Mechanochemical Synthesis of Visible Light Sensitive Titanium Dioxide Photocatalyst

1 Centre of Instrumental Techniques, Institute of Inorganic Chemistry of the AS CR, v.v.i., 250 68 Řež, Czech Republic 2 Department of Mechanochemistry and Reactivity of Materials, Instituto de Ciencias de Materiales de Sevilla C.S.I.C., Americo Vespucio s/n, Isla de la Cartuja, 41092 Sevilla, Spain 3 Institute of Physical Chemistry and Chemical Physics, Faculty of Chemical and Food Technology, Slovak University of Technology in Bratislava, Radlinského 9, 812 37 Bratislava, Slovakia


Introduction
Titanium dioxide attracted a great attention since Fujishima and Honda discovered in 1972 the photocatalytic splitting of water on TiO 2 electrode [1].In the last two decades, the application of nanosized TiO 2 focused on environmental remediation, especially water detoxification and air purification (both indoor and outdoor), commanded attention to these phenomena [2][3][4][5][6][7][8].TiO 2 is an efficient photocatalyst, but UV light is necessary for its activation owing to a rather wide band gap energy (∼3.2 eV).Solar energy contains only about 4% UV, light and much of the rest is visible light.In order to utilize the visible (solar) light efficiently for the photocatalytic reactions, titanium dioxide has to be modified in order to lower the energy necessary for excitation of the electron into the conduction band.There have been described several approaches how to modify TiO 2 with this respect.Several authors substituted Ti 4+ in TiO 2 by metal ion implantation (e.g., Cr 3+ or V 3+ , lanthanides, etc.) [9][10][11].It was demonstrated that the absorption band of Cr 3+ -doped TiO 2 shifted and that NO x decomposed giving rise to N 2 , O 2 , and N 2 O as the result of the photocatalytic reaction under visible light irradiation with wavelength larger than 450 nm.
N-doped titania photocatalyst sensitive to visible light irradiation was described in 2001 by Asahi et al. [12,13].The authors proved that nitrogen doping leads to narrowing of band gap by mixing the N 2p and O 2p states and consequently to induction of the visible-light responsive photocatalytic activity.It was also forecasted that other kinds of anions such as C, S, and F would result in a similar effect to nitrogen.Afterwards, various types of visible light sensitive TiO 2 photocatalysts doped with anions such as sulphur and carbon have been widely studied [14][15][16][17][18][19][20][21][22].The structure of N-doped TiO 2 as well as the effect of Ndoping on TiO 2 photoactivity has been discussed in many recent papers, for example, [23][24][25] and reviews [26,27].At present, there exist hundreds of papers and patents describing successful preparation of visible light sensitive International Journal of Photoenergy photocatalysts.However, most preparation methods used are high temperature processes or using expensive precursors or preparation instruments.According to our experience, the enhanced photoactivity in the visible light region is often compensated by sinking the activity at the UV light irradiation; therefore, the overall photoactivity of such material at sun light is often significantly lower in comparison to highly effective photocatalysts like, for example, DEGUSSA AEROXIDE TiO 2 P 25.Hence, synthesis of highly active visible light sensitive photocatalyst should be still considered as an actual problem to be solved.The measurement of rate of decomposition of various dyes is often used for determination of visible light activity of the photocatalysts.As it was demonstrated recently, such approach can lead to serious errors owing to possible dye-sensitization of the photocatalyst to visible light or injection of the photoexcited electrons from dyes directly to the photocatalyst particles [28,29].In this study the photocatalyzed decomposition of 4-chlorophenol is used as model reaction owing to stability of this compound under the reaction conditions to UV light as well as availability of many data in the literature using 4-chlorophenol as the model compound to determine the titania photocatalyst activity in the UV region.
It is well known that the activation of chemical reactions by mechanical energy can lead to many interesting applications, ranging from waste processing to the production of advanced materials with novel microstructures and enhanced mechanical properties [30].Among other phenomena, mechanical energy can induce phase transformation and growth of nanocrystals in titania powder as well as change their surface properties [31][32][33][34].It has been observed that mechanical treatment also affect significantly the photoactivity of titanium dioxide [35,36].
Recently we studied mechanically activated anatase → rutile phase transformations using phase pure anatase [37] supplied by Aldrich.We observed that the anatase → rutile phase transformation was significantly hindered by presence of ammonium carbamate, whereas for the pure sample the transformation was complete within one hour of grinding, at presence of ammonium carbamate after 2.5 hours only 40% of anatase has been transformed into rutile.Ammonium carbamate addition also caused decrease of crystallite size in the ground products at otherwise similar grinding conditions.We also observed that the anatase → rutile phase transformation of nanocrystalline anatase at the same grinding conditions proceeds significantly more slowly and the products are yellow coloured.

