Characterization of the thin layer photocatalysts TiO₂ and V₂O₅- and Fe₂O₃- doped TiO₂ prepared by the sol–gel method

Tetraisopropyl orthotitanate Ti (OC₃H₇)₄ (TTIP) was added to an ethanol-HNO₃ solution with a pH of about 3–4; the solution was stirred for 1 h. A transparent mixture was obtained (sol). The mixture was added to a quantity of water, stirred and heated with reflux at 80 °C for about 2 h, after which a highly viscous mixture was obtained. By keeping the mixture overnight one can achieve a transparent soft gel mass. After filtration, washing, drying and calcining at 500 °C for 3 h the powder product was obtained and denoted as Ti.

TTIP was dissolved in ethanol solvent in the appropriate ratio. Two milliliters of ammonium vanadate solution in non-ionic water (with concentrations ranging from 0 to 0.005 M) was added. The mixture was acidized by a solution of HNO₃ to reach a pH of about 3–4, and agitated at room temperature until a clearly homogeneous and highly viscous solution was obtained. The content of V₂O₅ (V₂O₅/TiO₂) was adjusted to be 0.028, 0.084, 0.110, 0.188, 0.802 and 1.604 mol% to obtain the catalyst samples 028V–Ti, 084V–Ti, 110V–Ti, 188V–Ti, 802V–Ti and 1604V–Ti, respectively.

Three milliliters of TTIP was dissolved in 50 ml of HNO₃-ethanol mixture (pH 3). The obtained solution must be clear and transparent. The sol was made by adding Fe(NO₃)₃ solution over 90 min. The content of Fe₂O₃ (Fe₂O₃/TiO₂) was adjusted to be 0.025, 0.05, 0.10, 0.50, 1.00 and 2.00 mol% to obtain the catalyst samples 025Fe–Ti, 050Fe–Ti, 100Fe–Ti, 500Fe–Ti, 1000Fe–Ti and 2000Fe–Ti, respectively. The sol was kept at room temperature until a complete hydrolysis of the TTIP and the gel had formed (24 h). The obtained precipitates were filtered and washed with ethanol and distilled water. The samples were dried at 110 °C for 1 h and then calcined and crystallized at 550 °C for 4 h.

Pyrex glass sticks of 6 mm diameter or closed Pyrex glass tubes of 19 mm diameter, 230 mm length, were used as supports for the catalysts. The surfaces of the supports were treated with 1% hydrofluoric acid for about 12 h, washed with distilled water to completely remove the acid, then soaked in a 1% solution of potassium hydroxide in methanol for 12 h. They were then washed with distilled water again, followed by washing with alcohol or acetone. They were then dried and calcined at 450 °C. For the glass sticks the area of coated and lighted catalyst film is 68 cm², the weight of the material coated on a glass stick is 15 mg and the total weight of the catalyst material used in the reaction is 30 mg (two sticks). For the closed glass tubes the weight of the material coated on a tube is 5 mg, 300 nm thick and the lighted area is 144.5 cm².

The TiO₂ or modified TiO₂ catalyst films coated on the pyrex glass sticks and tubes are obtained by using a dipping method from suspension (for P25) or colloidal solutions (for the rest of the samples).

The suspension solution was prepared as follows. The TiO₂-Degussa P25 powder was dispersed in water to form a suspension solution. The glass sticks were dipped into the suspension solution to achieve a film. The necessary amount of catalyst material in the film was obtained after several dipping operations. After being removed from the solution, the sample was dried naturally in air and then dried at a temperature of about 100–105 °C for 2 h. After the last dipping operation, when the necessary amount of material has been reached, the drying process was followed by a calcination process at 450, 550 or 900 °C for about 4 h in air with a flow velocity of 6 l h⁻¹ and the samples were denoted as P25-450, P25-550 or P25-900.

The creation of the catalyst films on the Pyrex glass tubes using the method of dipping in colloidal solutions was done in a similar way, albeit the solutions containing the catalyst materials were colloidal ones of TiO₂ or modified TiO₂. After the last dipping operation, when the necessary amount of material had reached, the drying process was followed by calcination at 450 °C for 4 h in air with a flow velocity of 6 l h⁻¹.

