Theoretical and Experimental Studies on the Visible Light Activity of TiO 2 Modiﬁed with Halide-Based Ionic Liquids

: Formation of a surface complex between organic molecules and TiO 2 is one of the possible strategies for the development of visible light-induced TiO 2 photoactivity. Herein, three ionic liquids (ILs) with the same cation and di ﬀ erent anions (1-butylpirydynium chloride / bromide / iodide) have been applied for the surface modiﬁcation of TiO 2 and to understand the role of anions in visible light-induced activity of ILs-TiO 2 systems. Photocatalytic screening tests (the measurements of phenol photodegradation reaction rate) revealed that anion type a ﬀ ected visible light activity ( λ > 420 nm) of TiO 2 obtained by the ILs-assisted solvothermal method. Density functional theory (DFT) calculations demonstrated that interactions between halogen anions and oxygen vacancies (OV) on the surface of the TiO 2 particles could be responsible for the speciﬁc wavelength-induced excitation and ﬁnally for the observed photoactivity of titania under visible light. Finally, our theoretical calculations have been proven by experiments using monochromatic light (the apparent quantum e ﬃ ciency was measured) and the properties of obtained samples were characterized using scanning electron microscopy (SEM), X-ray powder di ﬀ raction analysis (XRD), UV-Vis spectroscopy and X-ray photoelectron spectroscopy (XPS).


Introduction
Semiconductor photocatalysis, with a dominant role of TiO 2 as a representative catalyst, was, up to date, applied to a variety of environmental and energy problems. However, since titania absorbs wavelengths shorter than 400 nm (due to the large band gap), its practical application requires usage of UV irradiation sources. In this regard, a lot of effort was made to design and tailor physical, chemical and optical properties of TiO 2 to extend its absorption to the visible range. Various strategies, such as surface photosensitization [1], doping with heteroatoms [2], charge transfer (CT) metals [5] were employed to utilize a larger fraction of the solar spectrum in the processes involving the TiO2 photocatalyst.
Another, actually unexpected, way to improve TiO2 photoactivity under visible irradiation turned out to be application of ionic liquids (ILs) for the preparation of TiO2 microparticles [6]. These salts, with melting point below 100 °C, are usually composed of organic, asymmetric cations, and organic or inorganic anions. In contrary to other liquids, ILs are formed of supramolecular network of ions interacting via Coulombic forces, dipole−dipole and π-π interactions, hydrogen bonds, etc. [7,8]. These interactions, also present between IL moieties and titania precursor, affect size and surface architecture of TiO2 nano-and microparticles surrounded by the IL's protective layer, which in turn, modify titania photoreactivity [8,9].
In this regard, a controllable preparation of the functional TiO2 semiconductor employing ionic liquids is a new and effective approach [10,11]. Nevertheless, the mechanism of IL action and their role in the TiO2 photoactivity enhancement have not been fully explained yet. Application of various IL-assisted strategies to obtain the IL-TiO2 composites and characteristics of these materials revealed that ILs could improve the Vis response of this wide-band-gap semiconductor by: (i) doping of the crystalline structure with nonmetal elements (e.g., N, B, F) [12]; (ii) favoring oxygen vacancies (OV) and Ti 3+ species formation on the titania surface [13]; (iii) forming surface complex charge transfer [14] and (iv) affecting transport of photogenerated charges [15]. Moreover, new energy levels can be formed between TiO2, being an n-type semiconductor, and the halogen anion of IL [16]. In addition, the mechanism of the IL-TiO2 photoactivity is strictly dependent on the IL structure and its thermal stability (due to possible thermal decomposition of IL under synthesis conditions) [16].
The aim of this work is to further explain the mechanism of phenol photodegradation mediated by the IL-TiO2 microparticles, which were prepared in the presence of three halide-based 1butylpirydynium ionic liquids ([BPy][X]; [X]: Cl − , Br − , I − ). The theoretical model has been confirmed by experiments using monochromatic light.