Experimental
2.1.Chemicals.For the grinding experiments we used nanocrystalline anatase prepared by homogeneous preci-pitation of TiOSO 4 aqueous solutions with urea at 100 • C. TiOSO 4 (30 g) was dissolved in 100 mL hot water acidified with 98% H 2 SO 4 .The pellucid liquid was diluted into 4 L of distilled water and mixed with 200 g of urea.The mixture was heated to 100 • C under stirring and kept at this temperature for 6 hrs until pH 7 was reached and ammonia escaped from the solution.The precipitates formed have been thoroughly washed with distilled water, filtered off, and dried at laboratory temperature.The precipitate consists of crystalline anatase nanoparticles with diameter 4-6 nm, the nanoparticles are aggregated into firm spherical clusters ∼2 μm diameter, specific surface area of the sample is 200-300 m 2 /g.The material contains ∼0.1-0.5% by wt. of strongly bound SO 2− 4 and does not contain any nitrogen containing compound, for example, urea or NH + 4 salts (sample TIT 144).The detailed description of the preparation procedure and apparatus used as well as properties of the precipitate is given in our previous reports [38][39][40].
The standard materials DEGUSSA AEROXIDE TiO 2 P 25 photocatalyst and nanocrystalline anatase BDH [41] were used for comparing the photocatalytic activity of samples.
4-chlorophenol for photocatalytic activity tests (99%) was purchased from Fluka (Germany) and used without further purification.

The Grinding Experiments.
A high grinding energy planetary ball mill Pulverisette 7 (Fritsch, Germany), has been used.Steel grinding bowls with 45 mL volume and 40 mm in inner diameter charged with 7 steel balls of 15 mm in diameter were used.Ball weight/sample weight ratios close to 20 and a rotation speed of 700 rpm have been used.The sample "TIT 144 ground" was prepared by grinding the sample TIT 144 for 1.7 hrs with addition of 20% by wt. of ammonium carbamate.In order to clean the surface from possible rests of the grinding additive, the sample was subjected to subsequent annealing at 400 • C for 1 hr in an air atmosphere.

Physical Characterization of the Materials.
The X-ray diffractions diagrams were recorded with a PHILPS X'pert Pro equipped with a X'celerator.Cu K α radiation and a nickel filter were used.A voltage of 40 kV and an intensity of 40 mA were employed.The percentages of anatase, rutile, and the high pressure Titania phase (TiO 2 II) with structure α-PbO 2 were determined by the direct method [42] from Transmission electron micrographs were obtained using a JEOL JEM 3010 microscope operating at accelerating voltage of 300 kV.The samples for electron microscopy were prepared by grinding and subsequent dispersing of the powder in propyl alcohol and applying a drop of very dilute suspension on carbon-coated grids.The suspensions were dried by slow evaporation at ambient temperature.
The SEM micrographs were obtained using Philips XL 30 CP equipped with EDX, SE, BSE, and Robinson detectors.The powder samples were studied in native form without coating by conductive metal layers.
The B.E.T. surface area was measured by means of a Micromeritics equipment, model ASAPS 2010, using nitrogen as adsorbate at the liquid nitrogen temperature.
The X-ray Photoelectron Spectra (XPS) were recorded with a Vacuum Generators, model Escalab 210, equipment by using a Mg K α with a pass energy resolution of 50 eV.The spectra of the powder sample were recorded at room temperature under a residual pressure lower than 10 −10 mbar.

