Pickering Emulsions of Fluorinated TiO2: A New Route for Intensification of Photocatalytic Degradation of Nitrobenzene

Fluorination of the TiO2 surface has been often reported as a tool to increase the photocatalytic efficiency due to the beneficial effects in terms of production of oxidizing radicals. Moreover, it is shown that the unique amphiphilic properties of the fluorinated TiO2 (TiO2-F) surface allow one to use this material as a stabilizer for the formulation of Pickering emulsions of poorly soluble pollutants such as nitrobenzene (NB) in water. The emulsions have been characterized in terms of size of the droplets, type of emulsion, possibility of phase inversion, contact angle measurements, and optical microscopy. The emulsified system presents micrometer-sized droplets of pollutant surrounded by the TiO2-F photocatalyst. Consequently, the system can be considered to be composed of microreactors for the degradation of the pollutant, which maximize the contact area between the photocatalyst and substrate. The enhanced photocatalytic activity of TiO2-F was confirmed in the present paper as the apparent rate constants of NB photodegradation were 16 × 10–3 and 12 × 10–3 min–1 for fluorinated and bare TiO2, respectively. At NB concentrations largely exceeding its solubility, the rate constant was 0.04 × 10–3 min–1 in the presence of both TiO2 and TiO2-F. However, unlike TiO2, TiO2-F stabilized NB/water emulsions and, under these conditions, the efficiency of NB photocatalytic degradation in the emulsified system was ca. 18 times higher than in the nonemulsified one. This result is relevant also in terms of practical applications because it opens the route to one-pot treatments of biphasic polluted streams without the need of preliminary physical separation treatments.


■ INTRODUCTION
Nitrobenzene (NB) is a toxic and biopersistent aromatic compound. 1 It affects the central nervous and cardiovascular systems, is suspected to be carcinogenic, 2 and poses serious concerns for its effects in the environment. 3−6 However, NB is implemented in many industrial processes for the production of perfumes, resins, dyes, explosives, and pesticides. 7−9 Therefore, it is often present not only in industrial wastewater streams but also in surface and ground water 10 because its high chemical stability and low water solubility lead to its accumulation in the environment. Conventional purification and decontamination rely on physical treatments such as separation techniques. 11, 12 These methods are generally expensive, are poorly efficient, and afford only the removal of the pollutants without chemical degradation.
Photocatalysis is a promising advanced oxidation process that has been often proposed for water remediation applications as it allows oxidization of almost all harmful organic species. Upon absorption of light of suitable energy, reactive oxygen species (ROS) possessing high oxidizing ability are generated. The transfer of the photocatalytic process from the laboratory to real applications is a current concern that deserves a great deal of research efforts. The photodegradation of poorly water-soluble substrates faces some problems whose solution still requires basic research. In these cases, polluted water cannot be directly used as in most photocatalytic applications 13 because of the low concentration of the dissolved substrate and its low affinity with the generally hydrophilic surface of the photocatalyst. These factors negatively affect the quantum efficiency of the process and, consequently, its economic feasibility. The use of different solvents has been proposed to overcome these problems in cases where photocatalysis was used for organic synthesis purposes. 14 For instance, the degradation of polycyclic aromatic hydrocarbons (PAHs) has been recently carried out in the green solvent dimethyl carbonate in the presence of a minimum amount of water necessary to induce the formation of oxidizing radical species. Such a shift of solvent is better restricted to the synthesis of compounds with high added value. Using solvents other than water for environmental remediation faces hard problems of costs and management, especially in view of large-scale applications.
A different approach consists of the formation of emulsions of the hydrophobic pollutant dispersed in water in which the photocatalyst can also act as a stabilizer of the emulsion. In this way, the preliminary physical separation of the pollutant can be by-passed and the degradation reaction can take place inside microreactor-like droplets of the substrate surrounded (and stabilized) by the photocatalyst itself. This configuration is beneficial in terms of photocatalytic activity as it creates a larger contact area between the catalyst and pollutant, optimizes the availability of the active sites, and offers all of the advantages of working in micrometer-sized reactors. 15 The organic pollutant plays the role of the "oil" in the emulsion. The emulsion can be obtained by applying external energy either in the form of mechanical stirring or sonication. When the emulsion is stabilized by solid nanoparticles located at the oil−water interface, a "Pickering emulsion" 16 is obtained.
