tio2 Fe N

K.S. Rane^{a, ,} , R. Mhalsiker^a, S. Yin^b, T. Sato^b, Kuk Cho^c, E. Dunbar^c and Pratim Biswas^c

^aDepartment of Chemistry, Goa University, Panaji City, Goa 403206, India

^bInstitute of Multidisciplinary Research for Advanced Materials, Tohoku University, 2-1-1 Katachira, Aoba-ku, Sendai, Miyagi 980-8577, Japan

^cEnvironmental Engineering Science, School of Engineering and Applied Sciences, Washington University in St. Louis, One Brooking Drive, Campus Box 1180, St. Louis, MO 63130-4899, USA

Abstract

Nitrogen substituted yellow colored anatase TiO_2−xN_x and Fe–N co-doped Ti_1−yFe_yO_2−xN_x have been easily synthesized by novel hydrazine method. White anatase TiO_2−δ and N/Fe–N-doped samples are semiconducting and the presence of ESR signals at g 1.994–2.0025 supports the oxygen vacancy and g4.3 indicates Fe³⁺ in the lattice. TiO_2−xN_x has higher conductivity than TiO_2−x and Fe/Fe–N-doped anatase and the UV absorption edge of white TiO_2−x extends in the visible region in N, Fe and Fe–N co-doped TiO₂, which show, respectively, two band gaps at 3.25/2.63, 3.31/2.44 and 2.8/2.44 eV. An activation energy of 1.8 eV is observed in Arrhenius log resistivity vs. 1/T plots for all samples. All TiO₂ and Fe-doped TiO₂ show low 2-propanol photodegradation activity but have significant NO photodestruction capability, both in UV and visible regions, while standard Degussa P-25 is incapable in destroying NO in the visible region The mid-gap levels that these N and Fe–N-doped TiO₂ consist may cause this discrepancy in their photocatalytic activities.

Graphical abstract

DRS of TiO₂, N/Fe–N doped TiO₂.

Keywords: Hydrazine; Oxalate precursor; Oxalate-hydrazinate precursor; Anatase; Arrhenius plots; Electrical conductivity; N–TiO₂; XPS-N 1s peak; Photodegradation; ESR; Fe–N co-doped TiO₂

Article Outline

1.

Introduction

2.

Experimental

2.1. Synthesis

2.1.1. Preparation of titanium hydroxide, oxalate and their hydrazinates

2.1.2. Preparation of iron containing titanium oxalates and their hydrazinates

2.2. Characterization

2.3. Photodegradation of aqueous 2-propanol and gaseous NO_x

2.4. Thermal products of the hydroxide, oxalate and their hydrazinates: codes

3.

Results and discussions

3.1. Anatase phase: Fe-doped TiO₂

3.2. BET surface area, nitrogen content, ESCA and DRS: band gap

3.2.1. Nitrogen: O₂–N₂ analysis and XPS

3.2.2. Diffuse reflectance spectra (DRS)

3.3. DC electrical conductivity, ESR

3.3.1. Electron spin resonance (ESR)

3.3.2. Mechanism of conductivity

3.3.3. Impurity levels in the wide band gap TiO₂

3.3.4. Photocurrent

3.4. Photo-oxidation of 2-propanol and photodestruction of NO_x

4.

Conclusions

Acknowledgements

References

1. Introduction

Preparation strategies of spinel oxides such as MFe₂O₄ (M=Ni/Mn/Zn/Mg), perovskites, La (Sr) MO₃(M=Al/Co/Mn) and anatase TiO₂ and their effect on the magnetic, electric, catalytic and sensing properties are our main focus in the field of solid-state chemistry. Metal oxides are synthesized, in general, by the thermal decomposition of precursors such as metal hydroxides, carboxylates and their hydrazinates [1], [2], [3], [4], [5], [6], [7], [8], [9], [10] and [11]. In one such preparation it was observed [10] that a white pigment-grade anatase TiO₂ synthesized from the thermal decomposition of titanium hydroxide and oxalate produced yellow colored anatase when the precursors were modified by hydrazination. Both the white and yellow TiO₂ are semiconducting [12] indicating defects in the lattice. The yellow TiO₂ showed the ultraviolet (UV) absorption edge extending into the visible (Vis.) region in diffuse reflectance spectra (DRS), while the white pigment is transparent in the region. A survey of literature [13], [14], [15], [16], [17] and [18] revealed a large number of papers dealing with photocatalytic activity of yellow colored TiO₂ in the Vis. region in mineralizing water and air-borne organic pollutants. The yellow color and the Vis. light absorption are attributed to the change in the band structure due to N doping. The N doping either decreases the band gap by mixing of nitrogen 2p states with oxygen O 2p states on the top of the valence band or creates N-induced mid-gap level. Hence a defective TiO_2−xN_x may be formed when N-containing metal organic precursors are thermally heated. In our case, the titanium hydroxide and oxalate on thermal decomposition give oxygen-deficient semiconducting TiO_2−δ, the precursors on modification with hydrazine, N₂H₄, yield the N-doped TiO_2−xN_x or both oxygen-deficient and N-doped TiO₂ semiconductors [12]. It seems nitrogen is easily introduced in the lattice of anatase by the hydrazine method of synthesis. Hydrazine method of synthesis had been adopted in our earlier studies in the preparation of γ-Fe₂O₃ [3] and [5], NiZnFe₂O₄ [2], MgFe₂O₄ [7], MnZnFe₂O₄ [8] and perovskites such as LaAlO₃, La (Sr)AlO₃, La (Sr)MnO₃ [9]. Metal and mixed metal hydroxides and carboxylates modified with the hydrazination found to decompose at lower temperatures than those without such modifications. The titanium hydroxide and oxalate on hydrazination were also expected to decompose at lower temperatures as compared to unhydrazinated ones. The observed color change of the TiO₂ and introduction of nitrogen in the lattice of the oxide is interesting. This aspect of the introduction of nitrogen in the lattice of the TiO₂ is of interest, as spinel and perovskite oxides synthesized in our laboratories by the hydrazine method may also have in their lattice nitrogen which may have some effect on their electric and magnetic properties. In this paper, the synthesis of nitrogen-doped titanium dioxide (white and yellow) is explored to determine its photocatlytic properties. Degradation of aqueous 2-propanol and NO_x is determined, and compared to that of Degussa P-25. As both forms of the oxide (white and N doped yellow) are semiconducting and metal ions doping [19], [20] and [21] modify the electronic and photocatlaytic properties of TiO₂, effect of Fe doping by oxalate method and Fe–N co-doping by hydrazine method is investigated to see whether this improves photocatalytic activity. Electron spin resonance (ESR) was used to understand the defect nature in TiO₂.

