Mixed Metal Oxide W-TiO2 Nanopowder for Environmental Process: Synergy of Adsorption and Photocatalysis

A mixed metal oxide W-TiO2 nanopowder photocatalyst was prepared by using the sol–gel method with a broad range of elemental compositions x = CW/(CW + CTi), including TiO2 and WO3. The material was structurally characterized and evaluated in adsorption and photocatalytic processes by testing its removal capacity of a representative pollutant methylene blue (MB) in aqueous solutions and under UV-A and sunlight illuminations. The nanopowders appeared to be more effective adsorbents than pure TiO2 and WO3 materials, showing a maximum at 15 mol% W, which was set as the tungsten solubility limit in anatase titania. At the same time, the photocatalytic decomposition of MB peaked at 2 mol% W. The examination of different compositions showed that the most effective MB removal took place at 15 mol% W, which was attributed to the combined action of adsorption and heterogeneous photocatalysis. Moreover, MB decomposition under sunlight was stronger than under UV-A, suggesting photocatalyst activation by visible light. The pollutant removal efficiency of the material with 15 mol% W was enhanced by a factor of ~10 compared to pure TiO2 at the beginning of the process, which shows its high potential for use in depollution processes in emergency cases of a great pollutant leak. As a result, a Wx=0.15-TiO2 catalyst could be of high interest for wastewater purification in industrial plants.


Introduction
The research on TiO 2 photocatalysts is steadily increasing since the discovery of UVlight assisted electrochemical water splitting [1].The extensive research has shed light on the relevant mechanisms and reaction pathways [2][3][4].As a result, more than four million tons of TiO 2 per year are nowadays produced to satisfy the growing research and industrial demand in environmental and energy-related fields [5,6].
W-TiO 2 composites have been prepared via several methods including physical mixing [68], sol-gel [7,19,21,29,[35][36][37][38]48,[69][70][71][72][73][74], solution combustion [50], hydrothermal [52,55,[75][76][77], ultrasound-assisted [28], electrospinning [47,78], impregnation [42], wet chemical [63] and incipient wetting [21,51].Among these methods, the sol-gel technique is proven to be highly attractive for synthesis because of its uncomplicated, low-cost procedure and flexible synthetic routes and the possibility of attaining highly homogeneous compositions [37,79].At the same time, the variation in adsorption in the mixed metal oxide remains underexplored.An available study [80] has applied an addition of W to TiO 2 Degussa P25 powder, which did not allow for the formation of perfectly homogeneous compounds at the nanoscale.A competition between an increase in the specific surface area and a decrease in the activity per adsorption site has been also confirmed in the compounds with W content ≤ 6 mol% [81].However, the elemental composition can greatly affect the specific surface area of these materials, as it has been shown in an example of Zr x Ti 1−x O 2 nanopowders [23].This offers an opportunity to control adsorption and photocatalysis processes in the same material.The full range of elemental W/Ti compositions has not been explored in previous studies with respect to the adsorption and heterogeneous reactions' effectiveness, covering UV-A (most suitable for pure titania) and sunlight illuminations for the photocatalyst activation.
In this work, we report on a simplified preparation procedure of mixed metal oxide W-TiO 2 nanopowders via the sol-gel method with a broad range of elemental compositions.The adsorption and photocatalytic activity of the prepared materials were tested with UV-A lamp and natural sunlight illuminations based on the removal of a methylene blue (MB) pollutant in aqueous solutions.The influence of the material composition on the process effectiveness was inspected.
