Abstract

This study was focused on the application of white and blue light emitting diodes (LEDs) as sources for the photocatalytic degradation of organic dyes in liquid phase with visible light. The photocatalytic activity of N-doped titanium dioxide, synthesized by direct hydrolysis of titanium tetraisopropoxide with ammonia, was evaluated by means of a batch photoreactor. The bandgap energy of titanium dioxide was moved in the visible range from 3.3 eV to 2.5 eV. The visible light responsive photocatalysts showed remarkably effective activity in decolorization process and in the removal of total organic carbon. Methylene blue was also used as a model dye to study the influence of several parameters such as catalyst weight and initial concentration. The effect of dye on the photocatalytic performance was verified with methyl orange (MO). The results demonstrated that the right selection of operating conditions allows to effectively degrade different dyes with the N-doped TiO2 photocatalysts irradiated with visible light emitted by LEDs.

1. Introduction

In recent years, an enormous interest has been devoted to heterogeneous photocatalysis using oxide semiconductors owing to its potential applications to both environmental applications and organic synthesis [117]. Titanium dioxide, TiO2, is characterized by chemical stability, no toxicity and cheap production cost, so that it represents one of the most important oxides employed in several fields of photochemistry, for example, in environmental remediation, photoelectrolysis of water and dye-sensitized solar cells [1822]. Since TiO2 absorbs only a small fraction of the solar spectrum emission, due to the value of its bandgap energy (3.0–3.3?eV), the metal-doped TiO2 has attracted considerable attention due to its reported activity in the visible light. Various metals such as Fe, Cr, Co, Mo, and V have been employed to tune the electronic structure and enhance the photocatalytic activity of titanium dioxide [2325]. However, metal doping can result in thermal instability and increase of carriers trapping, which may decrease the photocatalytic efficiency. Furthermore, the typical preparation of transition metal-doped TiO2 requires more expensive ion-implantation facilities [26, 27].

Recently, many efforts have been made to modify titanium dioxide with nonmetals, such as B, C, N, S, and F, to efficiently extend the photoresponse from the UV to the visible light region [2840]. Furthermore, some theoretical calculations have also been performed to suggest that anion doping of TiO2 has considerable effects on the band gap alteration. Since the first report by Kobayakawa et al. [41], considerable research has been done with respect to the preparation, characterization, photocatalytic activity, and mechanism of nitrogen-doped TiO2 in visible light. The N-doped TiO2 seems to be the most promising among the so-called second generation photocatalysts. N-doping can be obtained by various methods such as sputtering [42], treating of TiO2 powders in ammonia atmosphere [43], or hydrolysis with urea in the presence of organic or inorganic titanium-based compounds [44]. N-doped titanium dioxide nanoparticles (N-TiO2) demonstrated the ability to degradate using tungsten bulb lamp [45] also industrial wastewaters as tannery wastewater, which instead requires a complex combination of conventional treatments to be satisfactory depolluted, as reported in [46].

Until now, there are very few reports on the use of LEDs emitting in visible region for the degradation of methylene blue, typically performed at very low dye concentration (1?ppm) [47].

For the above reasons the objective of this study was to explore the possibility of using this type of LEDs as source of visible light for the photocatalytic degradation of dyes with higher concentration in the presence of N-doped TiO2 photocatalyst. The catalyst was obtained by direct nitration with NH3 during the hydrolysis of titanium tetraisopropoxide. The influence of several parameters, such as catalyst weight, methylene blue concentration, and dye type were investigated.

2. Experimental

2.1. Photocatalysts Preparation and Characterization

N-doped TiO2 photocatalysts were prepared by sol-gel method, according to the modified synthetic procedure developed by Sato [47]. Different amounts (25?mL, 50?mL, 75?mL, and 100?mL) of ammonia aqueous solution at 30?wt%, supplied by Carlo Erba, were added quickly to 25?mL of 97?wt% titanium tetraisopropoxide (TTIP by Sigma Aldrich) at 0°C, while the solution was vigorously stirred, leading to the formation of a white precipitate. The precipitate was carefully washed with water and centrifuged to be separated. Finally, the obtained powder was dried and calcined at 450°C for 30 minutes to get TiO2 in the anatase phase. X-ray diffraction (XRD) measurements of the obtained yellow samples were carried out with an X-ray microdiffractometer Rigaku D-max-RAPID, with Cu-K radiation. Laser Raman spectra were obtained at room temperature with a Dispersive MicroRaman (Invia, Renishaw), equipped with 785?nm diode-laser, in the range 100–2500?cm-1 Raman shift. Aggregates sizes of photocatalyst particles in aqueous suspension were measured by a DLS instrument (Brookhaven Plus 91). UV-Vis DRS measurements were obtained by means of a Perkin Elmer spectrometer Lambda 35 using a RSA-PE-20 reflectance spectroscopy accessory (Labsphere Inc., North Sutton, NH). All spectra were obtained using an 8° sample positioning holder, giving total reflectance relative to a calibrated standard SRS-010-99 (Labsphere Inc., North Sutton, NH). The reflectance data were reported as the value from Kubelka-Munk theory versus the wavelength. Band gap determinations were made by plotting versus (eV) and calculating the intercept of a line passing through .