Pickering emulsions can be of the oil-in-water (O/W), water-in-oil (W/O), or even multiple types. 17−19 They possess the basic properties of classical emulsions stabilized by soluble surfactants (emulsifiers). 20 However, the stabilization produced by solid particles confers specific properties. The high resistance to coalescence is the main benefit. 21−23 Furthermore, Pickering emulsions are stable in a wide range of experimental conditions; in particular, they are not very affected by temperature variations, changes in the composition of the oil phase, and the pH of the aqueous phase. 24 This makes them useful in pharmaceutical, 25,26 cosmetic, 27 and food industries. 28,29 In 2010, the pioneering work by Crossley et al. 30 paved the way for catalytic reactions in Pickering emulsions. 31−38 However, only a few examples of photocatalytic reactions carried out in Pickering emulsions have been reported. 39−41 As far as we know, fluorinated TiO 2 has never been proposed for that purpose, while it is instead widely studied as a photocatalyst because of its higher photoactivity than pure TiO 2 . Its high photoactivity has been mainly attributed to (i) the presence of a defective surface, which increases the lifetime of the photogenerated charges, 42 (ii) the greater production of highly oxidizing hydroxyl radicals due to the better availability of photogenerated holes, 43 and (iii) the possible production of singlet oxygen. 44 Otherwise, the wettability by oils and water and the dynamics of water molecules on the fluorinated TiO 2 surface have been poorly studied. It has been proposed, for example, that fluorination gives rise to an increase in the Lewis acidic strength of Ti 4+ surface sites, which retain adsorbed water molecules, 45 thereby increasing the hydrophilic character of the surface. Similar conclusions have also been reached by other authors. 46 UV irradiation increases the polarity of the surface. 47 However, Mino et al. 48 recently revealed the complexity of the topic by showing evidence of different behaviors of various exposed crystalline facets and the different contributions of bulk and surface doping. Interestingly, this study disclosed that the presence of terminal fluoride groups on the {001} facet of TiO 2 decreased the hydrophilicity of the material by ca. two times. Therefore, the physical chemistry of fluorinated TiO 2 in Pickering emulsions is not trivial from a scientific point of view due to the complex interactions of its surface with water and oils. 49 We recently reported on the changes induced by surface fluorination on the morphological, structural, surface, and electronic features of TiO 2 . 50 While surface fluorination does not affect the bulk structural properties and the morphology of TiO 2 nanoparticles, it induces significant changes in terms of surface and electronic properties. In particular, the surface charge turns more negative upon fluorination, and supplementary intraband-gap energy states located up to 1.3 eV above the valence band are generated by the local interaction of chemisorbed fluorine atoms with the semiconducting structure of TiO 2 . The correlation between the optoelectronic features and the photocatalytic activity has been discussed. 50 The unique surface properties of fluorinated TiO 2 deserve further investigations and can be exploited for specific applications. For instance, their influence on the selectivity of the industrially relevant photocatalytic synthesis of highadded-value compounds has recently been demonstrated. 51 This paper reports the influence of surface fluorination of a nanostructured sol−gel-derived TiO 2 material on the stability of nitrobenzene/water emulsions and on the consequent efficiency of the photocatalytic degradation of nitrobenzene under UV irradiation. Notably, the possibility to treat in onepot biphasic polluted streams, without the need of costly physical separation processes upstream, is of practical relevance.