The content of vanadium and iron in the catalysts was determined by atomic absorption spectroscopy (AAS) using the Shimadzu AAS 6800 apparatus. The thickness of the layer was determined by the alpha-step method. The characterization of the catalyst samples was carried out by the methods of Brunauer–Emmett–Teller (BET) adsorption (CHEMBET 3000), x-ray diffraction (XRD; XD 5A-Shimadzu), field emission scanning electron microscopy (FE-SEM; HITACHI S-4800), Fourier transform–Raman (FT-Raman; Perkin-Elmer 2000), UV–Vis absorption (Jassco V-550) and infrared (IR)-spectroscopy (BRUKER VECTOR 22).

The photocatalytic activity of the samples was tested in the gas phase deep oxidation of p-xylene at 40 °C in a micro-flow reactor. The radiation sources were a UV lamp (λ = 365 nm, power of 8 W) and 80 pieces of light emitting diode (LED) (λ in the range 400–510 nm, total power of 19.2 W). The activity of the catalysts was evaluated using results obtained from analysis on the gas-chromatograph GC Agilent 6890 Plus, a flame ionization detector, and the capillary column HP-1 with methyl siloxane (30 m, 0.32 mm, 0.25 μm).

For the sample TiO₂-Degussa P25 (Merck, Germany) with a pH value in the range 3.5–4.5, the purity of 99.5% and the compression density of 130 g l⁻¹ were used. The influence of the treatment temperature on the physico-chemical properties of the TiO₂-Degussa P25 (powder) are shown in table 1.

The XRD patterns of the P25 sample after treatment at 450 and 550 °C (figure 1) are similar and one can observe only the characteristic peaks of the anatase and rutile phases. The content of the anatase phase in TiO₂ was calculated according to the intensity of the characteristic peak of anatase I_A at the angle 2θ = 25.3° and the content of the rutile phase, according to the intensity of the characteristic peak I_R at the angle 2θ = 27.5°

Zoom In Zoom Out Reset image size

According to the XRD patterns of the catalyst samples, the percentage of rutile phase content was calculated as follows [14]:

$$\begin{equation} R(\% ) = \frac{{100}}{{1 + ({{0.8I_{\rm A} }/{I_{\rm R}}})}}, \end{equation} \tag{ 1 }$$

where, R is the percentage of rutile phase (%), I_A is the integral intensity of the reflection peak (peak 101) characteristic of the anatase phase, I_R is the integral intensity of the reflection peak (peak 110) characteristic of the rutile phase.

The size of the crystal grain can be calculated using the Scherrer formula [15]

$$\begin{equation} d_{{\rm TiO}_2 } = 57.3\frac{{K\lambda }}{{\beta _{{1/2}} \cos \theta }}, \end{equation} \tag{ 2 }$$

where d_TiO2 is the average size of the TiO₂ particles, λ is the wavelength of x-ray radiation (λ = 1.5406 Å), K is a constant, usually taken as 0.94, β_1/2 is the width of the characteristic peak (101) of anatase at full-width at half maximum (FWHM), in degrees, θ is the Bragg angle (101 peak of anatase), in degrees, and 57.3 is the conversion factor for transferring from degrees to radians.

The data in table 1 show that after calcination at 450 and 550 °C, the contents of the anatase and rutile phases are not different from those in the uncalcined samples. The value of the ratio of anatase/rutile in these samples is approximately 82/18. Thus, treatment at 550 °C did not cause any phase transition of TiO₂ P25, but increased the particle size and reduced the specific surface area. Compared to the uncalcined sample P25, the specific surface area of the sample treated at 450 °C was not changed significantly, while the surface area of the sample treated at 550 °C decreased about 2.6 times and the grain size increased from 30 to 40 nm.

After calcination at 900 °C for 4 h almost all the anatase phase was transferred to the rutile phase. On the XRD pattern of this sample (figure 1), the characteristic peaks of anatase could barely be observed. This conclusion was also verified by UV spectral analysis. The UV spectra (figure 2) of the sample P25 before and after heating at 450 and 550 °C are similar, while the spectrum of the sample calcined at 900 °C shifted to the region of higher wavelengths. The bending point on the UV spectra for the first three samples was observed at 390 nm, while the bending point on the spectrum of the last sample was observed at 415 nm. Thus one can calculate the value of the band gap energy using the formula

$$\begin{equation} E_{\rm G} = {{hc}/\lambda }, \end{equation} \tag{ 3 }$$

where h is the Planck constant, c is the speed of light and λ is the wavelength of the absorbed light.