2. Experimental

2.1. Synthesis

2.1.1. Preparation of titanium hydroxide, oxalate and their hydrazinates

The synthesis of titanium hydroxide, oxalate and their hydrazine modifications was carried out using methods standardized in our previous studies [10] and [11]. Titanium iso-propoxide in iso-propanol was added slowly into water, and the titanium hydroxide precipitate that formed was oven- and freeze-dried. For preparing titanium oxalate, commercial TiO₂ was fused with KOH in a Ni crucible and treated with concentrated HCl to precipitate out KCl. NH₃ was added to the filtrate to get titanium hydroxide, which was then separately treated with oxalic acid and ammonium oxalate. The hydrazinated titanium oxalate was prepared by adding the titanium hydroxide slurry to the well-stirred oxalic acid/hydrazine hydrate mixture in an inert atmosphere. Hydrazinated titanium hydroxide was synthesized by introducing titanium isopropoxide in isopropanol into a mixture of water and hydrazine under inert atmosphere. In another hydrazination process, the titanium hydroxide and titanium oxalate were spread over a Petri dish and kept over hydrazine hydrate, N₂H₅OH (99%), in a desiccator and the completion of the hydrazine uptake was monitored titrimetrically using KIO₃ as the titrant [22].

2.1.2. Preparation of iron containing titanium oxalates and their hydrazinates

In order to get 0.01, 0.05 and 0.1 at% of Fe-doped TiO₂, the titanium hydroxide that had been synthesized from commercial TiO₂, as described in Section 2.1.1, was dissolved in dilute HCl. A freshly prepared ferrous chloride in the requisite amount was then added and the mixture was then treated with ammonia to obtain iron-containing hydroxide. The precipitate was then separately treated with oxalic acid in air, and the mixture of oxalic acid and hydrazine hydrate in N₂ atmosphere to obtain the oxalate and hydrazinated oxalate precipitates, respectively. In another hydrazination process, the titanium oxalate containing Fe was spread over a Petri dish and kept over N₂H₅OH (99%) in a desiccator, and the completion of the hydrazine uptake was monitored titrimetrically using KIO₃ as a titrant [22].

2.2. Characterization

All samples were chemically analyzed for the quantitative presence of Ti⁴⁺, C₂O₄²⁻ and N₂H₄ by standard methods [22]. From the isothermal weight loss and infrared analysis (Shimadzu IR Prestige-21), chemical formulas were determined [10], [11] and [12]. Using thermal analyzer (NETZSCH DSC-DTA-TG STA 409PC), the thermal paths were identified to get an idea of total decomposition of the precursors. The decomposed products were identified from X-ray diffraction (XRD) studies on ITAL APD 2000 using CuK_α radiation (λ=1.5418 Å) and Ni filter. BET surface area measurements were made using SORPTOMETRIC model 1990 instrument and a home-built apparatus (3-point measurements). Scanning electron micrographs (SEM) were obtained on Hitachi S-4500. Transmission electron micrographs (TEM) were recorded with JEOL-JEM 100SX microscope, working at 100 kV accelerating voltage. Electron spin resonance (ESR) spectra were recorded on Varian E-line Century Series E-112 X-band ESR spectrometer using 100 kHz field modulator. tetracyanoethylene (TCNE) was used as a standard for “g” factor measurements. For electrical conductivity measurements, the oxide samples in the form of tablets of 8 mm diameter and 5 mm thickness were prepared by compressing with a hydraulic press at 10 ton of pressure. Both the sides of the pellet were pasted with silver paste and pressed between platinum electrodes and introduced into a home-built two-probe conductivity cell. The temperature variation of resistance was measured from room temperature to 600 °C using a Keithley Electrometer. Diffuse reflectance spectra (DRS) were recorded on Shimadzu UV 2450 UV–Visible Spectrometer in the wavelength range of 200–700 nm. The amount of nitrogen in the samples was determined using the oxygen–nitrogen analyzer [14] (HIROBA, EMGA, 2800). ESCA analysis was done at 298 K using MgK_α radiation with a V.G. Scientific ESCASCOPE photoelectron spectrometer. X-ray photoelectron spectroscopy measurements were carried out over a Kratas Axis Spectrometer at a vacuum of 3×10⁻⁹ Torr with non-monochromatic MgK_α radiation.

2.3. Photodegradation of aqueous 2-propanol and gaseous NO_x

For studying the oxidative degradation of 2-propanol, about 50 mg of the oxide sample was suspended in an aqueous solution in a quartz cell. Prior to UV irradiation, the suspension was stirred for 30 min in an oxygen atmosphere in the dark conditions. The sample was then irradiated at 295 K using UV light (λ>250 nm) from a 100 W high-pressure Hg lamp with continuous stirring in an oxygen atmosphere. At given interval of time, the aliquots were analyzed after filtering through Millipore to remove TiO₂ particles. The products were analyzed using gas chromatography.

