PHOTOCATALYTIC ACTIVITY OF BISMUTH VANADATES UNDER UV-A AND VISIBLE LIGHT IRRADIATION: INACTIVATION OF ESCHERICHIA COLI VS OXIDATION OF METHANOL

Abstract Four bismuth vanadates have been synthesized by using two different precipitating agents (NH3 and triethylamine) following a hydrothermal treatment at 100 °C for 2 h and at 140 °C for 20 h. Then, solids were characterized by X-ray diffraction, BET surface area, UV–vis spectroscopy and scanning microscopy techniques. The characterization of the synthesized materials showed a well crystallized scheelite monoclinic structure with different morphologies. All materials display optimum light absorption properties for visible light photocatalytic applications. The photocatalytic activity of the catalysts was investigated for the inactivation of Escherichia coli bacteria and the oxidation of methanol under UV–vis and visible light irradiation sources. Main results demonstrate that BiVO4 are photocatalytically active in the oxidation of methanol and are able to inactivate bacteria below the detection level. The activity of the catalyst decreases when using visible light, especially for methanol oxidation, pointing out differences in the reaction mechanism. In contrast with bacteria, whose interaction with the catalyst is limited to the external surface, methanol molecules can access the whole material surface.


Materials characterization.
BET surface area and porosity measurements were carried out by N 2 adsorption at 77 K using a Micromeritics 2010 instrument.
X-ray diffraction (XRD) patterns were obtained using a Siemens D-501 diffractometer with Ni filter and graphite monochromator. The X-ray source was Cu K radiation (0.15406 nm).
Crystallite sizes for BiVO 4 catalysts were estimated from the line broadening of (121) X-ray diffraction signals by using the XPert HighScore Plus software.
The morphology of samples was followed by means of field emission-SEM (Hitachi S 4800).
The samples were dispersed in ethanol using an ultrasonicator and dropped on a copper grid.
UV-vis spectra were recorded by using a Shimadzu AV2101 in the diffuse reflectance mode (R) and transformed to a magnitude proportional to the extinction coefficient (K) through the Kubelka-Munk function, F(R ∞ ). Samples were mixed with BaSO 4 that does not absorb in the UV-vis radiation range (white standard). Scans range was 250-800 nm.
Bismuth and vanadium chemical analysis of the catalysts after reaction was carried out by means of inductively coupled plasma atomic emission spectroscopy (ICP-AES) in a Varian Vista AX equipment. Concentrations were quantified using the spectral lines at 309.310 nm (V) and 306.771 nm (Bi) nm after calibration with certified standards.

Photocatalytic experiments.
The experimental setup for the photocatalytic reactions consists of an annular reactor 15 cm in lenght, 3 cm inner-tube diameter and 5 cm external-tube diameter operating in a closed recirculating circuit with a stirred reservoir tank. Experiments have been carried out using a catalyst concentration of 0.5 g L -1 , a total working volume of 1 L and a recirculation flow rate of 2.5 L min -1 . The catalyst concentration was preliminary optimized to ensure full absorption of the incident radiation both with TiO 2 and with the synthesized bismuth vanadates materials. More details about this reactor can be found elsewhere [21].
Illumination was carried out with a 6000 K Xenon car headlight lamp placed in the axis of the 6 reactor. Spectrum of the light incident to the reactor was controlled with the choice of material for the inner tube of the annulus. Experiments under UV-vis irradiation were carried out using a Pyrex glass tube, whereas visible light driven experiments were performed using a poly-methylmethacrylate (PMMA) inner tube to cut off the light below 380 nm. The radiation flux and spectrum entering the reactor were measured with a BlueWave spectroradiometer (StellarNet Inc.) with a UV-Vis-NIR cosine receptor calibrated in the range 300-1100 nm. Figure 1 shows the comparison betweeen the irradiation spectrum recorded between 300 and 500 nm when glass and PMMA inner tubes were placed in the reactor. As it can be noticed the glass tube keeps almost unaltered the emmision power and spectrum of the unfiltered xenon lamp, whereas PMMA removed most of the UV-A radiation. The integration of the irradiance spectra in the range 276 to 500 nm leads to values of 90 W m -2 when using a glass wall and 68 W m -2 for the PMMA tube. replicates of each decimal dilution were incubated at 37 ºC for 24 h before counting the number of bacterial colony forming units (CFU). Oxidation of methanol was followed by the formaldehyde formation analysed through a colorimetric method at 412 nm [22]. All experiments have been carried out in deionized water at natural pH and have been repeated twice to test the reproducibility of the results.
Before starting the experiments, the catalyst suspension was stirred and saturated with air for 15 min, being the dissolved oxygen the electron acceptor of the process. In the meantime, the lamp was switched on to stabilize its emission power and spectrum. Measurements of the model pollutants before and after this equilibration time show that adsorption is not significantly detected and can be therefore neglected.