W-TiO 2 nanopowders were synthesized via the sol-gel method.A schematic diagram of the W-TiO 2 nanopowder preparation procedure is depicted in Figure 1.The synthesis was conducted at the total precursor concentration of C Ti + C W = 0.3 mol/L with a hydrolysis ratio h = C H2O /(C Ti + C W ) = 1.25,where C H2O , C Ti and C W are molar concentrations of water, TTIP and tungsten (VI) chloride.The pure TiO 2 powder was prepared by dissolving TTIP in 2-propanol under stirring at 500 rpm for 15 min to obtain solution A. Then, the required amount of water was mixed with 2-propanol under stirring for 15 min to obtain solution B, which was slowly dropped into solution A under intense stirring, which triggered the nucleation of titanium oxo-alkoxy (TOA) nanoparticles [82].The obtained colloidal solution was dried in a fume hood at room temperature for 5 days to obtain white amorphous powders, which were further dried in an oven at 85 • C for 2 days and afterwards calcined in a Nabertherm furnace (Lilienthal, Germany) at 450, 500, 550 and 600 • C for 4 h.The pure WO 3 powder was prepared by dissolving an appropriate amount of WCl 6 in 2-propanol under stirring for 15 min to obtain solution A. Solution B was prepared by adding the required volume of water to 2-propanol and then stirring for 15 min.After that, solution B was slowly dropped into solution A under stirring for 15 min.The mixed solution was dried, evaporated and calcined in the same way as the pure TiO 2 .The mixed metal oxide W-TiO 2 nanopowders were prepared with different compositions x = C W /(C W + C Ti ) = 0.0025, 0.01, 0.02, 0.04, 0.06, 0.08, 0.15, 0.30 and 0.50.Solution A was prepared by following the preparation procedure of pure TiO 2 .For solution B, the required amount of WCl 6 was dissolved in 2-propanol under stirring for 5 min, and then water was added under stirring for 10 min.Solution B was then slowly dropped into solution A under stirring, then dried, evaporated and calcined in a furnace at 450, 500, 550 and 600 • C for four hours.Equal volumes of the solutions A and B were used in all synthesis procedures.The total volume of 150 mL was used for compositions x = 0, 0.0025, 0.01, 0.02, 0.04, 0.06 and 0.08, 100 mL for compositions x = 0.15 and 0.30, 50 mL for x = 0.50, and 20 mL for x = 1.

Characterization Techniques
The X-ray diffraction (XRD) pa erns were collected with a Co-Kα, INEL EQUINOX 1000 apparatus.The specific surface area of samples was measured by N2 physisorption at 77 K using a BELSORP-MAX apparatus from Microtrac (Haan, Germany).The specific surface area was obtained over the relative pressure range (0.05-0.25) using the Brunauer, Emmet and Teller (BET) equation.Before analysis, samples were degassed under primary vacuum (<150 mPa) at 393 K for 12 h using a BELREP VAC module from Microtrac.The Raman spectra were measured with a high-resolution micro-Raman spectrometer HR800 HORIBA Jobin Yvon (Palaiseau, France) operating at a wavelength of 473 nm, with spectral and spatial resolutions, respectively, of 0.25 cm −1 and 5 µm.The sca ered light was collected in a backsca ering configuration and recorded on a Peltier-cooled CCD camera.EDX measurements were performed with a Hitachi TM3000 TableTop Scanning Electron Microscope (Tokyo, Japan) equipped with a Swift ED3000 module for elemental analysis.
The time-resolved microwave conductivity (TRMC) measurements were conducted at the experimental installation described in Ref. [83].The method allows for the monitoring relaxation of the photoinduced charges by monitoring changes in the microwave power reflected by a semiconductor surface after light illumination, which induces electrical conductivity.In these experiments, the samples were irradiated with 360 nm light pulses of 8 ns duration and 1.5 mJ energy delivered by a cw OPO Nd:Yag laser (EKSPLA) operating at a 10 Hz repetition rate.The area of samples with a 5 mm diameter was illuminated and the recorded TRMC decay curves were accumulated over 200 laser pulses.The measurements were performed at room temperature and in an ambient atmosphere.
The adsorption capacity of 0.025 g of W-TiO2 nanoparticles was measured in aqueous solutions of 100 mL volume with 10 ppm MB.The solution was stirred at 500 rpm in the dark and then 5 mL of the solution was taken every 15 min into a centrifuge tube.This sample was centrifuged at 4300 rpm for 15 min using the CompactStar CS4 setup to precipitate the powders, and the concentration of the remaining MB in the clear solution was calculated after measurements of light absorption at λ = 664 nm using a UV-visible spectrophotometer HS 3300 (Humas, Daejeon, Korea).

