Synthesis of nano-TiO2 assisted by glycols and submitted to hydrothermal or conventional heat treatment with promising photocatalytic activity

TiO2 nanoparticles were successfully synthesized by the sol-gel method employing different glycols (ethylene glycol, diethylene glycol or polyethylene glycol 300), which were heat-treated in conventional oven or by hydrothermal via, obtaining photocatalysts with particle sizes and distinct crystalline structures. HRTEM analyses showed that the oxides submitted to hydrothermal treatment featured spherical morphology, being formed by partially aggregated particles with sizes varying between 2 and 5 nm. X-ray diffractograms and Raman spectroscopy confirm that anatase was predominant in all synthesized compounds, with presence of brookite phase for samples that received hydrothermal treatment or were synthesized in the presence of polyethylene glycol with heat treatment in conventional oven. The amount of brookite as well as the cell volume, deformation, network parameters and crystallinity were estimated by Rietveld refinement. The surface area and porosity of the materials were higher when the synthesis involved the use of hydrothermal treatment. These oxides are mesoporous with porosity between 14 and 31%. The oxide synthesized in the presence of ethylene glycol with hydrothermal thermal treatment (TiO2G1HT) exhibited the highest photocatalytic activity in terms of mineralization of azo-dye Ponceau 4R (C.I. 16255), under UV-Vis irradiation. This higher photocatalytic activity can be attributed to the formation of binary oxides composed by anatase and brookite and by its optimized morphological and electronic properties.

In the present study, photocatalysts based on TiO 2 were synthesized by the sol-gel method. 114 The influence of the use of different structural molds (ethylene glycol, diethylene glycol or 115 polyethylene glycol 300) as well as the effect of thermal treatments by conventional or 116 hydrothermal routes, was evaluated on their photocatalytic activity, and structural optical and 117 morphological properties. The photocatalytic activity was evaluated through the degradation of 118 the azo-dye Ponceau 4R, chosen due to its industrial application and undesirable effects on the 119 environment and human health (Oliveira et al., 2012;European Food 2020). The results 120 presented here aim to provide new insights into the synthesis of TiO 2 -based photocatalysts with 121 different crystalline phases and the influence of preparation conditions on the photocatalytic 122 properties of these systems. All chemicals were of analytical or HPLC grade and were used as received. Ultrapure water 129 obtained from an Elix 5 Milli-Q ® water purification system was employed in all experiments. 130 TiO 2 samples were synthesized by the sol-gel method, using different glycols (ethylene glycol, 131 diethylene glycol or polyethylene glycol 300) (Sigma Aldrich), and heat treatment in a 132 conventional oven or hydrothermal system. 133 The TiO 2 Gx photocatalyst was obtained from the mixture, under magnetic stirring, of 10 mL 134 of Ti (IV) isopropoxide (Aldrich, 97%) and 50 mL of glycol (where x = 1 when 886 mmol of 135 ethylene glycol (Vetec, 99.5%) were used, x = 2 for 527 mmol diethylene glycol (Vetec, 99.5%), 136 and x = 3 when polyethylene glycol 300 (Fluka) was used). After 2 hours of stirring, a mixture 137 containing 10 mL of ultrapure water and 90 ml of acetone (Synth, 99.5%) was added to the 138 suspension and kept under stirring for 2 hours. The white precipitate was separated with the aid 139 of a centrifuge (9000 rpm for 20 minutes), followed by washing several times with ethanol to 140 remove residues of glycol, followed by washing three times with distilled water.