Zoom In Zoom Out Reset image size

From table 1 it follows that the value E_G of the untreated sample and the samples calcined at 450 and 550 °C is 3.18 eV, but the sample calcined at 900 °C is characterized by a band gap energy of 2.99 eV. In addition, the slopes of the UV spectra 1, 2, 3 in figure 2 are gradual, while for the sample calcined at 900 °C the spectral curve is steep. This observation is consistent with the conclusion that the last sample calcined at 900 °C only contains the rutile phase. Thus, for the samples of TiO₂ P25 treated at different temperatures, it is appropriate to use UV lamps in photocatalytic reactions.

The IR spectra (figure 3) also indicate a distinct feature of the sample TiO₂ P25 treated at 900 °C compared to the samples calcined at 450 and 550 °C. As is known, the number of OH-groups on the surface of TiO₂ is proportional to the intensity of the peak at about 3420 cm⁻¹. On the samples treated at 450 and 550 °C one can observe two absorption peaks at 3418 and 1635 cm⁻¹. The band concentrated at 3418 cm⁻¹ can be ascribed to basic OH-groups [16] characteristic of the valence vibration of structural OH-groups with basicity. On the other hand, in the sample treated at 900 °C, the peak characteristic of the OH-group almost disappears. This fact can be explained by the fact that the TiO₂ treated at 900 °C was dehydrated completely. Thus, the activation of TiO₂ samples at 450 or 550 °C makes their surface cleaner and many more OH-groups should be formed.

Zoom In Zoom Out Reset image size

The influence of water vapor on the photo-oxidation of p-xylene when using TiO₂ P25 as a catalyst is shown in table 2. TiO₂ catalyst films coated onto Pyrex glass sticks are obtained by the coat-dipping method from suspension solutions.

As can be seen in table 2, the activity of the catalysts increased with water content up to 11.5 mg l⁻¹. However, a further increase in water content led to a decrease of p-xylene conversion. This observation is consistent with results obtained in the PCO of toluene on TiO₂ [17]. As is known, the role of water vapor is to continuously recover OH-groups consumed during the reaction. The presence of water is a necessary condition for the formation of ${\rm OH}^{\unicode{x2022}}$ radicals. However, if the temperature is low, at high concentrations water molecules compete with molecules of the reactant in adsorption, which prevents contact between the reactant and the catalyst leading to a reduction of conversion.

Similarly, the dependence of p-xylene conversion on oxygen concentration is extreme and the highest value was observed when oxygen concentration reached 300 mg l⁻¹, with a further increase of oxygen concentration tending to decrease p-xylene conversion. This result is similar to the previous data obtained by other authors in the photo-oxidation of benzene [18] and trichloroethylene [19]. According to these authors, in photo-oxidation, oxygen is an irreversible photogenerated electron acceptor, preventing the recombination of electrons in the conduction band with positively photogenerated holes in the valence band which extends the lifetime of the holes, leading to an increase of catalytic activity.

From above results one can propose appropriate conditions for the photocatalytic process in the complete oxidation of p-xylene as follows: TiO₂ catalyst treatment at 450 °C; reaction temperature 40 °C; and concentrations of p-xylene, water vapor and oxygen at 19, 11.5 and 300 mg l⁻¹ respectively.

The IR spectra as well as the Raman spectra of the TiO₂ samples modified by V₂O₅ did not show any characteristic peaks of vanadium oxide. Also in the IR spectra (not shown), the characteristic peak of OH-groups at the 3418 cm⁻¹ region appeared with very low intensities, much lower than those in the IR spectra of pure TiO₂ (see figure 3). This fact indicates the poverty of OH-groups on the modified samples. As seen in figure 4, in Raman spectra of the studied samples V–Ti, one can only observe peaks at 147, 398, 518 and 640 cm⁻¹, characteristic of the anatase phase of TiO₂, but no characteristic peaks of vanadium were observed, which is consistent with the results of the XRD analysis (figure 5). According to the XRD patterns, the phase composition of the TiO₂ samples modified by vanadium with a content of V₂O₅ below 0.1 and 1.604% is similar to that of pure TiO₂; the content of the anatase phase is above 90%, the rest is rutile phase. In samples with average contents of V₂O₅ (0.188–0.802%), TiO₂ exists in the entire anatase phase.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