■ EXPERIMENTAL SECTION
Preparation of the Photocatalysts. TiO 2 nanoparticles were synthesized by using a facile sol−gel technique. Nineteen milliliters of titanium tetraisopropoxide (Ti[OCH(CH 3 ) 2 ] 4 , 97%, Alfa Aesar) was added to 4 mL of methanol (MeOH, 99.9%, Aldrich) and the obtained solution was sonicated in an ultrasonic bath (Elma, T460/ H, 35 kHz and 170 W). The hydrolysis process was then performed by adding dropwise 74 mL of deionized water under reflux and magnetic stirring. The obtained white gel was filtered and washed several times using ethanol and deionized water. The resulting powder was dried at 100°C for 18 h to evaporate water and the volatile organic compounds. Finally, the powder, labeled simply as TiO 2 , was calcined in a muffle furnace at 400°C for 4 h. Fluorination of TiO 2 was carried out by dispersing 1 g of the synthesized TiO 2 nanoparticles into 50 mL of 4% sodium fluoride (NaF, p.a., Aldrich) aqueous solution at pH 3.2 (obtained by addition of HNO 3 ) and stirring the suspension at room temperature for 48 h. The nanoparticles were separated by centrifugation, washed several times with HNO 3 aqueous solutions at pH 3.2, and finally dried at 80°C for 2 h. The resulting powder was labeled as TiO 2 -F.
The amount of fluorine in samples was determined by means of ion chromatography after digestion dissolution of the TiO 2 material by fusion with potassium hydrogen sulfate and extraction of the melt with dilute sulfuric acid. 52 Collected samples were then analyzed for F − anions using ion chromatography (930 Compact IC Flex, Metrohm, Switzerland) equipped with a chemical suppressor and conductivity detection. Elution solvent was 8 mmol·L −1 sodium carbonate (Fisher Scientific, Illkirch, France) in ultrapure water. A Metrosep A Supp 5250/4.0 column with an adequate precolumn at a temperature of 35°C was used. The calibration curve was linear in the range from 0.06 to 2000 μmol·kg −1 (R 2 = 0.999).
Preparation of Pickering Emulsions and Characterization. Nitrobenzene NB (C 6 H 5 NO 2 , 95%, Aldrich; solubility in water, 0.19 wt % at 20°C) constituted the oil phase of the Pickering emulsions. In the beginning, aqueous dispersions of 1 wt % TiO 2 or TiO 2 -F were obtained by means of an UltraTurrax T25 rotor-stator device equipped with a S25N18G shaft (IKA, Germany) rotating at 22000 rpm for 5 min to de-aggregate the particles. Thereafter, NB (up to 20 Langmuir pubs.acs.org/Langmuir Article vol %) was added to the aqueous dispersion and stirred during 5 min with an UltraTurrax T25 under the same conditions. The size distribution of the TiO 2 nanoparticles was determined by dynamic light scattering using a Zetasizer NanoZS instrument (Malvern, U.K.). Emulsions droplets were observed by optical microscopy using a Leica DMLM (Germany) microscope equipped with a video camera. The concentrated emulsions were diluted and spread between a glass plate and cover slip for microscopic observation in transmission mode and image analysis. The size distribution was determined using analysis image software AnalySIS.
Droplet size measurements of O/W emulsions were carried out at small-angle light scattering using a Mastersizer 3000 instrument (Malvern, U.K.). The refractive indexes of the optical model used for data processing were 1.332 for water and 1.552 for NB.
The electrical conductivity of emulsions was measured with a Radiometer CDM 210 conductivity meter at 25°C.
Contact angles of NB/photocatalyst/water were measured by the sessile drop method on the flat photocatalyst surface using the Drop Shape Analysis System DSA10Mk2 (Kruss, Germany). A pellet of the photocatalyst was pressed under 7 kPa pressure. A drop of NB (volume, 5−6 μL) was deposited on the surface of the pellet immersed in water with a syringe needle and the drop picture was recorded with a CCD camera.