RESULTS AND DISCUSSION
According XRD data ( Figure 2), the preparation of BiVO 4 leads in all cases to well crystallized scheelite structure in the monoclinic phase (PDF 14-0688, corresponding to the I2/a space group) [20]. Table 1    m ( Figure 3b) are obtained. The hydrothermal treatment at 140ºC leads, in the case of NH 3 precipitated system, to lower size peanut-like aggregates (Figure 3c). The peanut-like and straw-like particles present microporosity in the nanometer range while for TEA precipitated one, it can be noticed very dense polyhedral particles with lack of porosity as can be appreciable in the micrographs (Figure 2d). This point can be also stated by the significantly low cumulative pore volume obtained for this sample with respect to the rest (Table 1). These results are in agreement with previous studies showing that the different preparation conditions clearly induce dramatic changes in the morphology [24][25][26]. In this case, these changes with the hydrothermal treatment appear to be more pronounced for samples precipitated by TEA.    The comparison of Figure 1 and Figure 4 confirms that there is a good overlapping between the emission spectra of the lamp and the absorption spectra of the catalyst. In both cases, using UV-vis or visible light, the materials should be able to absorb most of the radiation entering the reactor. However, it can be noticed that the decrease in the activity observed when using visible light is significantly more pronounced than the decrease in radiation energy available (90 W m -2 with the glass tube and 68 W m -2 with the PMMA tube).
Consequently, it seems that the reduction in activity of the materials when using visible radiation is not exclusively due to the lower photon absorption in comparison with the use of UV-vis, but to an inherently poorer activity of the material, probably due to the lower oxidation and reduction potential of the electron-hole redox pairs generated under visible irradiation. That could even compromise the possible formation of hydroxyl radicals, leaving only available the direct hole transfer oxidation mechanism, not really favored in the case of methanol due to its weak adsorption interaction with the catalyst surface.
Regarding the activity for bacterial inactivation, Figure 6 displays the evolution of the concentration of viable E. coli bacteria as a function of the irradiation time for samples prepared following hydrothermal treatment at 100ºC. These results confirm the possibility of achieving a six-order of magnitude reduction of CFU concentration and full inactivation of E. coli below the bacterial detection limit after few hours of irradiation, with an activity significantly higher than that observed only upon irradiation without the catalyst. Quantitative results of disinfection activity of the materials have been calculated by fitting the experimental data to a kinetic model based on a series event disinfection mechanism to calculate the inactivation kinetic constant [28]. Results are shown in Figure 7,  Those results are difficult to explain based on physicochemical data, as the small differences observed in band gap values, crystal particle size or surface area are not enough to explain the best activity observed in the N5-100 catalyst. Some authors have associated the different activity behaviour of bismuth vanadates, having the same crystal structure, with the particle morphology which in turn affect the properties of these materials [13,23,29,30]. Thereby, Fan et al. [23] studied the different morphologies of monoclinic BiVO 4 samples and they observed that the flower-like BiVO 4 particles exhibit a high photocatalytic performance which was related with a high separation efficiency of photo-carriers. The lamellar BiVO 4 particles reported by Ke et al [31] also showed the best photocatalytic activity for O 2 evolution. As them, other authors have related the photocatalytic activity of BiVO 4 samples to the particles morphologies [24,32]. In this case, peanut-like BiVO 4 particles present better photocatalytic results under visible light. These materials seem to favor higher visible light absorption and/or lower recombination process.