From the results in table 3 it follows that, using the sol–gel method, one can obtain samples of TiO₂ and TiO₂ doped by V₂O₅ with a highly crystalline structure and with 90% or more of the anatase phase. The particle size of pure TiO₂ is quite large (∼30 nm). When V₂O₅ is added, with the content not exceeding 0.1%, the particle size changes little, to about 28–29 nm. An increase of V₂O₅ content up to 0.188–1.604% leads to a reduction of particle size to 21–24 nm. In samples 188V–Ti and 802V–Ti the rutile phase does not exist. This result is consistent with the FE-SEM image of the 028V–Ti and 110V–Ti samples (figure 6). In figure 6 it can be seen that the catalyst particles of sample 110V–Ti have a very uniform size, of about 28–29 nm, indicating a state where vanadium ions are either inserted into the network of TiO₂ or are highly dispersed in the catalyst sample. The thickness of the TiO₂ layer (measured using the Alpha-Step IQ surface profiler) for all of the studied samples was identical in the range 283–288 nm regardless of the V₂O₅ content in the samples.

Zoom In Zoom Out Reset image size

The UV–Vis spectra of the catalyst samples (not shown) indicate that in the near-UV region (λ = 360–400 nm), the light absorption spectra of the modified samples tend to move forward to the region of longer wavelengths but this tendency is very weak. The spectral lines of the samples containing not more than 0.110% V₂O₅ have bending points on the UV–Vis spectra at slightly longer wavelengths of light absorption (table 3). This means that the samples mentioned are also able to absorb photons in the visible region and have lower values of band gap energy. However, the capacity to absorb light in the ultraviolet region of the samples containing from 0.188 to 1.604% of V₂O₅ is the same and approximately equal to that of the pure TiO₂ sample. The sample containing 0.028% of V₂O₅ is characterized by having the longest absorption wavelength, the lowest value of band gap energy and the brightest bluish color. For the rest of the TiO₂-modified samples, the absorption wavelength is shortened compared to that of the sample 028V–Ti, the value of band gap energy gradually increases, and the color gradually fades until these characteristics are similar to those of pure TiO₂. This can be explained as follows: in sample 028V–Ti, part of the vanadium is inserted into the lattice of TiO₂ in the form of V⁴⁺ ions to form links V–O–Ti as the defects, enabling the reduction of the band gap energy and, as a result, extending the light absorption of the catalyst towards the visible range.

Data on the activity of the V–Ti catalyst samples in p-xylene photo-oxidation are presented in table 3. From table 3 it follows that under the irradiation of UV light, the values of the initial conversion and yield obtained on modified catalysts in p-xylene conversion for 60 min on 1 g of catalyst are lower than those for pure TiO₂. Among the modified catalysts, sample 028V–Ti expressed the highest efficiency in p-xylene conversion as it had the lowest value of band gap energy.

Thus, doping V into TiO₂ samples using a sol–gel method led to the reduction of their particle size from 30 to 21 nm, but created samples with low concentrations of OH-groups, so that the photocatalytic activity as well as the stability of these catalysts decreased.

The XRD patterns (figure 7) of the catalyst samples show only the characteristic peaks of the anatase (2θ = 25.3°) and rutile (2θ = 27.5°) phases without any characteristic peaks of Fe₂O₃. This result indicates that some Fe³⁺ ions are probably able to replace Ti⁴⁺ ions in the crystal network of TiO₂.

Zoom In Zoom Out Reset image size

The Raman spectra of the Fe-doped TiO₂ catalysts (figure 8) show peaks at 407, 522, 648 and 853 cm⁻¹, characterizing the light absorption of anatase TiO₂ [20]. No characteristic peaks of Fe₂O₃ are observed. It is very probable that the presence of Fe³⁺ ions in the lattice resulted in a shift of the absorption peaks characteristic of TiO₂ to the region of higher wavenumber. The specific features of the Raman spectra of the Fe-doped TiO₂ catalysts are the absence of the characteristic peak of anatase at 144 cm⁻¹ and the appearance of a new peak at 853 cm⁻¹. These facts confirm the idea about the penetration of Fe³⁺ ions into the network of TiO₂ crystals.