Application of 1-butylpirydynium ionic liquids ([BPy][X]
; where [X] is Cl − , Br − or I − ) was dictated by the observation that 1-butylpirydynium chloride used for TiO2 preparation provided photocatalyst with tremendous visible light driven activity [16]. Therefore, in this study, TiO2 was synthesized by the solvothermal method in the presence of [BPy][X] containing chloride, bromide or iodide anions ( Figure 1).

Structure, Morphology and Optical Properties
To give more information about the IL-TiO2 photocatalysts used in this study basic material characteristics was performed. As shown in Figure 2, the X-ray diffraction analysis was applied to verify the crystalline phase of the obtained samples. A set of diffraction peaks at 2θ values near 25.3°, 37.7°, 48°, 54°, 55°, 63° and 68.5° were observed for all samples, which are attributed to the anatase phase formation. The analysis confirmed that the samples do not contain any impurities and anatase phase with high quality was formed. Ionic liquids addition did not cause any changes in the crystal structure of the obtained samples.

Structure, Morphology and Optical Properties
To give more information about the IL-TiO 2 photocatalysts used in this study basic material characteristics was performed. As shown in Figure 2, the X-ray diffraction analysis was applied to verify the crystalline phase of the obtained samples. A set of diffraction peaks at 2θ values near 25.3 • , 37.7 • , 48 • , 54 • , 55 • , 63 • and 68.5 • were observed for all samples, which are attributed to the anatase phase formation. The analysis confirmed that the samples do not contain any impurities and anatase Catalysts 2020, 10, 371 3 of 12 phase with high quality was formed. Ionic liquids addition did not cause any changes in the crystal structure of the obtained samples.
Catalysts 2020, 10, x FOR PEER REVIEW 3 of 13 Figure 2. The X-ray diffraction patterns for the IL-TiO2 and pristine TiO2 samples.
The surface morphology and microstructure of the reference TiO2 and IL-TiO2 samples were characterized by SEM ( Figure 3). It was demonstrated that pristine TiO2 as well as modified TiO2 had a spherical shape. For the samples prepared with addition of ionic liquids the dominant fraction were microparticles with average diameter from 1 to 2 µm, while in the case of reference TiO2, the largest contribution was particles with sizes 3-5 µm. Moreover, the microspheres of pristine TiO2 were more irregular in size and shape, compared to IL-assisted TiO2 samples. As it shown in Figure 3, the microparticles prepared with addition of ILs had smoother surface and more uniform particle size distribution. The average size of the microspheres was calculated based on the statistical size of at least one hundred particles in the case of each sample. It was revealed that samples [I] yielded mainly microparticles with diameter of 1-3 µm, with the highest contribution of 1-2 µm, in an amount of 70%, 45% and 58%, respectively. The percentage contribution of large particles, with diameter 3-5 µm was rarely reached for ILs-assisted samples and never exceeded 16%, however for the pristine TiO2 the amount of this fraction was 52%. The formation of the microspheres probably started with hydrolysis of the TBOT precursor, followed by the formation and growth of the TiO2 crystallites (nucleation). The XRD measurements (see Table S1) revealed that the average diameter, as well as, volume of crystallites of the IL-TiO2 photocatalysts were higher than that detected for pristine TiO2. These dimensions increased in the following order In the next step, the crystallites aggregated to create the spherical shape of the particles. Some small crystallites may detach from the particle surface and diffuse into solution because of the tendency to reduce the overall energy of the particle, which is called the Ostwald ripening mechanism. This mechanism is also related with the deposition of smaller particles on the larger crystals [17]. ILs as surface active agents may adsorb at the particle surface forming the surface layer, thereby limiting smaller particles deposition during Ostwald ripening, inhibiting further growth and agglomeration [10,16]. Growth of the particles is usually a diffusion-limited process, therefore a structure and content of the ionic liquid used in the reaction system influence TiO2 particles' growth [10,16,18]. Further, as shown in Figure 3, more uniform particle size distribution in case of the IL-TiO2 microparticles was probably also due to steric hindrance and ability of the alkyl substituents to take a variety of arrangements. The surface morphology and microstructure of the reference TiO 2 and IL-TiO 2 samples were characterized