However, two factors should be taken into account when comparing the results obtained with bacteria and organic molecules. One of them is related to the interaction between catalyst and methanol or catalysts and bacteria. For an effective photocatalytic degradation process, the direct contact between pollutants and photocatalysts is very important but in the case of microorganisms, the inactivation by h + and •OH radicals required close contact between the catalyst and the bacteria [19,33]. It is widely accepted that to reach bacterial inactivation is necessary to achieve a number of cumulative •OH radical attacks to disrupt the cell wall.
Therefore, it is expected that catalyst under visible light conditions take longer to reach the amount of cell damage necessary for bacterial inactivation, as it is confirmed in Figure 7. But, in any case, this mechanism implies a good contact between the bacterium membrane and surface catalyst. On the other hand, this catalyst surface/bacteria interaction can be influenced by the catalysts charge surface. It is known that the point of zero charge of BiVO 4 is around 2.5 [14,34]. This would indicate that in the reaction media, the BiVO 4 surface might be negatively charged. The importance of pollutant adsorption over the negatively charged BiVO 4 surface has been previously stated for rhodamine B and methylene blue [35]. In the present case, methanol adsorption won't be favoured, whereas electrostatic repulsion with the negatively charged bacteria surface would be produced unless an intermediate counter ion layer of cations is formed, favouring the interaction with the catalyst surface as suggested by Pablos et al. [36]. Assuming that the under visible light hydroxyl radical formation is hindered and short-distance hole transfer could be favored, this fact could explain why the decrease in the activity when using visible light is more pronounced for methanol oxidation than for E. coli inactivation.
The second factor that can affect the photocatalytic response of bismuth vanadates is the aggregate size and porosity of the materials (Table 1). Although all catalysts present low BET surface areas, the porosity of the catalysts depend on the preparation conditions ( Figure 3).
Chemicals can diffuse through the porosity and access deeper active sites inside of the catalysts particle, reaching a larger extension of surface area of the BiVO 4 , which is more notable for porous peanut-like and acicular particles. However, in the case of E. coli inactivation due to its huge size only the most external surface of the BiVO 4 particles is really available for the interaction with the bacteria. This fact can explain why the T5-140 catalyst, being one of the least porous catalysts and showing the smallest aggregate size, can show the highest photocatalytic activity. Thus, it can be assumed that the external surface and an intimate contact favoured by the smaller particle size can play an important role in the bacteria inactivation. However, the opposite behavior is not necessarily observed, showing the most porous materials not always the highest activities for methanol oxidation.

CONCLUSIONS
In summary, all the studied BiVO 4 materials have proved to be active under UV-visible and visible light irradiation, with only slightly different kinetic constants both for the photocatalytic oxidation of methanol and the inactivation of E. coli. Catalysts prepared with NH 3 display better photocatalytic response in the degradation of methanol whereas in the case of inactivation of bacteria the more active catalyst are BiVO 4 materials prepared with TEA. In both types of reaction, kinetic constants decrease when using visible light in comparison the use the full spectrum of UV-vis light of the Xe-lamp. However, whereas the decrease in the activity for the inactivation of bacteria is more or less in agreement with the reduction in the total radiation power, the decrease in the efficiency for methanol oxidation is much more significant, pointing out differences in the reaction mechanism in addition to lower radiation absorption. In contrast with bacteria, whose interaction with the catalyst is limited to the external surface, methanol molecules can access the whole material surface.