Zoom In Zoom Out Reset image size

The particle size values of the catalysts were calculated using Scherrer's equation. The figures in table 4 show that the influence of iron oxide content on the ratio of anatase/rutile phases is not simple. In the range from 0.05 to 0.1 mol% of Fe₂O₃ the percentage of the rutile phase increases from 24.1 to 41%, but at higher values of iron oxide content the proportion of rutile to anatase phase in the studied samples decreased regularly and reached the quantity of 18.2 mol%, close to that in the commercial TiO₂ P25. Also, iron oxide content influences the particle size and, as a rule, the specific surface area of the catalyst samples. Although the trend is not very clear, the common understanding is that the particle size should decrease with the content of iron oxide, while the values of surface area varied in the opposite direction.

The FE-SEM image (figure 9) of the 025Fe–Ti catalyst sample shows uniform particle size, but in the range from 0.05 to 1.0 mol% of iron oxide content the images characterize a mixture of fine particles and bulk TiO₂. In contrast, the image of sample 2000Fe–Ti shows very fine and uniform particles of TiO₂ crystals.

Zoom In Zoom Out Reset image size

The pure TiO₂ catalyst is characterized by the band gap energy value of anatase TiO₂ (E_G = 3.20 eV). The UV–Vis spectra of all of the Fe-doped TiO₂ catalyst samples (figure 10) show a red shift tendency and this shift depends on the iron content in the lattice. Concomitantly, the increase of iron content in the lattice resulted in a decrease of E_G. The 2000Fe–Ti sample expressed the biggest shift: the value of λ_max reached 464 nm and the value of E_G decreased to 2.63 eV. Also one can observe that the color of the catalyst samples changed from white to yellow with the increase of iron content in the lattice.

Zoom In Zoom Out Reset image size

The IR spectra (figure 11) of all of the Fe-doped TiO₂ samples show the characteristic peaks of OH-groups at the surface (3225 cm⁻¹), molecular water (1621 cm⁻¹), Ti–O (653–550 cm⁻¹) and Ti–O–Ti (495–436 cm⁻¹) as in the case of pure TiO₂. Nonetheless, it is interesting to note the appearance of a new peak at 2200 cm⁻¹, the absorption intensity of which regularly increases with iron content. The order is as follows: 2000Fe–Ti > 1000Fe–Ti > 500Fe–Ti > 100Fe–Ti > 050Fe–Ti > 025Fe–Ti. It is very probable that this peak could be assigned to the Ti–O–Fe vibration.

Zoom In Zoom Out Reset image size

The results on the catalytic activity of the TiO₂ catalysts modified by iron oxide in p-xylene photo-oxidation under UV and visible light irradiation are presented in table 5. From the data in table 5 one can see that in regime 'd', under UV light irradiation, the initial activity and efficiency of pure TiO₂ are higher than those of the modified catalysts. Samples 1000Fe–Ti and 2000Fe–Ti expressed the lowest values of efficiency in spite of having the same content of the anatase phase as of pure TiO₂ P25. In regime 'e', when the reaction occurred under the visible light of LEDs (λ = 470 nm), the modified catalysts gave higher conversion and yield than the sample of pure TiO₂. Obviously, the modification of TiO₂ by Fe increased its capacity for light absorption in the long wavelength region, so the modified samples showed an advantage in the visible region of radiation. According to the authors of [21], the superior catalytic activity of the Fe-doped samples is associated with the formation of a ferrioxalate complex. Ferrioxalate is considered as one of the best complexes of Fe³⁺-containing polycarboxylate. It can absorb orange light and use visible light efficiently. The Fe-Ti catalysts prepared by the flame spray pyrolysis method (FSP) can be regenerated by the oxidation of Fe (II) to Fe (III) under visible light radiation. Yu et al [22] proposed that the reason for the raised photocatalytic activity of TiO₂ films doped by Fe³⁺ is the generation of OH-groups as well as oO atoms.

The results in table 5 show that under visible light irradiation the initial activity of the modified catalysts increased with the content of Fe₂O₃ and reached a maximum value at 0.1% of the modifier concentration. As the content of Fe₂O₃ continues to increase to 2.0%, the conversion of p-xylene decreased. The optimum content of Fe₂O₃ in TiO₂ is 0.5% under UV radiation and 0.1% under the visible light radiation of LEDs.