Photodestruction of NO_x studies were carried out as described in Ref. [14] at >290, >410 and >510 nm. About 0.12 g of the oxide sample was placed in a quartz reactor cell of internal volume 373 cm³ and a 450 W high-pressure mercury lamp was used as the light source. The light wavelength was controlled by selecting various filters, i.e. Pyrex glass for cutting off the light of λ<290 nm, Kenko L41 Super Pro (W) filter for λ<400 nm and Fuji tri-acetyl cellulose filter for λ<510 nm. The light intensity at λ<290, 400 and 510 nm on the surface of the photocatalyst was identified as, 352, 337 and 243 μmol m⁻² s, respectively. For photocatalytic studies, a gas mixture of NO+Air+N₂ (for balancing the reaction mixture) was introduced into a constant flow reactor. The activity was determined by measuring the concentration of NO gas at the outlet of the reactor. For comparison, a photocatalytic reaction was also carried out using commercially available TiO₂ (Degussa P-25).

2.4. Thermal products of the hydroxide, oxalate and their hydrazinates: codes

Based on the thermal analysis and X-ray characterization [10], [11], [12] and [23] all the precursors mentioned in Section 2.1 were isothermally decomposed around 400 °C for 4 h and the oxide products were coded as follows: titanium oxalate (oxal); titanium oxalate hydrazinate (oxalhyd); titanium hydroxide, both oven and freeze dried (TipOD and TipFD) and their hydrazinates (TipODH and TipFDH); 0.01, 0.05 and 0.1 at% Fe-doped titanium oxalates and their hydrazinates: 0.01 Fe-anatase/oxal, 0.05 Fe-anatase/oxal, 0.1 Fe-anatase/oxal, Hyd_0.01Fe, Hyd_0.05Fe and Hyd_0.1Fe.

3. Results and discussions

3.1. Anatase phase: Fe-doped TiO₂

The titanium hydroxide, Ti(OH)₄·1.5H₂O and its hydrazinate, Ti(OH)₄·N₂H₄ and titanium oxalate, (NH₄)₂TiO(C₂O₄)₂·2H₂O and its hydrazinate, (N₂H₅)₂TiO(C₂O₄)₂ decompose completely at 400 °C [10], [11], [12] and [23]. The iron-doped sample precursors also decompose at the same temperature. In general, the hydrazine-modified samples decompose earlier than that of the unmodified ones, which may be due to explosive decomposition. The oxide products of the oxalate (oxal) and both oven-dried and freeze-dried hydroxides (TipOD/TipFD) are white, while the thermal product of titanium oxalate hydrazinate (oxalhyd) is yellow color. The d_hkl values that obtained by XRD (Fig. 1) match well with the ICDD Card [24] of anatase TiO₂. The thermal products of iron-doped oxalate and hydrazinate: 0.01Fe-anatase/oxal; 0.05Fe-anatase/oxal; 0.1Fe-anatase/oxal, Hyd_0.01Fe; Hyd_0.05Fe; Hyd_0.1Fe are brown in color and show anatase type phase (Fig. 1).

Fig. 1. X-ray diffraction pattern of anatase and Fe-doped anatase.

3.2. BET surface area, nitrogen content, ESCA and DRS: band gap

The BET surface areas of all the TiO₂ samples are found to be in the range of 42–107 m² g⁻¹. The white anatase obtained from the thermal decomposition of (NH₄)₂TiO(C₂O₄)₂.2H₂O showed a surface area of 65.9 m² g⁻¹, while the yellow oxide product of hydrazinated complex, (N₂H₅)₂TiO(C₂O₄)₂ indicated an enhanced value of 107.6 m² g⁻¹. Low temperature-explosive decomposition of the hydrazine complexes, in general, found in our studies [1], [2], [3], [4], [5], [6], [7], [8], [9], [10] and [11] lead to products with small particle size and large surface areas. A representative TEM of yellow TiO₂ and Fe–N co-doped, Hyd_0.01Fe, obtained by hydrazine method in Fig. 2 confirm a fine nature of particles in the nanometer-sized range, while the white anatase from titanium oxalate has agglomerated particles. SEM studies too reveal the presence of an agglomerated particles in the case of products of oxalate precursors.

Fig. 2. TEM of (a) yellow TiO₂ (b) white TiO₂ and (c) Fe–N co-doped TiO₂, Hyd_0.01Fe.

3.2.1. Nitrogen: O₂–N₂ analysis and XPS

The hydrazine that is released by low temperature decomposition of (N₂H₅)₂TiO(C₂O₄)₂ reacts with the atmospheric oxygen and liberates enormous [25]

(1)N₂H₄+O₂→N₂+H₂O, ΔH=−621 kJ mol⁻¹.