Characterization Techniques
The X-ray diffraction (XRD) patterns were collected with a Co-Kα, INEL EQUINOX 1000 apparatus.The specific surface area of samples was measured by N 2 physisorption at 77 K using a BELSORP-MAX apparatus from Microtrac (Haan, Germany).The specific surface area was obtained over the relative pressure range (0.05-0.25) using the Brunauer, Emmet and Teller (BET) equation.Before analysis, samples were degassed under primary vacuum (<150 mPa) at 393 K for 12 h using a BELREP VAC module from Microtrac.The Raman spectra were measured with a high-resolution micro-Raman spectrometer HR800 HORIBA Jobin Yvon (Palaiseau, France) operating at a wavelength of 473 nm, with spectral and spatial resolutions, respectively, of 0.25 cm −1 and 5 µm.The scattered light was collected in a backscattering configuration and recorded on a Peltier-cooled CCD camera.EDX measurements were performed with a Hitachi TM3000 TableTop Scanning Electron Microscope (Tokyo, Japan) equipped with a Swift ED3000 module for elemental analysis.
The time-resolved microwave conductivity (TRMC) measurements were conducted at the experimental installation described in Ref. [83].The method allows for the monitoring relaxation of the photoinduced charges by monitoring changes in the microwave power reflected by a semiconductor surface after light illumination, which induces electrical conductivity.In these experiments, the samples were irradiated with 360 nm light pulses of 8 ns duration and 1.5 mJ energy delivered by a cw OPO Nd:Yag laser (EKSPLA) operating at a 10 Hz repetition rate.The area of samples with a 5 mm diameter was illuminated and the recorded TRMC decay curves were accumulated over 200 laser pulses.The measurements were performed at room temperature and in an ambient atmosphere.
The adsorption capacity of 0.025 g of W-TiO 2 nanoparticles was measured in aqueous solutions of 100 mL volume with 10 ppm MB.The solution was stirred at 500 rpm in the dark and then 5 mL of the solution was taken every 15 min into a centrifuge tube.This sample was centrifuged at 4300 rpm for 15 min using the CompactStar CS4 setup to precipitate the powders, and the concentration of the remaining MB in the clear solution was calculated after measurements of light absorption at λ = 664 nm using a UV-visible spectrophotometer HS 3300 (Humas, Daejeon, Korea).
The photocatalytic activity was measured through the decomposition of 10 ppm MB in aqueous solutions of 200 mL volume, with 0.025 g of W-TiO 2 nanoparticles.The catalyst powder was added to the MB solution under stirring at 500 rpm in the dark until adsorption equilibrium was reached and the solution was then illuminated with a UV-A lamp VL-6.LC (λ = 365 nm, 24 W; VWR International, Rosny-sous-Bois, France) at a distance of 10 cm.The photocatalytic activity under sunlight illumination was tested between 9 and 12 am in Phnom Penh city (Cambodia) with the same nanoparticle concentration of 0.125 g/L.In these experiments, the full series of W-TiO 2 photocatalysts with different elemental compositions (or W content, equal to 100× in mol%) was placed in an identic aqueous solution of MB pollutant and simultaneously exposed to sunlight or a UV-A lamp.Samples of 5 mL of the solution were taken every 15 min and centrifuged at 4300 rpm for 15 min in order to separate catalyst particles, and absorption at λ = 664 nm was measured by the UV/visible spectrophotometer to evaluate the remaining MB concentration.The first-order rate constants of MB decomposition were measured for a comparison of the material activities.