In the decomposition of several organic compounds under visible light radiation, the authors of [23] found the maxima of methanol photo-oxidation at Fe(III) doping levels of 0.25 and 0.5 atom%. The extreme nature of the effect of Fe content is due to several reasons. When replacing Ti(IV) by Fe(III), the created Fe(III) centers should be considered as trapping sites within the TiO₂ matrix as well as on the surface of the TiO₂ particles [23]. Based on the favorable energy levels, Fe(III) centers may act either as an electron or as a hole trap. The authors of [23] explained the enhancement of conversion efficiency in methanol degradation on Fe(III)-doped titania by assuming that the Fe(III) center acts predominantly as a electron trap from which the electron is transferred to molecular oxygen more rapidly than in the undoped catalyst:

$$\begin{eqnarray} &&{\rm TiO}_2 : {\rm Fe}^{\rm III} \buildrel{{{\rm e}^{-}_{\rm cb}}}\over{\!\!-\!\!-\!\!\!-\!\!\!-\!\!\!-\!\!\!\longrightarrow} {\rm TiO}_2 : {\rm Fe}^{\rm II}\buildrel{{\rm O}_{2}}\over{\!\!-\!\!-\!\!\!-\!\!\!-\!\!\!-\!\!\!\longrightarrow} {\rm TiO}_2:{\rm Fe}^{\rm III} + {\rm O}_2^-.\nonumber\\ \end{eqnarray} \tag{ 4 }$$

From the reaction kinetics of methanol photo-oxidation on TiO₂ it should follow that the catalysis of the O₂-reduction by an electron relay would increase the photocatalytic activity.

Wang et al [23] demonstrated that, for methanol photo-oxidation the average number of Fe(III) centers per particle is approximately one. At optimal concentrations, Fe(III) in the particle is sufficient as a shallow electron trap for the optimal catalysis of the O₂-reduction. If more than one Fe(III) center is present per particle, a distribution of the catalytic activity will arise from their different positions in the bulk or at the surface of the photocatalyst, respectively. Alternatively some of the Fe(III) dopants might act as hole traps

$$\begin{equation} {\rm TiO}_2 : {\rm Fe}^{\rm III} + h_{vb}^ + \buildrel{{{\rm e}^{-}_{\rm cb}}}\over{\!\!-\!\!-\!\!\!-\!\!\!-\!\!\!-\!\!\!\longrightarrow}{\rm TiO}_2 : {\rm Fe}^{\rm IV} \end{equation} \tag{ 5 }$$

which leads to an enhanced recombination of the trapped charge carriers by the following reaction:

$$\begin{equation} {\rm TiO}_2 : {\rm Fe}^{\rm II} + {\rm TiO}_2 : {\rm Fe}^{\rm IV} \buildrel{{{\rm e}^{-}_{\rm cb}}}\over{\!\!-\!\!-\!\!\!-\!\!\!-\!\!\!-\!\!\!\longrightarrow}{\rm TiO}_2 : {\rm Fe}^{\rm III} + {\rm heat}. \end{equation} \tag{ 6 }$$

In addition, the surplus quantity of Fe³⁺ ions tends to cause the transition of the anatase phase into the rutile one [22]. This fact is another reason for the reduction of catalytic activity. Thus there are several speculative reasons to explain why two catalysts, 2000Fe–Ti and 1000Fe–Ti, with high Fe₂O₃ contents (1.0 and 2.0% of Fe₂O₃) have lower activity than the catalysts containing 0.1 and 0.5% of Fe₂O₃.

The efficiency of the studied catalysts in p-xylene conversion also varies in similar ways to the extent of conversion, but in the case of the 2000Fe-Ti sample, the efficiency is relatively high indicating its good stability under visible light radiation. This fact may be related to the consistency between the physico-chemical properties of the sample (λ_max = 464 nm, E_G = 2.672 eV) and the radiation energy of the LEDs that facilitate the process in producing electron–hole pairs, leading to the generation of ${\rm OH}^{\unicode{x2022}}$ radicals. Besides, the activity of this catalyst is not very high and the formation of intermediate compounds is relatively weak, so the carbonaceous deposit on the catalyst surface is small, consequently the stability of the catalyst should be improved. The results obtained above suggest the possibility of using combined radiation from UV and LED lighting for improving the efficiency of the catalysts in photocatalytic oxidation. The results of the study of 2000Fe–Ti using different modes of radiation are presented in table 6.