Diffusion Reflectance Spectra and Band Gap Mesurements.
UV-vis spectra were recorded in the diffuse reflectance mode with a Shimadzu AV2101 UV-vis spectrophotometer equipped with an integrating sphere, BaSO 4 was used as reference.The spectra were taken in the diffuse reflectance mode (R) and transformed to a magnitude proportional to the extinction coefficient (K) through the Kubelka-Munk function, F(R α ).Band gap values were obtained from the plot of the modified Kubelka-Munk function, (F(R α )E) 1/2 , versus the energy of the absorbed light E.

EPR Spectra Measurements.
The formation of hydroxyl radicals upon irradiation of TiO 2 aqueous suspensions was investigated by EPR spin trapping technique using DMPO.The instantly prepared TiO 2 suspensions containing DMPO were carefully mixed by slight air stream and immediately transferred to a quartz flat cell (WG 808-Q, Wilmad-LabGlass, USA) optimized for the TE 102 cavity.The samples were irradiated at 295 K directly in the EPR resonator, and EPR spectra were recorded in situ using EMX X-band EPR spectrometer (Bruker, Germany).As an irradiation source an HPA 400/30S lamp (400 W, Philips) was used.The radiation flux was focused to obtain high intensity in the active part of TE 102 cavity.A Pyrex glass filter (thickness of 1 mm) was applied to eliminate radiation wavelengths below 300 nm.A GG 435 filter (Schott Glaswerke, Germany) was used to cut off wavelengths shorter than 435 nm.The stock suspensions were prepared by dispersing 10 mg TiO 2 powder in 10 ml distilled water, and then carefully homogenized in an ultrasound bath (60 s, Ultrasonic Compact Cleaner TESON 1, Tesla, Slovak Republic) before use.The concentration of DMPO stock solutions was 0.2 mol l −1 .The experimental EPR spectra acquisition and their simulations were carried out using WIN EPR and SimFonia standard programs (Bruker).[38].Two light sources were used, NARWA 400 W Blue Lamp to determine the visible light activity and the TESLA 125 W Black lamp to determine the activity in the UV region.The light spectrum of these sources is shown on     4(c) and 4(d)).We can see on this figure that the sample ground with ammonium carbamate consists of ∼10 nm randomly oriented anatase nanocrystals.Probably, amorphous part is present also in the aggregates; however, the particular nanoparticles are well crystalline even in the near surface regions.The electron diffraction pattern, in agreement with the XRD measurements, correspond well to anatase crystal structure.

Results and Discussion
The EDS analysis of the TIT 144 ground sample (Figure 5) shows that the material, except for TiO 2 , contains also low amount of Fe (0.3%) and S (0.6%).It is known that both these admixtures can shift the band gap energy to lower values, but concentrations used in experiments described in the literature are usually significantly higher [45].The concentrations found in the TIT 144 ground sample are too low to explain the band gap shift to visible region and presence of high photoactivity under visible light radiation, as was demonstrated by EPR measurements and kinetics of the 4-CP decomposition.

Band Gap Determination.
The results obtained for samples TIT 144 and TIT 144 ground are given on Figure 6.Significant shift of band gap width to lower values in comparison with the initial sample (from 3.2 eV for the International Journal of Photoenergy starting sample TIT 144 to 2.05 eV for the sample TIT 144 ground) was observed.The results show that after grinding is carried out with an additive like ammonium carbamate that behaves as an ammonia supplier, the light absorption edge moves to the visible light region.

XPS Measurements.
The XPS spectrum of the N 1s band is shown in Figure 7.The analysis of this spectrum allows concluding that only nitrogen present on the surface as NH x radicals has been detected.The percentage of nitrogen determined from the XPS spectrum is 1.25%.