This energy is sufficient to exothermically decompose the complex devoid of hydrazine at much lower temperature than (NH₄)₂TiO(C₂O₄)₂·2H₂O, and the resultant particles thus formed are of fine nature. The nitrogen contents measured on all anatase and Fe-doped samples prepared by the hydrazine method showed the values in the range of 0.425–0.95%, and the highest value is observed for the yellow anatase obtained from titanium oxalate hydrazinate (oxalhyd), while the oxide products of the oxalate samples did not show any presence of nitrogen (Table 1). However, the 0.95% nitrogen content of yellow anatase reduced to 0.30% on calcining at 400 °C, the titanium oxalate hydrazinate in nitrogen atmosphere (oxalhyd 400 °C/N₂) and the similar but marginal decrease from 0.13% to 0.1% is observed for yellow TipODH. The hydrazine method of preparation introduces nitrogen in the lattice easily, as during decomposition of N₂H₄ that released reacts with atmospheric O₂ and produces the N₂ and H₂O (Eq. (1)). The nitrogen thus formed in situ easily gets trapped in the lattice of the oxide. Thus, the yellow anatase has the formula, TiO2_−xN_x and Fe–N-doped anatase may have the composition, Ti_1−yFe_yO_2−xN_x. Although the nitrogen content measured by oxygen–nitrogen analyzer indicates its presence in the yellow colored TiO₂ and Fe–N co-doped TiO₂, the ESCA study carried out on ESCASCOPE did not give clear indication of the presence of N 1s peak [18], [21], [26] and [27] at 396 eV. However, a peak centered 400.54 eV is observed in Kratas Axis Spectrometer, Fig. 3a for the yellow N-doped TiO₂ (oxalhyd) and 400.27 eV, Fig. 3a for Fe–N co-doped TiO₂ (Hyd_0.1Fe) which contain, respectively, 0.95% and 0.796% nitrogen (Table 1).

Nitrogen content, BET surface area, band gap and % of photodegradation of 2-propanol and NO of all the TiO₂, N/Fe/Fe–N-doped TiO₂

Fig. 3. XPS of N 1s, oxygen 1s and Ti 2p_3/2 core levels of yellow TiO_2−xN_x, oxalhyd (a,b,c) and 0.1 Fe–N co-doped TiO₂, Hyd_0.1Fe (a″, b″, c″).

Sakthivel and Kisch [13] and [28] observed no anionic-like nitrogen species around 396 eV, rather a N 1s peak at 404 eV attributed to hyponitrite type nitrogen. Valentin et al [29] observed N 1s core level at 400 eV and attributed it to a lower valence state of nitrogen. Gole et al. [16] and Chen and Burda [30] observed the N 1s level at 401-3 eV and considered it to be N–Ti–O linkage in the lattice. Thus, the presence of nitrogen in the lattice can be ascertained. Further, Satish et al. [31] from their observations of Ti 2p_3/2 level at 459.3 for TiO₂ and that for N–TiO₂ at 458.5 eV attributed the lower binding energy (BE) for N–TiO₂ to different electronic interaction of Ti with nitrogen anion compared to the oxygen anion. They suggested a considerable modification of the lattice due to N substitution. In our studies, we find the Ti 2p_3/2 level at 458.187 eV (Fig. 3c) for N-doped yellow TiO₂ (oxalhyd) and 458.194 eV (Fig. 3c for Fe–N co-doped TiO₂ (Hyd_0.1Fe). Although the other samples do not show a peak at 400 eV indicative for the presence of nitrogen, the Ti 2p_3/2 level found to be 458 eV suggests some modification in the lattice due to nitrogen substitution. The oxygen 1s peaks at 530.313 eV, for N-doped yellow TiO₂ (oxalhyd) and 530.256 eV for Fe–N co-doped TiO₂ (Hyd_0.1Fe) are shown, respectively, in Fig. 3b and Fig. 3b″.

3.2.2. Diffuse reflectance spectra (DRS)

The DRS of the TiO₂ samples are shown, Fig. 4. The absorption is at <400 nm for white TiO₂, while the yellow colored TiO₂ indicates the absorption edge extending into the visible region, >400 nm.

Fig. 4. DRS of TiO₂ and corresponding differential plots.

A band gap, E_g1, of 3.26 eV was observed for the white TiO₂ that was calculated from Eq. (2) [32]

(2)E_g=1239.8/λ.The wavelength λ of the absorption edge was taken from the peak position of the differential plots, which are also shown in the figure for ready reference. The yellow TiO₂, however, showed two peaks corresponding to band gaps: E_g1=3.26 and E_g2=2.63 eV, listed in Table 1. Two absorption bands at 400–408 nm (3.04–3.1 eV) and 530–550 nm (2.25–2.34 eV) are also observed [14] in the yellow powders obtained by planetary milling of the P-25 and hexamethylenetetramine (HMT). It is thought that the first and second edges are related to the band structure of the original titania and the newly formed N_2p band which is located above the O_2p valence band. The commercial white TiO₂, however, showed one band gap of 3.25 eV. All Fe-doped samples, Fig. 5, also show the absorption edge extending in the visible region and indicate two band gaps, Table 1.

Fig. 5. DRS of TiO₂, N/ Fe–N-doped TiO₂.

However, the first band appears at 3.31, 3.38 and 3.05 eV, respectively, for 0.01, 0.05 and 0.1 Fe-doped samples without nitrogen. On the other hand, the Fe-doped samples containing nitrogen, Hyd_0.01Fe, Hyd_0.05Fe, Hyd_0.1Fe have the first band at 2.8 eV which is much lower than 3.26 eV that observed for the yellow nitrogen-containing TiO₂. The second gap, however, is at 2.44 eV for all Fe-doped samples. These results suggest that iron doping modifies the first band gap of the white TiO₂ and introduces a second gap as in yellow N–TiO₂. But the second gap of N–TiO₂ is lowered by iron doping. The increase in the first band gap by Fe doping in TiO₂ from 3.26 to 3.38 eV and a decrease to 2.8 eV by co-doping of Fe and N suggest that there occur a lot of electronic structure changes due to doping. The study further suggests that a mid band gap is created. Anatase is a wide band gap (3.2 eV) insulator, and the changes in electronic structure must, therefore, reflect in their electrical characteristics. Hence the direct current electrical conductivity measurements were carried out on all samples to get some information on their semiconducting properties.