by SEM ( Figure 3). It was demonstrated that pristine TiO 2 as well as modified TiO 2 had a spherical shape. For the samples prepared with addition of ionic liquids the dominant fraction were microparticles with average diameter from 1 to 2 µm, while in the case of reference TiO 2 , the largest contribution was particles with sizes 3-5 µm. Moreover, the microspheres of pristine TiO 2 were more irregular in size and shape, compared to IL-assisted TiO 2 samples. As it shown in Figure 3, the microparticles prepared with addition of ILs had smoother surface and more uniform particle size distribution. The average size of the microspheres was calculated based on the statistical size of at least one hundred particles in the case of each sample. It was revealed that samples [I] yielded mainly microparticles with diameter of 1-3 µm, with the highest contribution of 1-2 µm, in an amount of 70%, 45% and 58%, respectively. The percentage contribution of large particles, with diameter 3-5 µm was rarely reached for ILs-assisted samples and never exceeded 16%, however for the pristine TiO 2 the amount of this fraction was 52%. The formation of the microspheres probably started with hydrolysis of the TBOT precursor, followed by the formation and growth of the TiO 2 crystallites (nucleation). The XRD measurements (see Table S1) revealed that the average diameter, as well as, volume of crystallites of the IL-TiO 2 photocatalysts were higher than that detected for pristine TiO 2 . These dimensions increased in the following order In the next step, the crystallites aggregated to create the spherical shape of the particles. Some small crystallites may detach from the particle surface and diffuse into solution because of the tendency to reduce the overall energy of the particle, which is called the Ostwald ripening mechanism. This mechanism is also related with the deposition of smaller particles on the larger crystals [17]. ILs as surface active agents may adsorb at the particle surface forming the surface layer, thereby limiting smaller particles deposition during Ostwald ripening, inhibiting further growth and agglomeration [10,16]. Growth of the particles is usually a diffusion-limited process, therefore a structure and content of the ionic liquid used in the reaction system influence TiO 2 particles' growth [10,16,18]. Further, as shown in Figure 3, more uniform particle size distribution in case of the IL-TiO 2 microparticles was probably also due to steric hindrance and ability of the alkyl substituents to take a variety of arrangements. The UV-Vis absorption spectra of the photocatalysts prepared with addition of different ILs as well as pristine TiO2 are presented in Figure 4. Pristine TiO2 absorbed irradiation mainly in the UV range (λ < 400 nm). Addition of ionic liquids to the reaction environment resulted in noticeably increased absorption in the visible irradiation range (λ > 400 nm) when compared with pristine TiO2. Moreover, depending on the type of added ionic liquid, absorption differed in the visible light range. Generally, the enhancement of Vis light absorption was in the following order:  The UV-Vis absorption spectra of the photocatalysts prepared with addition of different ILs as well as pristine TiO 2 are presented in Figure 4. Pristine TiO 2 absorbed irradiation mainly in the UV range (λ < 400 nm). Addition of ionic liquids to the reaction environment resulted in noticeably increased absorption in the visible irradiation range (λ > 400 nm) when compared with pristine TiO 2 . Moreover, depending on the type of added ionic liquid, absorption differed in the visible light range. Generally, the enhancement of Vis light absorption was in the following order: The Ti 2p XPS spectra revealed the signals attributed to Ti3+ species to increase for IL-assisted samples, which, in turn, indicate a higher content of oxygen vacancies (see Table 1, Figure S1) in the halogen-based ILs-TiO 2 particles. For example, for the TiO 2 _[BPy][Cl] sample, the amount of Ti 3+ states was 7.36 at.% while that for pristine TiO2 only 2.41 at.%. This finding prompted us to hypothesize that oxygen vacancies could play a basic role in the mechanism of excitation of IL-TiO 2 under visible irradiation, which will be discussed in the next section of this article. Moreover, the XPS results demonstrate that IL anions form a protecting layer surrounding the TiO 2 particle as a result of interaction with the TiO 2 surface (see Table S2). It was revealed that C-Nx bonds were present in the Catalysts 2020, 10, 371 5 of 12 IL-modified TiO 2 samples. Additionally, the presence of individual elements (Cl, Br, I) on the surface of the IL-TiO 2 samples was confirmed (Table S2). The Ti 2p XPS spectra revealed the signals attributed to Ti3+ species to increase for IL-assisted samples, which, in turn, indicate a higher content of oxygen vacancies (see Table 1, Figure S1) in the halogen-based ILs-TiO2 particles. For example, for the TiO2_[BPy][Cl] sample, the amount of Ti 3+ states was 7.36 at.