EPR Measurements.
The formation of highly reactive radical species under irradiation with UV or visible light on the surface of a photocatalyst is significant evidence of its photoactiovity.However, upon irradiation of aerated TiO 2 aqueous suspensions, no paramagnetic signals of radicals are monitored using cw-EPR spectroscopy, due to the extremely short life-times of the reactive radical species generated.However, the photoproduced short-lived reactive radical intermediates may be evidenced by the spin trapping agents.The EPR spin trapping method involves the interaction of reactive short-lived free radicals with a diamagnetic EPRsilent compound (spin trap), via addition to a spin trap double bond, to produce a stable free radical product (spin adduct).The nature of the radical added can be identified after the analysis and simulation of the experimental EPR spectra [46].The 5,5-dimethyl-1-pyrroline Noxide (DMPO) molecule possesses sufficient photostability under the given experimental conditions (λ > 300 nm), since in a TiO 2 -free solution, negligible concentration of spin adducts are formed.The addition of photoactive titanium dioxide caused the generation of a four-line EPR signal upon irradiation, with spin Hamiltonian parameters characterized by the hyperfine splitting a N = 1.495 mT, a β H = 1 : 472 mT and a g-value of 2.0057, which are typical for hydroxyl radical adduct ( • DMPO-OH).The relative integral intensities of • DMPO-OH adduct were evaluated by the double-integration procedure of experimental EPR spectra [47].It is evident from Figure 8 showing the results of EPR measurement under filtered irradiation that intensive formation of OH • radical occur at irradiation of the sample by both UV (λ > 300 nm) as well as visible (λ > 435 nm) light.This result documents well the photoactivity of the material even in the visible light region.

Decomposition of 4-Chlorophenol.
The kinetic analysis of photocatalyzed decomposition of 4-chlorophenol has been carried out in terms of pseudofirst-order kinetics at the photostationary state.The 1st order reaction constant for photodegradation of 4-chlorophenol in water suspension was calculated for reaction in both visible and UV light.
From Figure 9 we can conclude that in the UV region the sample TIT 144 ground shows lowest activity with respect to both industrial photocatalysts tested, BDH, and P25.However, the measured photoactivity is still remarkably higher in comparison to other visible light active photocatalysts we have tested earlier [14,16].The photocatalytic activity measured under visible light shows reverse sequence  of photocatalysts, the most active is the ground sample, whereas the P25 photocatalyst shows only very low activity in the visible light region and the BDH photocatalyst (pure nanocrystalline anatase) shows no measurable photoactivity in the visible light region with respect to decomposition of 4-chlorophenol.
As it can be concluded from EDS and XPS measurements, the sample TIT 144 ground contains as main impurities traces of N and Fe compounds, which were not detected in the initial sample.It is well known that both these elements could be responsible for increase of the photocatalytic activity under the visible light radiation [12,13,48,49].
For the explanation of the effect of nitrogen, it is supposed that nitride ions replace the O 2− in the anatase  structure and create Fermi level in the electronic structure [12,13].To excite the electrons from the valence band into this level requires less energy than the excitation to the conduction band.
Different with the intrinsic excitation of TiO 2 under UV irradiation, the excitation behaviour of doped TiO 2 under visible irradiation is suggested involving the Fe 3+ ion.The t2g level of 3d orbital of Fe 3+ is above the valence band of TiO 2 , which can absorb a photon with a wavelength around 415 nm to produce a Fe 4+ ion and a TiO 2 conductive band electron (1).Conductive band electrons react with adsorbed O 2 to form O 2− (2), while Fe 4+ reacts with surface hydroxyl group to produce hydroxyl radical.Therefore, 4chlorophenol can be photodegraded even under the visible light irradiation [49].Another additional possible reason is the photo-induced electron transfer transition between Fe 3+ themselves, which results in a broad absorption around 500 nm and produces Fe 2+ and Fe 4+ .Fe 2+ and Fe 4+ then react with O 2 and surface hydroxyl group to produce O 2 − and hydroxyl radical [49], which degrade 4-chlorophenol, Fe 3+ + hν −→ Fe 4+ + ecb − (1) The observed high photoactivity of the studied sample can be therefore ascribed to synergic effect of small crystal size International Journal of Photoenergy (resulting from effect of ammonium carbamate) and surface doping of TiO 2 by Fe and N (originating from ammonium carbamate and abrasive wear of the steel from the grinding apparatus).