3.3. DC electrical conductivity, ESR

The temperature variation of direct current electrical conductivity of the samples (Fig. 6) shows a linear decrease in the conductivity with the increase in temperature, thereby, satisfying the Arrhenius equation, Eq. (3)

(3)σ=Ae^-E/kTwhere, σ is the conductivity; E the activation energy for producing free electrons or holes; T the absolute temperature, k the Boltzmann's constant; and A the constant. From the Arrhenius plot of log σ vs. 1/T, activation energy of 1.8 eV was calculated from the linear range. These studies indicate that the oxides are semiconducting. It can be observed from Fig. 6 that the yellow colored TiO_2−xN_x (TipODH) shows higher conductivity as compared to the white TiO₂ (oxal). The semiconducting white TiO₂ may have non-stoichiometric composition TiO_2−δ. All Fe and Fe–N co-doped anatase with composition, Ti_1−yFe_yO_2−xN_x, have higher resistance as compared to the yellow N-doped TiO₂.

Fig. 6. Temperature variation of electrical resistivity of TiO₂ and Fe-doped TiO₂.

An electronic semiconductivity is observed [33] in anatase and rutile due to free electrons as current carriers and an activation energy of 1.7 eV is found for rutile from an Arrhenius plot. From the variation in conductivity with the oxygen pressure, it is considered that there occurs an oxygen vacancy which contributes to electronic conductivity. TiO₂ is a reducible oxide and therefore the valence of the Ti can be changed from stable Ti⁴⁺ to the trivalent, Ti³⁺, which has a 3d¹ electronic configuration and can be identified [34] using electron paramagnetic resonance, EPR.

3.3.1. Electron spin resonance (ESR)

Sol–gel-derived TiO₂ showed one single first-derivative absorption line (line width of 0.6 mT) with g=1.998±0.004 and the sample on hydrogen reduction indicated a pronounced peak. Based on these observations, the authors confirmed the presence of Ti³⁺ in TiO₂. The ESR spectra of all TiO₂ and Fe-doped TiO₂ samples, Fig. 7, show a band at g=1.994–2.0025 with ΔH=170–210 G.

Fig. 7. ESR of TiO₂ and Fe-doped TiO₂.

The measurements were carried out both at room temperature (RT) and liquid nitrogen temperature (LT). There seems to be broadening of the band for the nitrogen-doped sample. ESR performed on nano-colloid TiO₂ showed a resonance at g2.0035 which is attributed to an oxygen hole center created near the surface that significantly increased with nitridation [35]. UV-irradiated TiO₂ show {>Ti^IV–OH} radicals [36] with g₁ of 2.016 and g₂ of 2.002 that arise from hole trapping, and in addition another peak with g value of 1.991, which is attributed to {>Ti^III–OH} generated from trapping of excited electron. Fe-doped TiO₂ samples in the present investigations show a band with g=1.994–2.0025, but they also have an additional peak at low field of g=4.19 with ΔH=125 G. The EPR signal of soil sample for Fe³⁺ is observed at g=4.2 [37]. 35TeO₂ (65−x)V₂O₅×Fe₂O₃ (mol%) glasses show [38] ESR line at g=4.3 due to Fe³⁺. As the mol% of Fe₂O₃ increases the area under g=4.3 increases and broadening of the hyperfine starts to take place. Fe-doped TiO₂ in the present studies, thus, indicate the Fe³⁺ in the sample.

3.3.2. Mechanism of conductivity

Decrease in conductivity and increase in activation energy E_a with the increase of Fe/Ti mole ratio [39] is considered to be due to the fact that Fe ions dissolve in TiO₂, which is an n-type semiconductor. These facts suggest that the valence control can be described by the following Kroeger–Vink notation,

Fe₂O₃+1/2O₂→2Fe′Ti+4O^*O+2 h.

Electrons as the carrier of the electrical conductance decreased because of the formation of the holes, so E_a increased. Thin film anatase with activation energy of 0.65 eV reduced to 0.51 eV in the presence of trace amount of Fe [40]. As Fe doping concentration increased, the n-type anatase transformed into rutile at 0.32 at% Fe. The Fe doping is not a simple substitutional one and new oxygen vacancies are induced by the presence of iron atoms. Fe doped and Fe–N co-doped TiO₂ show the conductivity, Fig. 6, lower than that of the yellow TiO₂ (TipODH). The preliminary studies on the direct current electrical conductivity measurements in our present investigations, thus, do give indication of change in conductivity characteristics in the TiO₂ due to the presence of oxygen deficiency, nitrogen and Fe in the lattice.

There seems to be a correlation between the ESR studies and the electrical conductivity observations made here. The observation of activation energy 1.8 eV from the Arrhenius plots and midgap band 2.4 eV from DRS studies suggest possible presence of impurity levels in Fe–N-doped samples that cause the red shift in DRS. The red shift has been now considered to be due to N doping as well as by metal doping.

3.3.3. Impurity levels in the wide band gap TiO₂

There are controversies on locating the impurity levels in the band gap. Valentin et al. [41] in their studies on the photoactivity of N-doped anatase and rutile considered the N 2p localized states just above the top of the O 2p valence and in anatase these states cause a red shift of the absorption edge towards visible region, while in rutile this effect is offset by the concomitant N-induced contraction of the O 2p band giving blue shift.