% while that for pristine TiO2 only 2.41 at.%. This finding prompted us to hypothesize that oxygen vacancies could play a basic role in the mechanism of excitation of IL-TiO2 under visible irradiation, which will be discussed in the next section of this article. Moreover, the XPS results demonstrate that IL anions form a protecting layer surrounding the TiO2 particle as a result of interaction with the TiO2 surface (see Table S2). It was revealed that C-Nx bonds were present in the IL-modified TiO2 samples. Additionally, the presence of individual elements (Cl, Br, I) on the surface of the IL-TiO2 samples was confirmed (Table S2).    Figure S2). At this stage, it was revealed that addition of ionic liquid to the reaction environment caused increased photocatalytic activity, as shown based on the photodegradation reaction rate, in comparison to pristine TiO 2 (prepared in the same way but without ILs). To explain the effect of the anion type on the TiO 2 _[BPy][X] photoactivity we proposed a computation model in which we proved that halogen-oxygen vacancies are responsible for TiO 2 excitation under Vis irradiation. Moreover, the formation of oxygen vacancies on TiO 2 is equivalent to the occurrence of Ti 3+ centers that, in consequence, leads to the creation of unpaired electrons, which could form donor levels in the electronic structure of TiO 2 [19]. Described in detail, in this computational model a surface complex between the halide anion and TiO 2 vacancy plays a central role. Unlike halide anions, both TiO 2 particles and 1-butylpirydynium cation do not absorb in the visible spectrum. Therefore, one might suspect that the presence of halide anions makes TiO 2 photoactive in the visible light. In order to prove this assumption, we employed the described in the literature OV in an anatase particle, i.e., Ti 35 O 69 [20]. In Figure 5, the spin density originating from the two excess electrons associated with an oxygen vacancy are compared in Ti 35 O 70 and in the same particle where the vacancy interacts with the halogen anion and oxygen molecule. To explain the effect of the anion type on the TiO2_[BPy][X] photoactivity we proposed a computation model in which we proved that halogen-oxygen vacancies are responsible for TiO2 excitation under Vis irradiation. Moreover, the formation of oxygen vacancies on TiO2 is equivalent to the occurrence of Ti 3+ centers that, in consequence, leads to the creation of unpaired electrons, which could form donor levels in the electronic structure of TiO2 [19]. Described in detail, in this computational model a surface complex between the halide anion and TiO2 vacancy plays a central role. Unlike halide anions, both TiO2 particles and 1-butylpirydynium cation do not absorb in the visible spectrum. Therefore, one might suspect that the presence of halide anions makes TiO2 photoactive in the visible light. In order to prove this assumption, we employed the described in the literature OV in an anatase particle, i.e., Ti35O69 [20]. In Figure 5, the spin density originating from the two excess electrons associated with an oxygen vacancy are compared in Ti35O70 and in the same particle where the vacancy interacts with the halogen anion and oxygen molecule.  Due to this interaction spin density is partially transferred to the particle fragment distant to the position of the (OV)X − ···O 2 complex (see Figure 5; cf. A with B, C or D). Moreover, the excess electron is removed from the Ti 3+ cation which interacts with the halogen anion that suggests the oxidation of titanium cation to Ti 4+ due to complex formation. It is worth emphasizing that the halogen anions do not form stable complexes with O 2 in solution. For instance, the distance between the bromide anion and the closest of oxygen atom in the O 2 molecule amounts to as much as 4.19 Å, while the interaction energy is as low as 0.2 kcal/mol for the Br − . . . O 2 system optimized at the PBE0/def2TZVP level and using the PCM model of water. Therefore, one can conclude that formation of the Br − . . . O 2 complex, an indispensable condition for the photoinduced electron transfer between Br − and O 2 , is negligible in solution. In contrast, interactions between an oxygen vacancy, Br − and O 2 make the distance between the halide anion and oxygen short enough to enable the photoinduced electron transfer to be efficient.