Conclusions
Addition of ammonium carbamate to nanocrystalline anatase subjected to high intensity grinding retards growth of anatase crystals and their phase transformation to rutile.This effect can be used for modification of properties of anatase by grinding without undesirable growth of nanocrystals and phase transition to rutile.The grinding product obtained after grinding nanoparticulate anatase TIT 144 for 1.7 hrs as well as the product obtained after annealing to 400

Figure 1 :
Figure 1: The light spectrum of sources used for determination of photocatalytic activity.

Figure 2 :
Figure 2: XRD diagrams of the starting TiO 2 sample-TIT 144 (a) and the samples ground for 1.6 hours without additives (b) and with ammonium carbamate-TIT 144 ground (c), respectively.

Figure 4 :
Figure 4: TEM micrographs (a, b) and electron diffraction pattern (c, d) of the sample TIT 144 ground.

Figure 1 .
In order to remove weak UV lines in the Blue Lamp spectrum, filter formed by 0.1 mole•L −1 aqueous solution of NaNO 2 was used.The light sources were placed close to the quartz tubes.By means of the ferric oxalate actinometry, an average light intensity entering the volume of 50-70 mL of the irradiated solution was determined as I 0 = 5.3 × 10 −5 Einstein•dm−3 ) and 3(b) show size and shape of the TiO 2 nanoparticle aggregates.The starting sample TIT 144 is formed by aggregates of spherical particles ∼10 μm size whereas the ground sample TIT 144 ground (Figure 3(b)) consists of irregular-shaped particles ranging from nm to μm size.The details of structure of nanoparticles of the sample TIT 144 ground can be seen on the TEM micrographs (Figures 4(a) and 4(b)) and ED patterns (Figures

Figure 5 :Figure 6 :
Figure 5: EDS analysis of the sample TIT 144 ground with ammonium carbamate addition.

Figure 7 :
Figure 7: XPS peak of N 1s in the sample TIT 144 ground with ammonium carbamate addition.

Table 1 :
Phase composition (X A fraction of anatase; X R , rutile; X II , TiO 2 -II phase), specific area (S) and crystal size (D A , anatase; D R , rutile and D II , TiO 2 -II phase) of the different phases present at the as prepared sample and after grinding for 1.6 hours without additives (1.6 h) and with ammonium carbamate additive (1.6 h AC), respectively.
• C. The detailed description of the photo reactor is given in [43,44]esults show that the starting sample (TIT 144) is constituted by a single phase of anatase.The grinding of anatase leads to formation of TiO 2 -II and rutile, what would support that a mechanochemical transformation of anatase into rutile through the high pressure phase (TiO 2 -II) takes place as suggested in the literature[43,44].On the other hand, the comparison of the phase composition and specific surface data of samples "1.6 h" and "1.6 h AC" (sample TIT 144 ground) in Table1suggest that ammonium carbamate inhibits the phase conversion and the cold welding of particles during the grinding.Moreover, the crystal size measurements included in Table1suggest that ammonium carbamate promotes the recrystallization during the grinding of anatase, contrary to what occurs during the grinding without additives.
3.1.XRDResults.Figure2shows the XRD diagram of the titania starting sample (TIT 144) together with the XRD diagrams of this sample after grinding 1.6 hours without additives and with 20% of ammonium carbamate as additive (TIT 144 ground), respectively.The crystal size and the phase composition determined from XRD data are included in Table1together with the specific surfaces determined from B.E.T. measurements.3.2.Electron Microscopy.The SEM micrographs on Figures3(a •C contain Fe and N as trace impurities.The annealed sample exhibits remarkable photocatalytic activity under visible light as it was demonstrated by both measurement of kinetics of photocatalytic decomposition of 4-chlorophenol and formation of OH • radical proved by EPR measurements.