Structural Properties
Before discussing the functional properties of the prepared materials, some comments will be given on their structural properties.We notice that nanopowders prepared according to the described method are expected to consist of the smallest nanoparticles-nuclei. Indeed, this fact has been previously evidenced in TOA nanoparticles formed at low hydrolysis ratios h ≤ 1.5 [80,81].These TOA nuclei have the size of 2R = 3.2 nm and agglomerate into larger particles while keeping their individual characteristics [84].The nucleation growth process of tungsten-titanium oxo-alkoxy (WTOA) nanoparticles in the sol-gel synthesis procedure has not been yet documented.However, the WCl 6 precursor is known to readily hydrolyze.Keeping in mind the stronger Pauling electronegativity of W (2.36) compared to Ti (1.54), one can expect a faster hydrolysis rate compared to the polycondensation of tungsten species, similar to vanadium species.The last point argues for a similar tendency in the formation of composite WTOA nanoparticles to that of vanadiumtitanium oxo-alkoxy (VTOA) nanoparticles [85].Although perfect micromixing conditions were not controlled in the simplified synthesis procedure, an extremely low hydrolysis ratio well below that required for the complete nucleation slowed down the reaction kinetics, making quasi-perfect elemental dispersion possible at the nanoscale, probably forming single-shell and core-shell WTOA nanoparticles depending on the colloid composition.Future studies will clarify this issue.We also confirmed the elemental compositions x of W-TiO 2 nanopowders after the synthesis procedure by performing EDX measurements (Table S1 and Figure S1 of Supplementary Materials).
XRD patterns of the calcined W-TiO 2 nanopowders showed evidence of a general tendency to stabilize the metastable anatase phase.The most intense peaks of the anatase (2θ = 25.28 • ) and rutile (2θ = 27.44 • ) phases can serve as a guide to follow the phase transformation.As Figure 2a shows, the stable rutile phase appeared in pure TiO 2 after calcination at temperatures above 500 • C, which is in agreement with previous studies [86].
In contrast, no traces of rutile were observed in W-doped nanopowders even at 600 • C (Figure 2b).Besides the stabilizing interaction, this indicated a very homogeneous elemental mixing of W and Ti elements at the nanoscale, which led to no pure titania domains transforming to the rutile phase.
Furthermore, the anatase TiO 2 crystalline phase was preserved with W insertion up to 15 mol%, as evidenced Figure 3.In contrast, with a greater W loading, WO 3 crystallized in the monoclinic phase [87], as shown in the growing peaks at 23.5 • and 33.5 • .This behavior was common to all prepared powders, and the fraction of the WO 3 phase increased with an increase in the temperature from 500 • C to 600 • C. It has been earlier shown that additives (Li + , K + , Cu 2+ and Al 3+ ) inducing vacancy in the anion sublattice promote the anatase-rutile phase transition, while additives (S 5+ , P 5+ and Nb 5+ ) reducing the number of vacancies inhibit the phase transition [88].This effect has also been observed in 1-4 mol% W-doped TiO 2 [89].Our experimental findings support these observations in a larger domain of the material compositions.In contrast, no traces of rutile were observed in W-doped nanopowders even at 600 °C (Figure 2b).Besides the stabilizing interaction, this indicated a very homogeneous elemental mixing of W and Ti elements at the nanoscale, which led to no pure titania domains transforming to the rutile phase.
Furthermore, the anatase TiO2 crystalline phase was preserved with W insertion up to 15 mol%, as evidenced Figure 3.In contrast, with a greater W loading, WO3 crystallized in the monoclinic phase [87], as shown in the growing peaks at 23.5° and 33.5°.This behavior was common to all prepared powders, and the fraction of the WO3 phase increased with an increase in the temperature from 500 °C to 600 °C.It has been earlier shown that additives (Li + , K + , Cu 2+ and Al 3+ ) inducing vacancy in the anion subla ice promote the anatase-rutile phase transition, while additives (S 5+ , P 5+ and Nb 5+ ) reducing the number of vacancies inhibit the phase transition [88].This effect has also been observed in 1-4 mol% W-doped TiO2 [89].Our experimental findings support these observations in a larger domain of the material compositions.In contrast, no traces of rutile were observed in W-doped nanopowders even at 600 °C (Figure 2b).Besides the stabilizing interaction, this indicated a very homogeneous elemental mixing of W and Ti elements at the nanoscale, which led to no pure titania domains transforming to the rutile phase.