Indeed, regarding the abovementioned stabilization energy for the Br − . . . O 2 complex, the interaction energy corresponding to the absorption of a halide anion and oxygen molecule on the oxygen vacancy is much larger. As indicated by the data gathered in Table S3, depending on the halide anion, this interaction energy spans the range from 46 to 60 kcal/mol. Interestingly, the absorption energies calculated for a larger cluster, Ti 4 O 12 , are within 2 kcal/mol identical with those for Ti 2 O 6 (see Table S3), which suggests that already the smaller model should deliver reliable results. Moreover, the gas phase interaction energies calculated at the PM6 level for the full Ti 35 O 69 model (see Table S3) are similar to those obtained for the smaller models of oxygen vacancy. On the on hand, this fact confirms the validity of the Ti 2 O 6 model and on the other hand shows that the local geometry of the vacancy being in the closest vicinity to the halide anion and oxygen decides on the adsorption characteristics.
The ground state spin distribution suggests, thus, that due to the formation of the abovementioned complex, the system becomes prepared for the excited state electron transfer between a halide anion and molecular oxygen. This observation prompted us to propose a simplified vacancy model (in order to enable more accurate DFT model to be employed) comprising just two titanium cations, Ti 3+ (having an unpaired electron interacting with O 2 and Ti 4+ interacting with X − , see Figure 6). be efficient.
Indeed, regarding the abovementioned stabilization energy for the Br −… O2 complex, the interaction energy corresponding to the absorption of a halide anion and oxygen molecule on the oxygen vacancy is much larger. As indicated by the data gathered in Table S3, depending on the halide anion, this interaction energy spans the range from 46 to 60 kcal/mol. Interestingly, the absorption energies calculated for a larger cluster, Ti4O12, are within 2 kcal/mol identical with those for Ti2O6 (see Table S3), which suggests that already the smaller model should deliver reliable results. Moreover, the gas phase interaction energies calculated at the PM6 level for the full Ti35O69 model (see Table S3) are similar to those obtained for the smaller models of oxygen vacancy. On the on hand, this fact confirms the validity of the Ti2O6 model and on the other hand shows that the local geometry of the vacancy being in the closest vicinity to the halide anion and oxygen decides on the adsorption characteristics.
The ground state spin distribution suggests, thus, that due to the formation of the abovementioned complex, the system becomes prepared for the excited state electron transfer between a halide anion and molecular oxygen. This observation prompted us to propose a simplified vacancy model (in order to enable more accurate DFT model to be employed) comprising just two titanium cations, Ti 3+ (having an unpaired electron interacting with O2 and Ti 4+ interacting with X − , see Figure 6). In order to demonstrate that the cluster of Ti2O4 size properly describes the system under consideration we also analyzed Ti4O12 (cut out from the Ti35O69 model of oxygen vacancy (see Figure In order to demonstrate that the cluster of Ti 2 O 4 size properly describes the system under consideration we also analyzed Ti 4 O 12 (cut out from the Ti 35 O 69 model of oxygen vacancy (see Figure 5) which is two times larger than the original one. The TD-DFT calculations for Ti 4 O 12 , performed at the PBE0/Def2TZVP/PCM level of theory, resulted in a similar transition to those obtained for Ti 2 O 6 . Namely, the leading transitions in the long wavelength region of absorption occur at similar values and similarly correspond to electron transfer from the halogen anion to the adsorbed oxygen molecule (see Figure S3 were the transitions in both clusters are compared).