141
For the preparation of heat-treated photocatalysts in a conventional oven (TiO 2 GxM), after 142 washing the powder was dried at 70°C under reduced pressure and sintered at 400°C for 2 hours. 143 After centrifugation and washing the decanted oxide prepared using hydrothermal treatment, 144 TiO 2 GxHT, was submitted to the hydrothermal reactor under a pressure of approximately 13.8 145 bar at 200°C for 4 hours. Subsequently, it was dried at 70°C for 24 hours. High resolution electronic transmission images were obtained using a Jeol, JEM-2100, 152 Thermo scientific Transmission Electron Microscope. The particle size and spacing between 153 crystalline planes were calculate with the free software "ImageJ". 154 X-ray diffraction analyses (XRD) using a Shimadzu XRD600 powder diffractometer 155 operating at 40 kV and 120 mA, employing Cu Kα (λ= 1,54148 Å) radiation. The diffractograms 156 were collected between 10°≤ 2θ ≤ 90° under a rate of 0.5º min -1 . Crystalline silicon was used as 157 the diffraction standard. X-ray diffratogram of the oxides were refined by the method of Rietveld 158 using the FullProf software, with fitting criteria (Factor S -Goodness of Fit) was employed as 159 the ratio between the weight factor (R wp ) and the expected factor (R exp ), which should be closer 160 to 1. The fit parameters can be found in the Supplemental Information (Table S1).

161
N 2 adsorption-desorption isotherms were obtained using an ASAP 2010 analyzer 162 (Micrometrics). The specific area were analyzed using the Brunauer, Emmett and Teller (BET) 163 model and the Barrett-Joyner-Halenda (BJH) model for the porous volume (Barrett, Joyner & 164 Halenda, 1951). 165 Raman spectra were acquired at room temperature using a Bruker spectrometer model RFS 166 100/S, samples were excited at 1064 nm with laser operating at 100 mW. Diffuse reflectance 167 spectra of the synthesized oxides were acquired using an UV-1650PC Spectrometer (Shimadzu), 168 at room temperature and potassium bromide was used as reference. The band gap energy being 169 estimated by the Kubelka-Munk function (Patterson, Shelden & Stockton, 1977). In all photocatalytic assays, 100 mg L -1 of the catalyst was added to 31 mg L -1 dye Ponceau 174 4R (trisodium (8Z)-7-oxo-8-[(4-sulfonatonaphthalen-1-yl)hydrazinylidene]naphthalene-1,3-175 disulfonate, CI 16255, Sigma-Aldrich, 75%) aqueous solution (pH = 6.9) under magnetic 176 stirring. The experimental setup was previously described in detail (Oliveira et al., 2012). 177 Information about the radiation source and experimental data were available in (Machado et al., 178 2008;Santos et al., 2015). 179 The photocatalytic system was kept at 40 ± 2 ºC and under stirring for 30 minutes in the dark 180 to reach the adsorption equilibrium. Control measurements in the dark were performed and in the 181 absence of a catalyst to evidence the role of TiO 2 in the photochemical reaction. Aliquots were 182 taken at 20 minutes intervals, filtered and analyzed by spectrophotometry, following the 183 discoloration at 507 nm using a Shimadzu spectrophotometer model 1650PC and by Total 184 Organic Carbon (TOC) measurements, using a Shimadzu TOC-VCPH/CPN analyzer.  The XRD data ( Fig. 2) confirm that all samples are composed mostly of nanocrystals of 216 anatase, with the (101) phase preferably exposed. In the case of HT processing at 200°C for 4h, 217 the presence of crystalline anatase phases and traces of brookite was observed, being confirmed 218 by the presence of peaks at 2θ equal to 25.38° (101) and 30.80° (121), respectively. Under the 219 treatment conditions to which these materials were submitted, the formation of the rutile phase 220 was not observed. The formation of the brookite phase was probably a crucial factor for the 221 inhibition of the transformation of anatase into rutile.