From the results in table 6 it can be seen that the conversion extents of p-xylene on 2000Fe–Ti under combined radiation (mode III) at every experimental point are higher than under other modes of radiation, although the intensity of the combined radiation (659 lux) is lower than that in mode I (872 lux) about 25%. This may also be explained by the fact that when using UV lamps and LEDs there is a combination of two elements: the strong intensity of UV light and the suitable wavelength emitted by LEDs, at the same time side reactions (creating carbon deposits) should be diminished and the catalyst stability is improved, so that the photocatalytic activity in this mode could be kept stable over 60 minutes. That is why the conversion yield of p-xylene for 60 min on catalyst 2000Fe–Ti under combined lighting (LED + UV) is two times higher than under the UV light radiation and five times higher than under the LED radiation. This result confirms the advantages of the combined mode in the utilization of lighting sources for the reaction of p-xylene oxidation in the gas phase, as well as creates a scientific basis for using sunlight in waste gas treatment.

Since the reaction was conducted under different conditions, the activity of the catalysts can only be compared by using the values of the reaction rate. The reaction rate is calculated as follows [24]:

$$\begin{equation} r = \frac{{C_{{\rm xyl}}^0 VX}}{m}, \end{equation} \tag{ 7 }$$

where C⁰_xyl is the initial concentration of p-xylene in the gas mixture, mmol l⁻¹; X is the p-xylene conversion to CO₂ and H₂O; m is the mass of the catalyst, g and V is the velocity of the flow, l h⁻¹.

The current concentration of the substances is calculated as follows:

$$\begin{equation} C_j = \frac{{P_j }}{{RT}}, \end{equation} \tag{ 8 }$$

with P_j being the partial pressure of the j component, hPa; R = 0.082 l atm mol⁻¹ K⁻¹ = 82 l hPa mol⁻¹ K⁻¹ = 0.082 hPa mol⁻¹ K⁻¹.

Thus the reaction rate in a gradientless system at atmospheric pressure is calculated by the following equation

$$\begin{equation} r = \frac{{P_{{\rm xyl}}^0 VX}}{{24.436m}}, \end{equation} \tag{ 9 }$$

where P⁰_xyl is the initial partial pressure of p-xylene, hPa. The results of the calculations are shown in table 7.

The addition of 0.028% V into TiO₂ using the sol–gel method (which, as noted above, insignificantly extends the region of light absorption and slightly reduces its particle size while creating samples with low concentrations of OH-groups), makes the catalytic activity of the modified catalysts in p-xylene (conversion extent as well as yield) photo-oxidation lower than that on the pure TiO₂ sample.

The effect of TiO₂ modification with iron oxide by the same method is very different. In this case the region of light absorption is moved to the visible area (λ = 464 nm) and consequently the value of the band gap energy of the modified catalyst is reduced to 2.67 eV, so that it should be possible to use visible light for the activation of photocatalytic reactions. Under ultraviolet light Fe–Ti catalyst samples with Fe content from 0.025 to 0.500% possesses the same activity as the pure TiO₂ sample, but with the content of iron oxide 2.00% the catalyst possesses a lower value of activity compared to the original sample. However, under visible light the catalytic activity of the TiO₂ catalysts modified with Fe, expressed as the reaction rate, is higher by 1.85 to 3.7 times compared to that of the pure TiO₂; the values of the conversion yield for all of the modified catalysts are also two to three times higher and tend to increase with Fe₂O₃ content. Certainly this state is associated with the effect of Fe₂O₃ in extending the region of light absorption towards the longer wavelength zone and narrowing band gap energy, as well as reducing particle size of the modified catalysts. Based on the obtained results, a combined utilization of UV and visible radiation for the activation of catalysts was applied and this experiment has given a very interesting successful result: the activity of the 2000Fe–Ti catalyst rose significantly as can be seen in table 7.

The fact that combined UV + LED lighting increased the activity and stability of catalysts in photo-oxidation opens the possibility for the application of Fe₂O₃-modified TiO₂ as highly active catalysts in exhaust treatments under sunlight. The advantages of thin film catalysts include high efficiency and easy recovery. Catalysts can be recovered by processing in air at 500 °C.