The doping of metal oxides with other elements is a well-recognized process in which the dopants produce electronic mid-gap states. Ultraviolet absorption in TiO₂ is due to its wide band gap of 3.2 eV and the band narrowing enables it to extend the absorption edge into the visible region as considered by Asahi et al. [21] for the N-doped TiO_2−xN_x due to the mixing of N- 2p and O- 2p orbitals. Ihara et al. [42] on the other hand considered not only nitrogen in the lattice but also the oxygen vacancies and electronic levels slightly below the conduction band edge were responsible for the visible light response. The ESR signal observed at g=2.003 supports the oxygen vacancy, as this signal is assigned to electron trapped at the oxygen defect site and it was observed that the signal gets strengthened when visible light illumination was applied. But, from their exhaustive photocurrent studies of non-doped and N-doped TiO₂, Nakamura et al. [27] gave a clear experimental evidence to the mechanism for visible light response by showing a N-induced mid gap level 0.75 eV above the valence band, which lies 2.45 eV below the conduction band. The UV radiation produces the holes in the O 2p valence band which is 3.20 eV below the conduction band (the normal band gap of TiO₂) and visible light illumination produces the holes in the mid gap level that is 2.45 eV below the conduction band. Such mid gap bands are observed by us in our DRS studies of N-doped yellow TiO₂ and Fe doped (0.01/0.05/0.1 Fe-TiO₂) as discussed above (Table 1). However, the Fe–N co-doped samples (Hyd_0.01Fe, Hyd_0.05Fe, Hyd_0.1Fe) do show the mid gap band at 2.40 eV, but there takes place a decrease in main band gap too from 3.2 band to 2.8 eV. Thus, there seems to be narrowing of the main band as well as introduction of the mid level gap in the Fe–N co-doped samples.

3.3.4. Photocurrent

The difference in the incident photon to current efficiency (IPCE) enhancement between UV and visible light illumination observed in the study of photocurrent measurements by Nakamura et al [27] to different reactivities of valence band and mid-gap level holes. Photocurrent measurements were thus considered to be important to get the insight in the mechanism of photocatalysis under UV and visible regions. The measurement of photocurrent during the reaction is of greater relevance than observing the products constitution in understanding the mechanism of photocatalysis involving the mid-gap and valence band holes. Such mid-gap states are indeed important in improving semiconducting properties of insulators, but in photocatalytic applications they are not desired, as these states trap the photogenerated holes and cause a decrease in the oxidation power of the catalyst. A reduced photocatalytic oxidation activity for 2-propanol and enhanced destruction activity for NO_x under visible irradiation as compared to the standard Degussa P-25 was observed. The results of such studies are discussed taking the clue of the presence of mid-gap level in the samples.

3.4. Photo-oxidation of 2-propanol and photodestruction of NO_x

The photodegradation of aqueous 2-propanol on all samples was carried out using a UV lamp (λ>245 nm) at 295 K. The irradiation time was 2 h and in some cases the percentage degradation was measured at different intervals: 20, 40, 60 and 90 min. The gas phase NO_x destruction study was done at >290, >410 and >510 nm. The results of such studies are shown in Table 1.

About 20% 2-propanol degradation is observed for yellow N-doped TiO₂ (oxalhyd and Tiphyd), while the white TiO₂ (oxal) showed 48% conversion. The standard Degussa P-25 sample, however, showed higher conversion of 58%. A low percentage photo-oxidation of 2-propanol observed [18] on TiO_2−xN_x with 0.005x<0.02 under visible light irradiation compared to UV irradiation is considered to be due to the formation of an N-induced narrow band slightly above the valence band. Similar low percentage degradation observations have been made by Nakamura et al. [27] from their photocurrent measurement studies on methanol, while from the analysis of the oxidized product of photocatalytic reactions of methylene blue [21], 2-propanol [18] and acetone [21] and [43] on N–TiO₂ it was observed that these reactions are possible under visible irradiation. Such discrepancies are due to no clear understanding of the mechanism of the reactions. Mrowetz et al. [44] fail to observe photocatalytic oxidation of HCOO⁻ into CO₂⁻, or of NH₃-OH⁺ into NO₃⁻ under visible illumination by N-doped TiO₂ and authors of the opinion that any photo-catalytic oxidation of aqueous organic contaminants by such catalysts is conjectural one. However, Nakamura et al. [27] conclude their observations by considering N-induced (occupied) mid-gap level slightly above the valence band that is responsible for the visible light response. The white TiO₂ (oxal) is semiconducting non-stoichiometric and has one optical (DRS) band gap, E_g1=3.26 eV, while the yellow N-doped TiO_2−xN_x shows two band gaps: E_g1=3.25 and E_g2=2.63 eV. And this presence of the mid-gap band may be the center for annihilating the charge carrier formed in the photoreaction causing the decrease in the degradation activity. The Fe-doped (0.01/0.05/0.1 Fe–TiO₂) and Fe–N co-doped (Hyd_0.01Fe; Hyd_0.05Fe; Hyd_0.1Fe-TiO₂) show less than 10% degradation, Table 1, thereby, indicating further decrease in the activity due to the presence of mid-gap band. The Fe–N co-doped TiO₂ may have this very low activity due to the mid-gap band at 2.42 eV, but they have their main band gap at a lower value of 2.8 eV, which may cause further annihilation of the charge carriers. On the other hand, all above samples have [45] and [46] significant NO destruction ability >290, >410 and >510 nm, Fig. 8, while the standard Degussa P-25 has no activity, in the visible region, >410 nm, Table 1.

Fig. 8. NO_x destruction capacitiy of TiO_2−δ and Fe/Fe–N-doped TiO₂ in UV and visible regions; (open triangle=oxide from oxalate and shaded triangle=oxide from hydrazine method).

The mid-gap bands seem to be advantageous in NO destruction activities; however, reasons as to why they are not so for the degradation of 2-propanol need to be addressed. UV light in large band gap substances such as TiO₂ produces electron–hole pairs. The light-induced hole in the valence band combines with the water molecule to form hydroxyl group with powerful oxidation ability to degrade 2-propanol, while the electrons in the conduction band in the presence of oxygen and traces of moisture in the air (of the reaction mixture consisting of NO+Air and N₂ for balance in the present investigation) are trapped to generate active OOH^* radical. The nitrogen monoxide reacts with the reactive oxygen radical to produce HNO₂ or HNO₃. The mid-gap level may allow visible light to generate electrons and hence this is advantageous in NO destruction in our present studies. Hole produced in mid-gap level may not react with water molecule to produce hydroxyl oxidant. However, further elaborative studies need to be carried out to explore any possible correlation between the activation energy that was obtained from conductivity measurements and the mid-gap bands that were observed from the DRS studies in understanding the photocatalytic activities of anatase by choosing the other dopants such as Nb, Pt and Cr. The influence of surface area and crystal size of anatase on doping needs to be studied to determine their influence on catalytic activity.