Furthermore, the anatase TiO2 crystalline phase was preserved with W insertion up to 15 mol%, as evidenced Figure 3.In contrast, with a greater W loading, WO3 crystallized in the monoclinic phase [87], as shown in the growing peaks at 23.5° and 33.5°.This behavior was common to all prepared powders, and the fraction of the WO3 phase increased with an increase in the temperature from 500 °C to 600 °C.It has been earlier shown that additives (Li + , K + , Cu 2+ and Al 3+ ) inducing vacancy in the anion subla ice promote the anatase-rutile phase transition, while additives (S 5+ , P 5+ and Nb 5+ ) reducing the number of vacancies inhibit the phase transition [88].This effect has also been observed in 1-4 mol% W-doped TiO2 [89].Our experimental findings support these observations in a larger domain of the material compositions.Changes in the nanopowder morphology accompanied the particles' crystallization.In particular, a clear decreasing trend in crystallite size was observed with an increase in W content, as Figure 4 shows based on the Scherrer equation.The decrease in the crystallite size of Ti 0.8 W 0.2 O 2 compared to pure titania has been previously reported and attributed to the incorporation of W into the TiO 2 lattice [90].
Two particularities of the prepared materials can be observed from the Raman analysis in Figure 5.The first one is that the low-energy E g mode in the range of ~150 cm −1 shifts to higher frequencies upon W insertion, which has been previously attributed to the unit cell expansion [91].However, (in disagreement with this last study) no linear shift in this Raman mode with the increase in the W content, saturating at 15 mol% W, was observed in our work.In contrast, the shift appeared to be significant between 1 and 6 mol% W, with no subsequent evolution upon greater W insertion.The second particularity concerns the blue shift in the low-frequency A 1g /B 1g and high-energy E g modes with an increase in W content above 6 mol%.This may be an indication of the cell deformation toward rutile geometry.In agreement with the XRD analysis, the Raman spectra (broad bands at 260, 320 and 690 cm −1 , marked by (*) in Figure 5) confirm the material segregation above 15 mol% W, resulting in anatase TiO 2 /monoclinic WO 3 [92] polymorph powders.In contrast, the dissolution of W into te anatase TiO 2 matrix is complete for the lower W contents.Because of that, we tentatively ascribed 15 mol% to the solubility limit of tungsten in anatase titania.
Nanomaterials 2024, 14, x FOR PEER REVIEW 7 of 18 Changes in the nanopowder morphology accompanied the particles' crystallization.In particular, a clear decreasing trend in crystallite size was observed with an increase in W content, as Figure 4 shows based on the Scherrer equation.The decrease in the crystallite size of Ti0.8W0.2O2compared to pure titania has been previously reported and a ributed to the incorporation of W into the TiO2 la ice [90].Two particularities of the prepared materials can be observed from the Raman analysis in Figure 5.The first one is that the low-energy Eg mode in the range of ~150 cm −1 shifts to higher frequencies upon W insertion, which has been previously a ributed to the unit cell expansion [91].However, (in disagreement with this last study) no linear shift in this Raman mode with the increase in the W content, saturating at 15 mol% W, was observed in our work.In contrast, the shift appeared to be significant between 1 and 6 mol% W, with no subsequent evolution upon greater W insertion.The second particularity concerns the blue shift in the low-frequency A1g/B1g and high-energy Eg modes with an increase in W content above 6 mol%.This may be an indication of the cell deformation toward rutile geometry.In agreement with the XRD analysis, the Raman spectra (broad bands at 260, 320 and 690 cm −1 , marked by (*) in Figure 5) confirm the material segregation above 15 mol% W, resulting in anatase TiO2/monoclinic WO3 [92] polymorph powders.In contrast, the dissolution of W into te anatase TiO2 matrix is complete for the lower W contents.Because of that, we tentatively ascribed 15 mol% to the solubility limit of tungsten in anatase titania.Two particularities of the prepared materials can be observed from the Raman analysis in Figure 5.The first one is that the low-energy Eg mode in the range of ~150 cm −1 shifts to higher frequencies upon W insertion, which has been previously a ributed to the unit cell expansion [91].However, (in disagreement with this last study) no linear shift in this Raman mode with the increase in the W content, saturating at 15 mol% W, was observed in our work.In contrast, the shift appeared to be significant between 1 and 6 mol% W, with no subsequent evolution upon greater W insertion.The second particularity concerns the blue shift in the low-frequency A1g/B1g and high-energy Eg modes with an increase in W content above 6 mol%.This may be an indication of the cell deformation toward rutile geometry.In agreement with the XRD analysis, the Raman spectra (broad bands at 260, 320 and 690 cm −1 , marked by (*) in Figure 5) confirm the material segregation above 15 mol% W, resulting in anatase TiO2/monoclinic WO3 [92] polymorph powders.In contrast, the dissolution of W into te anatase TiO2 matrix is complete for the lower W contents.Because of that, we tentatively ascribed 15 mol% to the solubility limit of tungsten in anatase titania.