Photocatalytic Experiments. The photocatalytic experiments were carried out in a cylindrical Pyrex batch photoreactor (diameter: 10 cm) equipped with a mechanical stirrer rotating at 500 rpm. Emulsion (160 mL) containing 0.1 vol % NB and 1 wt % catalyst was illuminated by a PLL lamp UVA (8 mW·cm −2 ) placed horizontally below the photoreactor. The lateral walls were covered with aluminum foil. Cooling water circulated in a thimble surrounding the reactor to maintain the temperature at an ambient value. Samples were withdrawn at fixed time intervals, mixed with 50 vol % acetonitrile/water (50/50) solution to break the emulsion trapped by solid particles, and filtered (Millipore, PTFE, 0.45 μm) before analysis.
Notably, the amount of catalyst chosen for the photodegradation tests was selected from a compromise between issues related to photocatalysis and to the stability of the emulsion. The amount of photocatalyst allowed one to neglect the backscattering, which would be relevant only for small values of the optical thickness. 53 Nevertheless, as the system was not optimized with respect to the radiant field distribution, no information on the intrinsic reaction kinetics could be retrieved. Therefore, the apparent kinetic constants reported throughout the text are intended to provide only a numerical comparison between degradation kinetics obtained under the same experimental conditions.
The time evolution of the photocatalytic degradation of NB emulsion was followed by using a HPLC system model 1290 Infinity. This apparatus was equipped with a binary pump (Agilent Technologies, G4220B), autosampler (Agilent Technologies, 1290 Sampler), photodiode array (PDA) detector (Agilent Technologies, 1260), and C18 NUCLEOSIL 100-5 column (Machery Nagel: 4.6 × 250 mm). The injected sample volume was 20 μL. The eluent circulating at a flow rate of 1 mL·min −1 consisted of ultrapure water (40%) and acetonitrile (60%). The resulting data were collected by using Agilent Mass Hunter software.
■ RESULTS AND DISCUSSION Emulsification Experiments. The formulation process of Pickering emulsions requires a precise control of the intervening parameters to effectively tailor type, stability, and droplet size. Preliminary experiments aimed to show the feasibility of the formulation of nitrobenzene/water (NB/W) emulsions with pristine TiO 2 and TiO 2 -F as the stabilizers. Figure 1A,B presents NB/W emulsions obtained in the presence of pristine TiO 2 and TiO 2 -F, respectively. In both cases, the phase containing NB sedimented, leaving a supernatant aqueous phase due to the higher density of NB (1.2 g·mL −1 ) with respect to water.
When pristine TiO 2 was used as the emulsifier ( Figure 1A), fast sedimentation of TiO 2 particles took place and a release of NB was observed. Therefore, pristine TiO 2 nanoparticles were unable to stabilize the NB/W emulsions. In fact, the hydrophobic NB molecules show low tendency to adsorb onto the hydroxylated and highly hydrophilic TiO 2 surface. Conversely, TiO 2 -F could stabilize NB/W emulsions quite efficiently under the same conditions ( Figure 1B). The droplet size was ca. 128 μm and no release of pure NB was observed during at least 2 months (Figure 2). The NB droplets undergo sedimentation quite fast because of the density of NB higher than water and quite large droplet size. The mean diameter of the NB droplets, however, did not change during storage, demonstrating that coalescence did not occur.   These results were confirmed by the optical microscopy images shown in Figure 3 where it is evident that NB droplets of ca. 130 μm in size are nicely dispersed in a continuous water phase.
The different behavior of the TiO 2 -F sample with respect to the pristine TiO 2 material can be explained by its surface properties. The fluorinated material was prepared by a simple ligand exchange of surface hydroxyl groups with F − ions, 54,55 as expressed by eq 1.