2.2. Photocatalytic Activity Tests under Visible Light

In a typical activity test a defined amount of photocatalyst was suspended in 100?mL of MB solution. The suspension was left in dark condition for 2 hours to reach the adsorption equilibrium, and then photocatalytic reaction was initiated under visible light up to 3?h. The experiments were performed with a pyrex cylindrical photoreactor (?cm) equipped with an air distributor device (?cm3/min (STP)), a magnetic stirrer to maintain the photocatalyst suspended in the aqueous solution, and a temperature controller. The photoreactor was irradiated by a strip composed by 30 white light LEDs (nominal power: 6?W) with wavelength emission in the range 400–800?nm or by a similar number of blue light LEDs (nominal power: 6?W) with wavelength emission in the range 400–550?nm. The LEDs strip was positioned around the reactor so that the light source uniformly illuminated the reaction volume. The curve inside Figure 1 represents the emission spectrum of white LEDs, while the fraction of curve coloured in blue evidence the spectrum emission of blue LEDs. On the left side of the same figure a schematic picture of the photocatalytic reactor is presented. Slurry samples were collected at fixed time intervals, and centrifuged for 20 minutes at 4000?rpm for removing photocatalyst particles. The centrifuged samples were analysed to determine the change of dyes concentration, measured with a Perkin Elmer UV-Vis spectrophotometer at ?nm. A standard calibration curve was obtained for different MB concentration and allowed to convert absorbance to concentration (mg/L) units. Total organic carbon (TOC) of solution was evaluated from CO2 obtained by catalytic combustion at °C. CO2 produced in gas-phase was monitored by continuous analyzers, measuring CO, CO2 (Uras 14, ABB), and O2 (Magnos 106, ABB) gaseous concentrations.

3. Results and Discussion

The list of catalysts and their bandgap energy is reported in Table 1.

The reflectance measurements of N-doped TiO2 showed that the absorption onset shifted from 380 to 480?nm determining a decrease of bandgap values from 3.3?eV (the typical bandgap of undoped TiO2) to 2.5?eV (Figure 2), which is an unexpected value for the only anatase phase. This change in bandgap is therefore attributed to the presence of nitrogen in the crystal structure phase. This result confirms the ability of photocatalysts to absorb visible light.

XRD analysis (not reported) revealed that for undoped TiO2 and for N-TiO2 samples the only crystalline phase is anatase and it did not change after the doping process. The average size of TiO2 crystallites was calculated using the Scherrer equation [49] on diffraction plane (101) and the obtained values are reported in Table 1. The crystallite average size was 16?nm for N_1 and N_3 samples, while for N_2 and N_4 it was 15?nm. In contrast, DLS overestimated the size of TiO2 particles due to agglomeration in the suspension. For N-TiO2 samples, the degree of aggregation increased with decrease in crystallites size due to the high surface energy of smaller particles.

The results obtained from XRD are confirmed by Raman spectra (Figure 3), since the band positions are in complete accordance with those reported in previous studies for anatase powder [5061]. In fact the Raman modes at 141, 194, 394, 515, and 636?cm-1 are assigned to the anatase phase.

The behaviour of MB decolourisation under visible light generated by white LEDs is represented in (Figure 4) for undoped TiO2 and doped TiO2 with different nitrogen content.

In dark conditions a decrease of MB concentration was observed during the first hour of the test and it was unchanged in the second hour, indicating that the adsorption equilibrium of dye on catalyst surface was reached. The curves show that N-TiO2 catalysts have different amounts of MB adsorbed in dark. To explain this last result, the specific area was estimated (Table 1); for the sample N_1 and N_3 it was 75?m2/g, while for N_2 and N_4 it was 80?m2/g. As expected, the amount of organic dye adsorbed increases the higher is the specific surface area, fairly accordingly to the differences in the area values, as shown in Figure 4.

After the dark period, the solution was irradiated with visible light and the reaction started to occur. Figure 4 shows that undoped TiO2 is slightly effective for MB decolourization, the / reduction being about 8%, a value similar to that of photolysis reaction (7%). On the contrary, all the N-doped TiO2 photocatalysts exhibited higher photocatalytic activity under visible light irradiation. The order of decolorization activity of N-doped TiO2 after 180?min was as following: N_2 > N_4 > N_3 ~ N_1. The final value of MB conversion depends on dye concentration after the dark period.