222
Rietveld analyses of the diffractograms (Fig. S1) confirm the decrease in crystallite size for 223 the oxides obtained after hydrothermal heat treatment (HT), which agrees with the HRTEM 224 images. These quantitative data also confirm the greater presence of brookite phase in samples 225 submitted to hydrothermal treatment. Besides that, it is observed that the percentage of the 226 brookite phase remains practically constant even with the use of different glycols in the synthesis 227 process. Already for materials prepared with heat treatment in a conventional oven, it turns out 228 that the use of different glycols leads to greater deformations only for the TiO 2 G3M sample, PeerJ Mat. Sci. reviewing PDF | (MATSCI-2020:09:53207:1:1:NEW 13 Jan 2021)

Chemistry Journals
Analytical, Inorganic, Organic, Physical, Materials Science 229 where polyethylene glycol was used in the synthesis, causing the formation of 17.47% brookite 230 phase (Tay et al., 2013). 231 The TiO 2 G3HT sample had a higher portion of brookite when compared to the TiO 2 G3M 232 sample, because since brookite is featured by its low symmetry, its formation is more efficient 233 under mild conditions such as shorter period and lower preparation temperature, as occurs in 234 hydrothermal treatment conditions (Lin et al., 2012). The formation of mesoporous structures was confirmed by N 2 adsorption-desorption 241 isotherms (Fig 3). Isotherms follow the type III for samples with 100% anatase phase (TiO 2 G1M 242 e TiO 2 G2M). The other samples, with brookite content, have type IV with a pronounced 243 hysteresis loop of types H3 and H4, according to the IUPAC classification. This suggests that 244 these materials are mesoporous solids formed by agglomerated or aggregated particles (Gregg & 245 Sing, 1982). The presence of brookite causes a decrease in the average pore diameters, 246 suggesting that the presence of structural defects influences the adsorption capacity and porosity 247 of the material. The values of surface area and porosity of these materials are presented in Table  248 1.   Table 1 Morphologic and electronic parameters to oxides synthesized. 254 [ Table] 255 256 The diffuse reflectance spectra, expressed in terms of F(R) vs. photon energy (E), are 257 presented in Fig. 4. The indirect band gap value (E g ) was obtained by extrapolating the linear 258 segment to the X axis, Table 1. However, a simple inspection of the spectra suggests that the 259 band gap values calculated in this way calculated in this way are deviated from the actual values, 260 since the radiation absorption is not canceled (E<E g ), except from the point where F(R) → 0. 261 This suggests the existence of permitted states with energies lower than the estimated E g , that is, 262 E g(real) < E g . Thus, considering the lower threshold of the conduction band, which occurs when 263 F(R)0, that is, states with energies less than or equal to the energy associated with this 264 threshold, are prohibited. In view of this, E g(real) was also calculated (Table 1). Based on this 265 information, it appears that all photocatalysts absorb radiation more intensely in the near-UV 266 region. However, these photocatalysts, despite the high band gap energies, have significant 267 photocatalytic activity in the visible region, as suggest the estimated values of E g (real) . The 268 TiO 2 G1HT, TiO 2 G2HT and TiO 2 G1M photocatalysts show a radiation absorption profile shifted PeerJ Mat. Sci. reviewing PDF | (MATSCI-2020:09:53207:1:1:NEW 13 Jan 2021)

Chemistry Journals
Analytical, Inorganic, Organic, Physical, Materials Science 269 to the visible region, with E<E g , being therefore able to uptake photons in a large range of 270 wavelengths. Related to these factors, the high surface area, crystallinity and mixture of 271 crystalline phases are added, which end up favoring the photocatalytic potential of these oxides.