4. Conclusions

i) Nitrogen-substituted yellow colored anatase, TiO_2−xN_x and Fe–N co-doped TiO₂ have been readily synthesized by novel hydrazine method.

 

ii) White non-stoichiometric anatase form TiO_2−x, yellow TiO_2−xN_x, Fe-doped and Fe–N co-doped Ti_1−yFe_yO_2−xN_x are semiconducting and the ESR signals observed at g1.994–2.0025 supports the oxygen vacancy, as this signal is assigned to electron trapped at the oxygen defect site and g4.3 indicates the presence of Fe³⁺ in the lattice.

iii) UV absorption edge of anatase white TiO₂ extends to the visible region in yellow TiO_2−xN_x, Fe-doped and Fe–N co-doped TiO₂ and two band gaps are observed at 3.25/2.63 in N-TiO₂ and 3.31/2.44 and 2.8/2.44 eV, respectively, for Fe-doped and Fe–N co-doped TiO₂ suggesting possible mid-gap bands in the N, Fe and Fe–N doped anatase.

iv) The observation of activation energy 1.8 eV from the Arrhenius plots and mid-gap band 2.4 eV from DRS studies suggest possible presence of impurity levels in Fe–N-doped samples that cause the red shift in DRS.

v) All TiO₂ and Fe-doped TiO₂ show low 2-propanol photodegradation activity, but significant NO photodestruction capability in both UV and visible region, while the standard Degussa P-25 has no activity in the visible region. The midgap bands may cause the discrepancy in their catalytic activities.

Acknowledgments

This work is partly supported by UGC (major project of KSR), DST (major project of Dr. B.R. Srinivasan and K.S.R.), DST-FIST and UGC-Special Assistance Programs of the Department of Chemistry, Goa University. K.S.R. is grateful to Dr. Gundu Rao, RSIC, IIT-Bombay for ESR spectra, Prof. M. Anpo and Dr. Neppolian of Osaka Prefecture University, Japan for ESCA and photodegradation of propanol investigations and useful discussions and Dr. Bindu Varughese of University of Maryland for XPS data.

References

[1] K.S. Rane, A.K. Nikumbh and A.J. Mukhedkar, J. Mater. Sci. 16 (1981), p. 2387. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (20)

[2] V. Moye, K.S. Rane and V.N. Kamat Dalal, J. Mater. Sci. Mater. Electron. 1 (1990), p. 212. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (15)

[3] V. Borker, K.S. Rane and V.N. Kamat Dalal, J. Mater. Sci. Mater. Electron. 4 (1993), p. 241. View Record in Scopus | Cited By in Scopus (15)

[4] K.S. Rane, V.M.S. Verenkar and P.Y. Sawant, J. Mater. Sci. Mater. Electron. 10 (1999), p. 133. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (10)

[5] K.S. Rane and V.M.S. Verenkar, Bull. Mater. Sci. 24 (2001), p. 39. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (15)

[6] K.S. Rane, V.M.S. Verenkar and P.Y. Sawant, Bull. Mater. Sci. 24 (2001), p. 331. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (8)

[7] V.M.S. Verenkar, Beneficiation of Goan Ore Rejects to get pure iron oxide and utilization of the iron ore rejects to synthesize ferrites-high-tech magnetic materials, Ph.D. Thesis, Goa University, Goa, India, 1997.

[8] P. Y. Sawant, Physico-chemical methods to determine the trace elements in Goan Ore Rejects and beneficiate the ore to get pure iron oxides useful for high-tech ferrite manufacture, Ph.D. Thesis, Goa University, Goa, India, 1998.

[9] R. M. Pednekar, Synthesis and characterization of metal and mixed metal oxides of spinel and perovskite structure, Ph.D. Thesis, to Goa Univeristy, Goa, India, November 2005.

[10] R. Mhalsikar, K. S. Rane, A novel precursor technique to prepare yellow color TiO₂, in: Proceedings of the third National Symposium and Conference on Solid State Chemistry and Allied Areas, Indian Association of Solid State Chemists and Allied Scientists-ISCAS, IIT Delhi, India, p. 73 (P-30).

[11] K. S. Rane, Tailoring of electric, magnetic magnetic materials, in: Proceedings of the 19th Annual Symposium and Prof. Ram Chand Paul Second Symposium on Recent Trends in Chemistry, Punjab University, Chandigarh, India(Invited Lecture IL -4).

[12] R. Mhalsikar, R. Pednekar, K.S. Rane, Electrical characteristics of TiO₂ synthesized from different precursors, in: D. Bahadur, S. Vitta, O. Prakash (Eds.), Inorganic Materials: Recent Advances, Narosa Publishing House, New Delhi, India, Paper presented at the International Symposium in Inorganic Chemistry, IIT-Bombay, Mumbai, India, December 11–13, 2002, pp. 469–471.