The decolourization of MB does not necessary correspond to the oxidation and mineralization of the molecule; in fact the reduced form of MB (LMB, leuco) which is colourless can be produced in the presence of light [49]. The lack of coincidence among the best performing samples with respect to either decolorization or mineralization can be the result of different routes followed by MB during irradiation. Consider the following: The MB molecule can be transformed into LMB through reduction by electrons in the conduction band (4) or oxidized by interactions with the valence band holes or native OH species, starting with a de-methylation step to be finally mineralized (2). For long reaction times also LMB can be further degradated and mineralized.

The analysis of gases coming from the photoreactor showed the presence of CO2 and SO2 during the visible light irradiation, confirming the occurrence of MB mineralization (Figure 5). In fact, a significant TOC reduction was obtained. Its final value increased up to 97% by increasing the doping level (Figure 6). In this case, the order of activity was: N_4 N_3 N_2 N_1.

The photocatalyzed decolorization process can be described by a first-order kinetic equation (6) with respect to the concentration of MB [62, 63]. Consider where:

concentration of MB at any given time;

concentration of MB after dark adsorption:

irradiation time;

apparent kinetic constant.

The obtained results are plotted in Figure 7.

The order of kinetic constants of N-doped TiO2 after 150?min of light irradiation is .

Many reports suggest that the amount of catalyst plays a major role in the degradation of organic compounds or dyes in photocatalytic systems. To avoid the use of a catalyst excess, it is necessary to identify the optimum loading for an efficient removal of dye or organic compound. So it is necessary to optimize the amount of catalyst with respect to the highest photocatalytic activity.

To study the effect of the amount of catalysts, different quantities of powder between 0.05 and 1.5?g were used. The initial concentration of MB dye (7.5?ppm) was kept the same in all these experiments. Figure 8 shows the effect of the amount of catalyst on the conversion of MB in the presence of white LEDs. The progress of the conversion is linear up to 0.3?g, while for further increase of catalyst amount, the conversion stabilizes. These data indicate that 0.3?g of powders are completely exposed to the radiation. This phenomenon may be explained considering that with an increase of catalyst loading in the aqueous medium, the light penetration through the solution becomes difficult. Therefore 0.3?g of photocatalyst loading is considered to be an optimal value.

Also the dye concentration plays a key role in the photocatalytic degradation. The effect of MB initial concentration on the photocatalytic activity was checked in the range between 4 and 95?ppm with N_4 catalyst and optimal catalyst loading (0.3?g in 100?mL of solution). The trend of the curves was similar and it is shown in Figure 9. After 120 minutes of dark adsorption, the photocatalytic test started. At fixed reaction times, the increase of MB initial concentration determined a decrease of the photocatalytic activity. This could be due to the increase of colour intensity of the solution that reduces the light penetration into the aqueous medium, meaning that the path length of photons inside the solution decreases. After 180 minutes of irradiation the final value of the decolorization was 100% in the case of 4?ppm and 15% in the case of 95?ppm MB initial concentration. The same effect was observed by Matthews during the photocatalytic degradation of MB with TiO2 catalysts [48].

Moreover, the efficiency of the system was also evaluated with blue LEDs as source of visible light and methyl orange (MO) as organic dye. Figure 10 shows the comparison between the obtained results in presence of white and blue LEDs. As expected from the evaluation of N_4 bandgap energy (2.5?eV), only a fraction of radiation is used in the decolorization process. In fact, the curves related to the MB concentration during the irradiation time showed a similar trend because the range of radiation used by the photocatalyst was that one having an emission wavelength lower than 440?nm. Similar results were obtained for methyl orange (MO).

The photocatalytic stability of N-doped TiO2 was evaluated with recycling experiments (Figure 11). In the first cycle, the amount of MB adsorbed was lower than that one obtained with fresh catalyst because, using the same catalyst coming from the first test, a certain quantity of dye is already adsorbed. The rate of dye removal depends on the initial concentration in solution, so it is higher, the higher is residual MB after dark adsorption. However, it must be noted that the photocatalytic activity remained high and the level of final MB removal was the same at the end of the test, indicating that no deactivation occurred and the dye concentration decreased continuously under visible light irradiation.

4. Conclusions

In this study, the evaluation of the photodegradation of methylene blue (MB), as dye model system, in presence of white and blue light emitting diodes (LEDs) as source of visible light, was evaluated. The photoreactivity was assessed over N-doped TiO2 photocatalysts in batch experiments, evaluating the abatement of methylene blue, and in addition of methyl orange. The best catalyst was N_4 that showed remarkable efficiencies in decolorization process and in the removal of total organic carbon. The influence of several parameters such as catalyst weight, dye concentration and type, light source was assessed.