For such models, constrained DFT optimizations were performed. Namely, the Ti 2 O 6 fragment of a particle was fixed (a stiff structure of anatase particle does not deform because of the formation of a surface complex with oxygen and the X − anion) at the geometry obtained by Kim et al. [20], whereas all degrees of freedom related to X − and O 2 were fully optimized at the PBE0/Def2-TZVP level and the PCM model of water [21,22]. Then, the electronic transitions were obtained at the TD-PBE0/Def2-TZVP/PCM level. In the long wavelength region of absorption, the leading transitions (>90%, see Figure 6) correspond to the excitation of the electron from the lone pair in the halogen anion to the π* orbital localized in the oxygen molecule (nX − → π*O 2 ). These transitions were calculated at 370/403, 448 and 582 nm for the complexes containing Cl − , Br − and I − anions, respectively (see Figure 6). This type of excitation transfers an electron form X − to the adsorbed oxygen molecule. Since a monopole-monopole interaction between Ti 4+ and X − , a major force binding X − to Ti 4+ , does not exist in such an excited state, the transfer of the X atom to water is facilitated, which makes the formation of O 2 •− irreversible.
In order to confirm the theoretical calculations, the apparent quantum efficiency of the photocatalysts was estimated. The apparent quantum efficiency reflects the ratio of the rate of Catalysts 2020, 10, 371 8 of 12 the electron consumption from the initial rates of benzoquinone formation (first product of phenol oxidation) to the flux of incident photons. For each sample, the apparent quantum efficiency (AQE) was measured at three different wavelengths, i.e., at the absorption maxima predicted by the DFT model (403, 448 and 582 nm for Cl − , Br − and I − , respectively, see Figure 6) and additionally for the wavelengths longer and shorter by 30 nm with respect to the mentioned above maxima. Additionally, the AQE for pristine TiO 2 was examined. The AQE for the TiO 2 samples modified with halide-based ionic liquids is presented in Table 3. In the case of the photocatalysts containing chloride or bromide anions, AQE decreases with the increasing wavelength, which is typical for titania [23,24]. Described in detail, the relatively high photoactivity of TiO 2 _[Bpy][Cl] for wavelength of 370 nm (Table 3) resulted from: (i) the fact that pristine TiO 2 , as a wide band-gap semiconductor, is excited itself mainly under the UV region of irradiation and, (ii) additionally from the existence of the absorbing complex at λ = 370 nm, which is consistent with theoretical studies (Figure 6). The activity of photocatalysts tested at longer wavelengths (403, 420, 430 nm (Table 3)) was mainly due to the existence of a surface complex, which was predicted by theoretical calculations. It is worth mentioning that pristine TiO 2 was still slightly photoactive under visible range of irradiation (see Figure S4). This might be due to existence of oxygen vacancies (Ti 3+ states) or with the presence of impurities on the TiO 2 surface formed during the synthesis. Action spectra measurements of phenol photodegradation over pristine TiO 2 as well as IL-assisted TiO 2 are presented in Figure S4. However, the activity above 450 nm (Br − /I − modified samples) should be solely caused by the presence of halogen, since only defect-rich and modified (surface and doped) titania samples possess such activity. Interestingly, for the sample synthesized in the presence of iodide anion ([BPy][I]), the maximal AQE of 1.31% was measured at 582 nm (0.87% at 550 nm and 0.47% at 610 nm (see Table 2)), which agrees well with the results of our quantum chemical calculations (see Figure 6) and indicates the evident effect of iodide anion on the TiO 2 photoactivity. In case of such long wavelengths, the intrinsic activity of titania cannot affect the activity of TiO 2 _[BPy] [I].
Additionally, in order to gain insight into the ionic liquid mode of action, the thermal decomposition of 1-butylpirydium chloride under solvothermal reaction was investigated by using HPLC. The decomposition experiment was performed under the identical conditions as the synthesis of IL-TiO 2 samples i.e., 180 • C, 24 h. It was found that the decomposition efficiency of the [BPy] [Cl] was 50% [25]. The HPLC results are consistent with the XPS analysis, namely Ti-Nx species formation (Table S2). On the basis of the XPS results, we can deduce that N atoms were probably introduced into the TiO 2 lattice during the solvothermal synthesis. Since nitrogen has a comparable atomic size as oxygen, small ionization energy and high stability, the N doping can be an effective way to enhance the absorption of TiO 2 under the visible range of irradiation. It is related to change in the banding structure of TiO 2 due to nitrogen atoms' substitution and formation of surface oxygen vacancies. More details about different ILs decomposition in the above mentioned conditions were described in our publication in which we developed computational methodology allowing for prediction properties of newly synthesized IL-TiO 2 materials before their synthesis [25].