272
The electronic properties of the particles change significantly by reducing their size. Thus, 273 new properties can be expected in nanoparticles when compared to bulk (Hodes, 2007). The 274 variation of energy as a function of size promotes the quantum confinement and is characterized 275 by an increase in the indirect band gap energy (E g ), as can be seen for TiO 2 G1HT, which has 276 smaller particle and crystallite sizes, as estimated by HRMET and DRX analyses, and E g (3.30 277 eV) greater than that of the extended solid (3.20 eV for TiO 2 ) (Kumar & Devi, 2011).  The catalysts were also evaluated using Raman spectroscopy (Fig. 5). All samples exhibit 284 vibration modes typical of anatase (3E g + 2B 1g + A 1g ). A 1g symmetry mode was not visualized, 285 probably due the overlap with the band corresponding to the second mode, of B 1g symmetry 286 (Iliev et al., 2013;Fang et al., 2015). A slight change in the signs is observed depending on the 287 type of heat treatment used (Fig 5 -Inset). The bands referring to samples thermally treated by 288 hydrothermal route are broader than those observed for the calcined oxides in a conventional 289 oven. This broadening can be directly correlated to the concentration of oxygen vacancies on the 290 photocatalysts, as previously shown by Parker e Siegel (Parker & Siegel, 1990). Thus, Raman 291 analysis indicates that the synthesis of oxides treated by the hydrothermal route, induces the 292 formation of oxygen vacancies on the oxide surface, increasing the system disorder.  Figure 5 Raman spectra, at room temperature, for the synthesized TiO 2 photocatalysts. Inset: 297 Expanded normalized Raman spectra between 100 and 200 cm -1 in the main E g peak region 298 attributed to the broadening of the band according to the type of heat treatment.
299 300 Photocatalytic activity 301 The photocatalytic activity of the different synthesized oxides was evaluated in terms of the 302 degradation of the azo-dye Ponceau 4R. The control experiment, in the absence of any 303 photocatalyst, reveals extremely low levels of dye discoloration (4.0%) and mineralization (13%) 304 after 140 minutes of irradiation (Fig. S2). The degradation efficiency presented by the different 305 photocatalysts is summarized in Table 2.   Table] 310 311 The oxides thermally treated by hydrothermal via were more efficient than conventional heat 312 treatment in promoting the degradation and mineralization of the dye under study. Calcination in 313 a conventional oven led to an increase in the crystallinity of the materials, as seen by the XRD 314 data, and a decrease in the surface area, which ended up compromising the photocatalytic 315 activity of these oxides.

316
The photocatalytic performance exhibited by the samples synthesized in the presence of 317 different glycols and thermally treated by hydrothermal via can be attributed to the coexistence 318 of anatase and brookite the high surface area, mesoporosity, and more appropriate particle sizes. 319 Crystalline materials with smaller particle sizes are more likely to exhibit expressive 320 photocatalytic properties (Ohno et al., 2001). 321 Although the TiO 2 G3M photocatalyst also presents itself as a mixture of polymorphs anatase 322 and brookite, it did not show significant photocatalytic activity, probably related to its smaller 323 surface area.

324
The increase in photocatalytic activity of samples that present anatase and brookite can be 325 explained by the synergism between these polymorphs. Although anatase and brookite present a 326 very close E g (Machado et al., 2012;Patrocinio et al., 2015), theoretical calculations have 327 shown that the energies of the conduction and valence bands of anatase phase are slightly lower 328 than the corresponding energy levels of brookite (Li et al., 2008), suggesting a certain ease of 329 migration of electrons from brookite to anatase. Thus, the holes are more available for oxidation 330 reactions. In addition, the energy barrier between these polymorphs will tend to hinder the 331 recombination among charge carriers. Therefore, with an extended life span, holes in the 332 brookite valence band have a greater chance to oxidize organic matter, while electrons "trapped" 333 in anatase may favor reduction reactions, leading to an increase in the photocatalytic activity (Li, 334 Ishigaki & Sun, 2007;Patrocinio et al., 2015). 335 A complex degradation mechanism is expected in heterogeneous photocatalysis (Hoffmann et 336 al., 1995;Ahmed et al., 2010). The reactions occur initially at the solid-solution interface and 337 involve reactive species generated on the surface of the excited photocatalyst or by direct 338 interaction between the excited photocatalyst and the substrate (Oliveira et al., 2012;Santos et 339 al., 2015). In the degradation under study, the discoloration of the dye is probably related to the 340 homolytic scission of the azo group. Hydroxyl radicals (HO • ), formed in the solid-solution 341 interface, may be responsible for this process (Kumar & Devi, 2011). Table 2 presents data on 342 the percentage of discoloration in the reactions mediated by the oxides synthesized in this study. 343 The best performances occurred using oxides submitted to hydrothermal treatment.

