is larger than that of ca. 3.3 by the oxide ions in the **SG** thin film. The substitutional N atoms could be removed from the anatase lattice by post-annealing the **ED** thin film. Consequently, the total negative charge of the **ED-PA15** thin film, whose photoreactivity is the highest, decreased to ca. 3.0. Longer annealing treatment replenished oxide ions in the anatase thin films from their surfaces and the photoreactivity decreased (Figure 8) [6]. Thus, it was elucidated that O deficiency is an important factor to consider when designing anatase photoreactivity. It is also notable that the O-deficient anatase lattice is rather robust because the stoichiometric Ti2O3 did not appear at all.

Heat Treatment in Molecular Precursor Method for Fabricating Metal Oxide Thin Films 309

Rutile is the most stable crystal form of titania. Since Nishimoto *et al.* showed that anatase is more sensitive to UV light than rutile in photoreactions, rutile was believed to be inferior to anatase in terms of photoreactivity [81]. Anatase is important for photocatalysis in pollutant degradation and in the development of photofunctional materials such as films with hydrophilic surfaces under UV-light irradiation. The poor photoreactivity and photosensitivity of rutile is generally believed to be due to its crystal structure. Rutile is primarily known as a useful pigment for white paint, due to its chemical stability [82, 83].

Because the band edge of a rutile single-crystal is 3.0 eV, rutile has the potential to respond to Vis light. Using this knowledge and the results of previous experiments on anatase responses to Vis light, this section describes an attempt to achieve direct fabrication of Odeficient rutile thin films with high photoreactivity using a MPM. The first Vis-lightresponsive thin film created from O-deficient rutile is discussed here. This material works without application of an electric potential, due to its unprecedentedly high photosensitivity under UV-light irradiation. The present findings should facilitate widespread practical use

The thin films were formed by heat-treating the precursor films after spin-coating onto a quartz glass substrate. **SED** and **SSG** were applied in an Ar gas flow. The transparent precursor films formed by spin-coating the solutions and pre-heating in a drying oven at 70 °C for 10 min were heat-treated at 700 °C for 30 min in a furnace made from a quartz tube with an Ar gas flow rate of 0.1 L min–1. When **SED** was used, a transparent rutile thin film **R** was formed. When **SSG** was used, a transparent anatase thin film **A** was formed. The film

Each structure was characterized using XRD, Raman spectroscopy, and transmission electron microscopy (TEM). The selectivity was due to the O-vacant sites in the oxide thin films formed at different levels due to the differences between the amounts of oxygen in the two precursors. In this case, the oxygen source required to structure titania was available only in the precursor films when these thin films were fabricated. Therefore, crystallization into rutile, which has many O-vacant sites, and the accompanying rapid elimination of

In contrast, the amount of oxygen available to Ti4+ in titanoxane polymers, though significant, was insufficient to develop stoichiometric TiO2 from **A**. The oxygen defects in an anatase lattice generally lower the temperature of the phase transformation from anatase to rutile [84, 85]. Thus, selective formation occurred according to the differing degrees of O

The photoreactivities of the thin films were evaluated by the decoloration rates of MB solutions, which served as a model for organic pollutants in water. The results measured under Vis- and UV-light irradiation are summarized in Table 4, along with those measured under dark conditions (reference values). The data show the effects of adsorption on the samples, vessels, and self-decoloration of MB under each condition. Moreover, the

organic residues from the **R** precursor film, occurred because of the heat-treatment.

**7. O deficiency in rutile thin film** 

of rutile in light-related applications.

thickness was 100 nm in both cases.

deficiency.


**Table 3.** The crystallite size of anatase in **ED**, **EDair**, **SG** and post-annealed thin films [6]. The crystallite size of anatase was measured with a typical Scherrer-Hall method by employing a peak assignable to only (1 0 1) of anatase, because other peak intensities due to anatase were too low to measure accurately. The crystallite size of anatase in **ED** and **ED-PA5** could not be obtained because the (1 0 1) peak of anatase was also too weak to determine the crystallite size.

**Figure 8.** Plausible route of the O-deficient anatase lattice formation from the precursor complex skeleton through the heat treatment in an Ar gas flow and the sequential post-anneal [6].