Equation 1 does not express a simple adsorption equilibrium of F − ions because fluoride modification is stable upon washing during the preparation procedure and reverse process, a nucleophilic substitution by hydroxide ions, requires high NaOH concentrations (ca. 1 M), and does not reach completion. 56 The surface density of F − ions on the surface of TiO 2 has been determined by means of ion chromatography. By considering that the specific surface area of the TiO 2 sample is 103 m 2 ·g −1 and its average hydroxylation degree ranges between 4.8 and 6.1 Ti−OH·nm −2 , 57 a maximum fluorine grafting was achieved at pH = 3.2. The fluorine surface density was 3 fluorinated surface sites per nm 2 , corresponding to 50− 60% of the total surface sites. 50 Therefore, the surface of the fluorinated TiO 2 presents two functionalities almost equally distributed, which endow it with amphiphilic features.
Contact Angle Measurements. The stabilization of Pickering emulsions requires that partial wetting condition of the solid particles by water and NB is fulfilled. The wettability of the photocatalyst was assessed by means of contact angle measurements. The classical technique of contact angle measurement consists of the optical observation of a small drop of oil deposited onto the solid surface immersed in water (or of a drop of water deposited on the surface immersed in oil) and the determination of the contact angle in water on the picture. Furthermore, this technique affords information on the type and stability of Pickering emulsions. 18,58 In particular, contact angles close to 90°p rovide the most efficient stabilization of Pickering emulsions. Contact angles lower than 90°are indicative of hydrophilic particles that can better stabilize oil-in-water emulsions, while particles with contact angles higher than 90°better stabilize water-in-oil emulsions.
It was not possible to calculate the contact angle of NB over pristine TiO 2 because the drop of NB floated and rolled on the surface of the solid. This indicated total wetting by water and the absence of any interaction between TiO 2 and NB. This result explains why it was not possible to stabilize Pickering emulsions containing NB by using pristine TiO 2 as the emulsifier. On the contrary, the measurement was possible on the TiO 2 -F sample as shown in Figure 4.
A drop of NB on TiO 2 -F immersed in water presents a contact angle of 102.3 ± 0.6°. This indicates partial wettability and lower hydrophilicity of the TiO 2 -F surface with respect to the pristine TiO 2 . Such a contact angle close to 90°allows the stabilization of either O/W or W/O Pickering emulsions by TiO 2 -F particles, depending on the relative amounts of oil and water. 59 Hence, NB/water emulsions should be obtained when the NB content is smaller than that of water.
Characterization of the Emulsion. Two crucial features of the emulsion have been investigated: the size of the droplets and the type of emulsion as a function of the amount of the solid nanoparticles (TiO 2 -F) and of the volume fraction of the oil phase (Φ NB ).
The type of emulsion has been identified by varying the volume fraction of NB ( Figure 5A). Conductivity measurements performed on emulsions with different volume fractions of the oil phase gave evidence of a phase inversion phenomenon from O/W type (NB droplets in water, high conductivity) to W/O type (water droplets in NB, low conductivity) when Φ NB exceeded ca. 50 vol %. Notably, the possibility to switch between two emulsion types is a significant advantage of the present system with respect to analogous systems stabilized by surfactants; therefore, it is possible to stabilize with the same type of particle both direct and inverse emulsions.
The droplet size is a relevant parameter characterizing the ability of adsorbed particles to stabilize the emulsion. The stability of Pickering emulsions is governed by the reduction of the free energy of the system caused by the transfer of the photocatalyst from the aqueous dispersion to the NB/W interface. This transfer produces a variation in the size of the droplets and the consequent increase in the interfacial area. Therefore, the droplet size mainly depends on the amount of stabilizer (TiO 2 -F). Figure 5B shows the changes in the mean  . This can be explained by a lack of efficiency of the emulsification process because the UltraTurrax disperser was unable to reduce the NB droplet size down to the sizes that would have been expected by extrapolation of the curve in the first regime. As a consequence, TiO 2 -F nanoparticles were in excess once the plateau region was reached.
In the plateau region, the Pickering emulsion is comprised of micrometer-sized NB droplets surrounded by TiO 2 -F, which dually acts as a stabilizer and, once irradiated, as a photocatalyst. Therefore, the stabilized droplets can be seen as microreactors in which the photocatalytic reaction rate can be maximized due to the reduced mass transfer limitations and to the enhanced contact area between the substrate and photocatalyst.