344
The mineralization process follows a Langmuir-Hinshelwood kinetics (Hoffmann et al., 345 1995), being of pseudo-first order in relation to the dye, as show in Fig. 6. The rate constants are 346 listed in Table 2. In these assays, it was found that only 4.0% of the dye was adsorbed in the photocatalyst (Fig.  356 S2), suggesting that the observed mechanism occurs mainly through the photodegradation of 357 organic matter, certainly by the action of reactive oxygen species (ROS), such as HO • or O 2 •-, 358 with the predominant action of the HO • radicals, a very strong oxidizing agent (standard 359 reduction potential of HO • /H 2 O 2.38 V vs. NHE) (Hoare, 1985). Accordingly, and based on the 360 characterization of the photocatalysts, we can propose a mechanism, shown in equations (1-9) 361 which must occur at the solid-solution interface, where TiO 2 (A) is the anatase polymorph and 362 TiO 2 (B) is the brookite polymorph. As result of the photoexcitation of the catalyst (1), the e -/h + 363 pairs are generated; recombination processes (2) compete with the electron trapping in the 364 polymorph anatase (3) and holes in the brookite polymorph (4 and 5), generating the reactive 365 species responsible for the degradation of the dye (5, 6 and 7). In the valence and conduction 366 bands, the oxidation (8) and reduction (9) reactions occur, respectively, resulting in degradation 367 products. The TiO 2 G1HT oxide, present the best photocatalytic performance (k app = 5.9 × 10 3 min -1 ; R 380 = 0.9824), because the availability of reactive species becomes proportionally higher as the 381 concentration of the dye decreases, since the concentration of these species is practically constant 382 during the photocatalytic process (França et al., 2016). Therefore, P4R undergo fragmentation at 383 the beginning of the reaction (Fig 6 -Inset), which should favor a faster mineralization.

384
Analyzing the spectrum presented in the Inset of Fig 6, it can be seen that at the end of the 385 photocatalytic process, the band centered at 500 nm, referred to an electronic transition with a 386 major component π → π* (Oliveira et al., 2012), involving the naphthalenic structures and the 387 azo group, associated with the coloring of the dye, decreases significantly. The formed products 388 should not present new or significant absorption bands in the analyzed region, suggesting that the 389 degradation not only induces a quick discoloration of the dye (Table 2), as they should also cause 390 a significant fragmentation of the dye structure, whose fragments should not absorb significantly 391 in the monitored region of the electromagnetic spectrum.

392
The coexistence of anatase and brookite in the TiO 2 synthesized with different glycols and 393 treated by a hydrothermal via at low temperature, minimized the recombination rate of the e -/h + 394 pairs, thus allowing the holes to be available for oxidation reactions. In addition, the correlation 395 of physical and chemical factors, such as high surface areas and porosity, high photon absorption 396 capacity in the UV-visible region and crystallinity considerably improved the photocatalytic 397 activity of these oxides. In this study, we present the preparation of TiO 2 mesoporous nanoparticles using the sol-gel 402 method with different glycols as structural molds. The use of ethylene glycol associated to 403 further hydrothermal heat treatment proved to be the most effective way to obtain nanoparticles 404 with improved photocatalytic activity. The results showed that materials submitted to 405 hydrothermal heat treatment presented smaller particles and greater porosity, with formation of 406 approximately spherical nanoparticles and with sizes up to 5 nm and formation of a binary 407 mixture of anatase and brookite phases. The use of different glycols influenced the size of the 408 particles, promoting the formation of smaller particles. The existence of a junction between 409 different phases of the same semiconductor, accompanied by a decrease in the size of the 410 particles, favored the charge transfer processes and contributed to the delay of the recombination 411 processes, significantly improving the photocatalytic activity, verified by the degradation of the 412 azo-dye Ponceau 4R under UV-Vis light irradiation. This type of photocatalyst that can harness 413 both UV and visible light is a promising candidate for applications in photochemistry, sensors 414 and solar cells, which has motivated us to develop oxides and nanocomposites based on TiO 2 415 with a wide spectrum of applications.