Adsorption Capacity
The MB adsorption by W-TiO 2 nanopowders of different compositions calcined at 550 • C is shown in Figure 6 in terms of C/C 0 , where C and C 0 are, respectively, the remaining and initial MB concentrations in the solution.It can be seen that W-TiO 2 materials possess a stronger MB adsorption capacity than pure TiO 2 .The strongest adsorption was observed in the material with 15 mol% W, which was about 50 times above that of pure TiO 2 .This adsorption enhancement is significantly higher than that previously reported in W-TiO 2 powders prepared through ammonium metatungstate liquid deposition onto TiO 2 Degussa P25 powder, where an eightfold increase in MB adsorption has been reported in a 6.5 mol% W material [80].This difference in both optimal composition and enhancement factor can be clearly attributed to the more homogeneous W distribution in the material in our synthesis conditions, where W species interact with much smaller reactive TOA nuclei of 3.2 nm size [84] compared to tens of nanometers of TiO 2 Degussa P25 [93].The highest adsorption capacity of W-TiO 2 nanopowder was 32 mg/g of MB, attained after 15 min of the experiment.This corresponds to the adsorption capacity of the coal-based activated carbon in treating high-salt printing and dyeing wastewater [94], which, however, showed significantly slower kinetics.This shows good potential of the prepared materials in environmental processes.
pure TiO2.This adsorption enhancement is significantly higher than that previously reported in W-TiO2 powders prepared through ammonium metatungstate liquid deposition onto TiO2 Degussa P25 powder, where an eightfold increase in MB adsorption has been reported in a 6.5 mol% W material [80].This difference in both optimal composition and enhancement factor can be clearly a ributed to the more homogeneous W distribution in the material in our synthesis conditions, where W species interact with much smaller reactive TOA nuclei of 3.2 nm size [84] compared to tens of nanometers of TiO2 Degussa P25 [93].The highest adsorption capacity of W-TiO2 nanopowder was 32 mg/g of MB, a ained after 15 min of the experiment.This corresponds to the adsorption capacity of the coal-based activated carbon in treating high-salt printing and dyeing wastewater [94], which, however, showed significantly slower kinetics.This shows good potential of the prepared materials in environmental processes.The adsorption abilities of W-TiO2 nanopowders calcined at different temperatures of 450, 500, 550 and 600 °C were further compared.The equilibrium absorbance Nads/N0 according to the Lagergren model [95] and Langmuir isotherm correction of the photocatalysts' mass (equal to 0.125 g/L, as that used in the photocatalytic MB decomposition experiments) is plo ed in Figure 7.As one can see, the adsorption capacity of the prepared W-TiO2 nanopowders increased with an increase in W content up to 15 mol% and then reduced with the greater W loadings.The maximum adsorption capacity of 0.25 mM/g is 2.5 times larger compared to that previously reported [80].At the same time, we noticed that the adsorption of the prepared materials was not significantly affected by the calcination temperature.The adsorption abilities of W-TiO 2 nanopowders calcined at different temperatures of 450, 500, 550 and 600 • C were further compared.The equilibrium absorbance N ads /N 0 according to the Lagergren model [95] and Langmuir isotherm correction of the photocatalysts' mass (equal to 0.125 g/L, as that used in the photocatalytic MB decomposition experiments) is plotted in Figure 7.As one can see, the adsorption capacity of the prepared W-TiO 2 nanopowders increased with an increase in W content up to 15 mol% and then reduced with the greater W loadings.The maximum adsorption capacity of 0.25 mM/g is 2.5 times larger compared to that previously reported [80].At the same time, we noticed that the adsorption of the prepared materials was not significantly affected by the calcination temperature.We noticed that the specific surface area (σ) of the prepared materials measured by using the BET method (Figure S2 of Supplementary Information) confirmed the tendency in their MB adsorption capacity: σ = 50 ± 3 m²/g in nanopowders with 0 ≤ CW/(CW + CTi) ≤ 4 mol% calcined at 550 °C, while it sharply increased to 171 m²/g in nanopowders with 15 We noticed that the specific surface area (σ) of the prepared materials measured by using the BET method (Figure S2 of Supplementary Materials) confirmed the tendency in their MB adsorption capacity: σ = 50 ± 3 m²/g in nanopowders with 0 ≤ C W /(C W + C Ti ) ≤ 4 mol% calcined at 550 • C, while it sharply increased to 171 m²/g in nanopowders with 15 mol% W. Following the decreasing tendency in crystallite size (Figure 4), these measurements showed the important modifications in the porosity of the material with the composition W x=0.15 -TiO 2 , which has already been reported in mixed oxide materials [96].Otherwise, the peak of the porosity correlates with the presumed solubility limit of W in the TiO 2 matrix (see the XRD measurements) and may be connected to the phase separation phenomena at the component segregation point.This issue needs further investigation.