Photocatalytic Activity. The photocatalytic experiments were performed using NB/water/TiO 2 -F Pickering emulsions containing 0.1 vol % NB (1200 ppm). Figure 6 shows the mean size distribution of the NB droplets along with an optical microscopy image (inset) of the emulsion.
A stable Pickering emulsion with small NB droplets nicely dispersed in water is evident. Most of the droplets have a diameter ranging between 1.5 and 3 μm, while a smaller fraction presents diameters ranging between 0.3 and 0.6 μm. This smaller fraction corresponds to the amount of TiO 2 -F in excess that was not adsorbed on the surface of the NB droplets.
The presence of fluorine dually acts in improving the photocatalytic system. In fact, it enhances the photocatalytic activity with respect to pristine TiO 2 and stabilizes the emulsion, favoring the formation of micrometer-sized droplets.
The superior photocatalytic activity of the fluorinated TiO 2 has been tested by choosing an initial NB concentration, which could dissolve in water, i.e., 50 ppm, and performing its photodegradation in the presence of TiO 2 and TiO 2 -F. Results are reported in Figure 7.
The NB photodegradation is faster in the presence of the fluorinated sample. In particular, the apparent observed kinetic constants, determined by differentiating the experimental data at the initial time, by using a five-point formula for equally spaced points, 60,61 were 16 × 10 −3 and 12 × 10 −3 min −1 for the fluorinated and bare samples, respectively. This evidence, often reported in the literature, has been generally related to the higher availability of photogenerated holes, which eventually results in higher production of hydroxyl radicals. 43 Other authors attributed the enhanced activity to the presence of singlet oxygen generated through a prevailing energy transfer mechanism. 44 Moreover, the intraband-gap energy states within 1.3 eV above the valence band, whose existence has been theoretically predicted 62,63 and only recently experimentally demonstrated, 50 could also be beneficial, acting as traps, enhancing the life time of the photogenerated holes.  However, by taking into account the above reported surface features, the affinity of the photocatalyst surface with the organic substrate may also play a significant role. In fact, TiO 2 surface fluorination decreases its hydrophilicity, favoring the adsorption of NB and then its direct oxidation through the photogenerated holes.
To provide evidence of the contribution of the emulsification process on the photocatalytic activity, the photodegradation of NB has been carried out in emulsified and nonemulsified NB/water mixtures in the presence of TiO 2 -F. The results of the runs are illustrated in Figure 8 in terms of normalized concentration (NB concentration at time t/initial NB concentration × 100). For the sake of comparison, an experiment in the absence of the photocatalyst is also reported.
NB concentration always decreased in the presence of TiO 2 -F photocatalyst under UV irradiation. However, only 13% of the initial NB concentration was photodegraded after 48 h in the case of the nonemulsified NB/W system. The photocatalytic degradation was significantly higher in the emulsified systems, reaching 75% after the same reaction time. The apparent rate constants, determined as above, were 0.7 × 10 −3 and 0.04 × 10 −3 min −1 for the emulsified and nonemulsified TiO 2 -F systems, respectively. In other words, the emulsified system showed an initial NB degradation rate ca. 18 times greater than that retrieved for the nonemulsified system.
By comparing the results reported in Figures 7 and 8, it is evident that the stabilizing effect of TiO 2 -F plays a predominant role in determining the global efficiency with respect to the intrinsic higher efficiency of the fluorinated sample. In fact, the amphiphilic properties of the fluorinated surface give rise to adsorption equilibrium of the two phases, thus increasing the dispersive capacity of the TiO 2 -F nanoparticles and thermodynamically favoring the formation of micrometer-sized NB droplets. The resulting high contact surface area between the photocatalyst and substrate improves the photocatalytic activity. Indeed, by considering the photoreactor volume V = 160 mL, volume fraction of NB Φ NB = 0.1 vol %, and NB droplet size D = 2 μm, the total NB−water interfacial area in the emulsified system is 0.48 m 2 as calculated by eq 2.