Samples Preparation Method
The IL-assisted TiO 2 microparticles were synthesized by the procedure developed and optimized in our previous work [10]. Titanium(IV) butoxide (TBOT, Sigma-Aldrich, St. Louis, MO, USA) used as a precursor of TiO 2 was dissolved in an absolute ethanol (Avantor Performance Materials Poland S.A., Gliwice, Poland) with the purity 99.9%. Then, hydrochloric acid (Sigma-Aldrich, St. Louis, MO, USA) and distilled water were added. In the next step, the 1-butylpirydynium (Iolitec, Heilbronn, Germany) ionic liquid with different anions (chloride, bromide, iodide) was added in molar ratio 1:3 (IL:TBOT). The amount of IL added to the reaction system was selected based on our previous results [16]. The obtained reaction mixture was placed in a Teflon-lined stainless-steel autoclave and kept at 180 • C for 24 h. After the end of the set time, the reactor was cooled down at room temperature. The product was washed with ethanol (Avantor Performance Materials Poland S.A., Gliwice, Poland) and deionized water, dried at 50 • C for 12 h, and finally calcined at 200 • C for 2 h. For comparison, reference TiO 2 was synthesized using the same procedure without addition of IL.

Characteristic of the Material
The morphology (particle size and shape) of the IL-TiO 2 microparticles was measured by a scanning electron microscope (SEM, JEOL JSM-7610F, Jeol Ltd., Tokyo, Japan) working in a high vacuum mode. Powder X-ray diffraction (XRD) studies of the photocatalyst samples were carried out using a Rigaku MiniFlex 600 XRD system (Rigaku Corporation, Tokyo, Japan) equipped an X-ray generator with a copper target (operated at 40 kV and 40 mA). A Nicolet Evolution 220 UV-vis spectrophotometer (ThermoFisher Scientific, Waltham, MA., USA) was used to obtain the diffuse reflectance UV-vis absorption spectra of the samples, for which the baseline was performed using BaSO 4 . The surface composition was analysed by X-ray photoelectron spectroscopy (XPS) using a PHI 5000 VersaProbe TM spectrometer (ULVAC-PHI, Chigasaki, Japan) with monochromatic Al Kα irradiation (hν = 1486.6 eV) from an X-ray source operating at 100 µm spot size, 25 W and 15 kV. The high-resolution (HR) XPS spectra were recorded with the hemispherical analyser at the pass energy of 23.5 eV, the energy step size of 0.1 eV and the photoelectron take off angle 45 • with respect to the surface plane. Casa XPS software (v.2.3.19, Casa Software Ltd, Wilmslow, United Kigdom) was used to evaluate the XPS data. The binding energy (BE) scale of all detected spectra was referenced by setting the BE of C 1s signal to 284.8 eV.

Photocatalytic Testes (Screening Tests)
To investigate the photocatalytic activity of the composites of TiO 2 prepared in the presence of IL, the decomposition rate of phenol (Avantor Performance Materials Poland S.A., Gliwice, Poland) 0.21 mmol·dm −3 in an aqueous solution was measured. The mixture of 0.125 g of the photocatalysts in 25 mL of phenol solution was stirred using a magnetic stirrer in a cylindrical reactor with a quartz window. The aeration of the suspension (5 dm 3 /h) was maintained prior to and during the photocatalytic process. The photoirradiation was provided by a 1000 W Xenon lamp (6271H, Oriel), capable of emitting both the UV and visible light. For all measurements, an optical filter was used, with wavelengths λ > 420 nm. During the illumination, aliquots of the aqueous suspension with a volume of 1 cm 3 were successively sampled. Removal of the fine particles of the photocatalyst was performed by filtering the samples through a 0.2 µm syringe filter. To determine the phenol concentration, the colorimetric method (λ max = 480 nm) was used with the UV-vis spectrophotometer (Nicolet Evolution 220, ThermoFisher Scientific, Waltham, MA., USA). The blind test (in the absence of the photocatalysts or illumination) proceeded the photocatalytic degradation runs. The absence of either the photocatalyst or illumination resulted in lack of phenol degradation.