[13] S. Sakthivel and H. Kish, Chemphyschem 4 (2003), p. 487. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (217)

[14] S. Yin, H. Yamaki, M. Komatsu, Q. Zhang, J. Wang, Q. Tang, F. Saito and T. Sato, J. Mater. Chem. 13 (2003), p. 2996. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (89)

[15] S. Yin, Q. Zhang, F. Saito and T. Sato, Chem. Lett. 32 (2003), p. 358. View Record in Scopus | Cited By in Scopus (94)

[16] J.L. Gole, J.D. Stout, C. Burda, Y. Lou and X. Chen, J. Phys. Chem. B 108 (2004), p. 1230. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (351)

[17] H. Noda, K. Oikawa, T. Ogata, K. Matsuki and H. Kamata, Chem. Soc. Jpn. 8 (1986), p. 1084 (TiCl₄+(NH₂)₂·H₂O yellow colr due to oxygen vacancy).

[18] H. Irie, Y. Watanabe and K. Hashimoto, J. Phys. Chem. B. 107 (2003), p. 5483. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (685)

[19] M. Anpo, Catal. Surv. Jpn. 1 (1997), p. 169. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (227)

[20] M. Anpo and M. Takeuchi, J. Catal. 216 (2003), p. 505 (and references therein). Article | PDF (401 K) | View Record in Scopus | Cited By in Scopus (470)

[21] R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki and Y. Taga, Science 293 (2001), p. 269. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (2616)

[22] I.A. Vogel, A Text Book of Quantitative Inorganic Analysis, Longman, UK (1978).

[23] K. S. Rane, W. Weisweiler, H. Langbein, M. Crocoll, F. Kennfack, R. Pednekar, in: CGS Pillai, KL Ramkumar, PV Ravindran, V Venugopal, (Eds.), Proceedings of the 13th National symposium on Thermal Analysis-Thermans-2002, Indian Thermal Analysis Society, Mumbai, India, p. 145.

[24] ICDD Card 02-0387.

[25] E.W. Schmidt, Hydrazine and its Derivatives, Wiley, New York (1984).

[26] O. Diwald, T.L. Thopson, G. Goralski, S.D. Walck and J.T. Yates Jr., J. Phys. Chem. B 108 (2004), p. 52. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (143)

[27] R. Nakamura, T. Tanaka and Y. Nakato, J. Phys. Chem. B 108 (2004), p. 10617. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (247)

[28] S. Sakthivel, M. Janczarek and H. Kisch, J. Phys. Chem. B 108 (2004), p. 9384.

[29] C.D. Valentin, G. Pacchioni, A. Selloni, S. Livraghi and E. Giamello, J. Phys. Chem. B 109 (2005), p. 11414.

[30] X. Chen and C. Burda, J. Phys. Chem. B 108 (2004), p. 15446. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (199)

[31] M. Satish, B. Vishwanathan, R.P. Vishwanath and C.S. Gopinath, Chem. Mater. 17 (2005), p. 6349.

[32] B. Oregan and M. Gratzel, Nature 353 (1991), p. 737. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (1163)

[33] M.D. Earle, Phys. Rev. 61 (1942), p. 56. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (23)

[34] E. Munoz, J.L. Boldu, E. Andrade, O. Navaro, X. Bokhimi, T. Lopez and R. Gomez, J. Am. Ceram. Soc. 84 (2001), p. 392. View Record in Scopus | Cited By in Scopus (17)

[35] S.M. Prokes, J.L. Gole, X. Chen, C. Burda and W.E. Carlos, Adv. Funct. Mater. 15 (2005), p. 161. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (58)

[36] T.K. Kim, M.N. Lee, S.H. Lee, Y.C. Park, C.K. Jung and J.-H. Boo, Thin Solid Films 475 (2005), p. 171. Article | PDF (664 K) Article | PDF (157 K) | Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (40)

[37] B. Sutter, T. Wasowicz, T. Howard, L.R. Hossner and D.W. Ming, Soil Sci. Soc. Am. J. 66 (2002), p. 1359. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (10)

[38] R. Singh, J. Phys. D 17 (1984), pp. L57–L60.

[39] N. Uekawa, Y. Kurashima, K. Kakegawa and Y. Sasaki, J. Mater. Res. 15 (2000), p. 967. View Record in Scopus | Cited By in Scopus (14)

[40] A.R. Bally, E.N. Korobeinikova, P.E. Schmid, F. Levy and F. Bussy, J. Phys. D 31 (1998), p. 1149. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (84)

[41] C.D. Valentin, G. Pacchioni and A. Selloni, Phys. Rev. B 70 (2004), p. 085116.

[42] T. Ihara, M. Miyoshi, Y. Iriyama, O. Matsumoto and S. Sugihara, Appl. Catal. B 42 (2003), p. 403. Article | PDF (140 K) | View Record in Scopus | Cited By in Scopus (285)
T. Ihara, M. Miyoshi and M. Ando, J. Mater. Sci. 36 (2001), p. 4201. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (95)

[43] Y. Sakatani, J. Nunoshige, H. Ando, K. Okusako, H. Koike, T. Takata, J.N. Kondo, M. Hara and K. Domen, Chem. Lett. 32 (2003), p. 1156. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (58)

[44] M. Mrowetz, W. Balcerski, A.J. Cloussi and M.R. Hoffmann, J. Phys. Chem. B 108 (2004), p. 12269.

[45] K. S. Rane, R. Mhalsikar, T. Sato, Shu. Yin, Kuk Cho, E. Dunbar, P. Biswas, Photocatalytic destruction of NO on N- and Fe–N co-doped anatase TiO₂, in: Proceedings of the fourth National Symposium and Conference of Solid State Chemistry and Allied Areas, Indian Association of Solid State Chemsts and Allied Scientists—ISCAS, Goa, 1–3 December 2005.

[46] K.S. Rane, Visible light sensitive TiO₂ photocatalyst, in: Proceedings of the Singapore International Chemical Conference-4 (SICC-4), Singapore, 8–10 December 2005, p. 194 (Inivted Talk- 81-IMSC-A0489).