Contrarily, in the nonemulsified system, the photocatalyst nanoparticles are mainly located at the planar interface between the two liquid phases (ca. 8 × 10 −3 m 2 ).
Notably, by considering that only half of the surface hydroxyl groups are substituted by fluorine, adsorption of water still takes place, contributing to the photocatalytic reaction in various ways. Photogenerated holes induce water oxidation and produce hydroxyl radicals, which can in turn oxidize NB. Moreover, water may compete for adsorption with oxygenated degradation products, thus avoiding poisoning of the photocatalyst. It is worth mentioning that even if the apparent rate constant of TiO 2 -F in emulsion (0.7 × 10 −3 min −1 ) is much smaller than that in solution (16 × 10 −3 min −1 ), in the first case, it was possible to treat 1200 ppm NB, while only 50 ppm could be treated in solution. At this high NB concentration, the "classical" nonemulsified TiO 2 -F system is virtually not active (0.04 × 10 −3 min −1 ). Therefore, the size distribution of the droplets is a critical parameter both for the photocatalytic activity and for the stability of the emulsion. The remarkable efficiency enhancement observed in the emulsified system allows one to treat in one-pot high amounts of insoluble pollutants without implementing physical separation methods, which, in any case, require costly chemical degradation downstream.
The oxidation by products of the photocatalytic oxidation of NB dissolved in TiO 2 aqueous suspensions has been already reported in the relevant literature as mainly derived from NB hydroxylation induced by hydroxyl radicals photocatalytically generated. 3,6,64 Accordingly, in the present investigation, 4-nitrophenol (4-NP) and 3-nitrophenol (3-NP) were the main intermediates after 180 min irradiation time for the NB degradation tests shown in Figure 7. Figure 9 reports the correspondent chromatograms and UV−vis spec-      Performing the reaction in NB/W emulsion (as in the runs shown in Figure 8) did not significantly affect the degradation mechanism. Figure 10 shows the intermediates identified after 48 h irradiation in the TiO 2 -F stabilized emulsion, along with their UV−vis spectra.
Small amounts of 2-NP have been detected along with the 4-NP and 3-NP isomers, with the latter being the most abundant as mentioned above. A small and broad peak is detectable between 2.25 and 3.25 min, which could not be identified due to the low intensity and probable overlapping of signals related to different species. However, it is plausible to ascribe this signal to di-hydroxylated compounds such as nitrocatechol or nitroresorcinol, which are more hydrophilic compounds than the nitrophenol isomers, and consequently have lower retention time. This is also in agreement with Bhatkhande et al., 6 which also detected these intermediates in smaller amounts with respect to the monohydroxylated products.

■ CONCLUSIONS
The present paper reports the possibility of exploiting the unique amphiphilic properties of the fluorinated surface of TiO 2 to stabilize Pickering emulsions of organic pollutants poorly soluble in water. The obtained system is comprised of small droplets of substrate surrounded by the photocatalyst that act as microreactors for the degradation of the pollutant. This configuration allows obtaining 18 times faster photodegradation with respect to the nonemulsified system. The characterization of the physicochemical properties of the emulsion shows an inversion of the type of emulsion (from oil in water to water in oil) obtained by changing the amount of organic phase in the system. Moreover, the stabilizing ability of the fluorinated TiO 2 nanoparticles due to their amphiphilic surface is highlighted. Therefore, the role of fluorination is dual. On one hand, fluorination affects both the electronic and surface properties of TiO 2 and enhances the photocatalytic activity due to the higher production of hydroxyl radicals. On the other hand, the peculiar surface features allow one to stabilize NB emulsions, which was not possible by using bare TiO 2 . The present results shed some light on the hydrophilicity of the fluorinated surface of TiO 2 and open the route for one-pot treatments of biphasic polluted streams without the need of preliminary physical separation. The authors declare no competing financial interest.