Irradiations (Action Spectral Measurements)
Each photocatalyst powder (15 mg) was suspended in the aqueous solution (3.0 mL) of phenol (C 0 = 20 mg·L −1 ) and placed in a rectangular quartz cell (10 mm 2 and 50 mm in height), then irradiated at monochromatic wavelengths for 60 min using a diffraction grating-type illuminator (Jasco, CRM-FD) equipped with a 300 W Xenon lamp (Hamamatsu, C2578-02). The light intensity was measured by an optical power meter (HIOKI 3664). The light intensity varied from 5.231 to 14.113 mW/cm 2 depending on the wavelength. During the experiments, the reaction mixtures were continuously stirred, and a portion (0.2 mL) of the reaction mixture was withdrawn every 20 min with a syringe, filtered (∅ = 0.2 µm, Whatman, Mini-UniPrep) to remove photocatalyst particles and subjected to a HPLC analysis. The HPLC Shimadzu LC-6A system (Shimadzu, Kyoto, Japan) was equipped with a WAKOSIL-II SC18 AR reversed-phase column (250 × 4.6 mm) and the UV detector (254 nm). A flow rate of a mobile phase composed of acetonitrile (29.5%), water (70%) and phosphoric acid (0.5%) was 1 mL·min −1 .

Computations
The TD-DFT (time-dependent density functional theory) calculations were performed with the use of PBE0 functional and Def2-TZVP basis set on the geometries optimized at the same level of theory. PBE0 functional [26] is known to produce reliable results [27] for the excited states while the Def2-TZVP basis set was chosen since it is available for all three halogens studied in the current work. The water environment was simulated by the PCM model [21,22]. All calculations were carried out with the Gaussian09 package (ver. D.01, Gaussian, Inc., Wallingford, CT, USA) [28].

Conclusions
In summary, based on our findings, we demonstrated that the improved, under visible irradiation, photoactivity of TiO 2 modified with the selected ILs, originates from the interactions of the IL's halogen anion (Cl − /Br − /I − ) and molecular oxygen with the TiO 2 surface. These interactions lead to the formation of a surface complex that allows photoinduced charge transfer (CT) to be accomplished. We concluded that TiO 2 enables the adsorption of halides on the oxygen vacancies. Based on our findings, we proposed a phenol photodegradation mechanism in which the O 2 •− radical anions are formed in the primary photochemical reaction. The latter process constitutes a photoinduced electron transfer (ET) between the halogen anion and adsorbed O 2 . Indeed, using the TD-DFT method, we demonstrated the presence of electronic transition of significant oscillator strength at specific wavelength (370 and 403 nm for Cl − , 448 nm for Br − and 582 nm for I − ) in the system comprising a model of OV and the adsorbed Cl − /Br − /I − and O 2 ((OV)X − ···O 2 ). The calculated wavelengths correspond well to the experimentally determined spectral region of the TiO 2 photoreactivity. Additionally, we concluded that the nitrogen doping can also take part in the mechanism of phenol decomposition under visible irradiation; however, the first mechanism likely dominates. In this regard, the mechanism of TiO 2 photoactivity enhancement is still unclear and needs further research, especially the experimental conformation. Explanation of the photodegradation mechanism is crucial to understand the mode of action of this materials and design new with improved activity.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4344/10/4/371/s1, Figure S1. XPS HR spectra of Ti 2p of the TiO 2 photocatalysts modified with halide-based ionic liquids. Figure Figure S4: Action spectra for the photocatalytic degradation of phenol in the presence of pristine TiO 2 and TiO 2 obtained using ionic liquid containing various anions.,