Photocatalytic Activity
After completing the adsorption measurements, the photocatalytic activity of the W-TiO 2 nanopowders was investigated.These measurements were performed with the light source "on" after attaining the equilibrium of adsorption kinetics (Section 3.2).The results of MB removal using photocatalysts annealed at 550 • C with UV-A lamp illumination are shown in Figure 8a.The MB degradation kinetics followed first-order reaction kinetics ln(C/C 0 ) = −kt, where C 0 , C, k and t are, respectively, the initial and remaining pollutant concentrations, the rate constant and the process time (see Figure S3 of Supplementary Materials).The obtained rate constants were used as a measure of the photocatalysts' activities, which are shown in Figure 8b versus the composition x of the synthetized compounds.The materials' activity exhibits two maxima, at low and high tungsten contents.While the strongest one observed at 2 mol% W has already been reported [80,91], the second maximum at 15 mol% W is new.The low doping of TiO 2 by different cations generally produced an activity increase, in which the contribution of the photoinduced charge separation seems to play an important role [24].On the other hand, the compositions corresponding to maximum adsorption (Figure 7) have not been reported as highly active so far.We also notice negligible direct MB photolysis by UV-A light.The photocatalytic rate constants of the prepared W-TiO2 nanopowders calcined at different temperatures are shown in Figure 9.One can see that the nanopowder activity does not correlate with that of the most active TiO2 component.Indeed, the pure TiO2 catalyst had the highest activity after calcination at temperatures of 450-500 °C, which corresponds to the pure anatase phase formation; in contrast, its calcination at elevated temperatures of 550 and, most evidently, at 600 °C resulted in an activity decrease, which is explained by the rutile phase formation [97].Furthermore, an addition of tungsten extends the optimal activity to 550 °C (2 mol% W).Except for the activity increase upon low The photocatalytic rate constants of the prepared W-TiO 2 nanopowders calcined at different temperatures are shown in Figure 9.One can see that the nanopowder activity does not correlate with that of the most active TiO 2 component.Indeed, the pure TiO 2 catalyst had the highest activity after calcination at temperatures of 450-500 • C, which corresponds to the pure anatase phase formation; in contrast, its calcination at elevated temperatures of 550 and, most evidently, at 600 • C resulted in an activity decrease, which is explained by the rutile phase formation [97].Furthermore, an addition of tungsten extends the optimal activity to 550 • C (2 mol% W).Except for the activity increase upon low cation doping of a few mol% (followed by the activity decrease) taking place in many cation-doped TiO 2 -based materials, an evident second greatest enhancement in activity appearing around the concentration of 15 mol% clearly showed the added value of the mixed-oxide material.The photocatalytic rate constants of the prepared W-TiO2 nanopowders calcined at different temperatures are shown in Figure 9.One can see that the nanopowder activity does not correlate with that of the most active TiO2 component.Indeed, the pure TiO2 catalyst had the highest activity after calcination at temperatures of 450-500 °C, which corresponds to the pure anatase phase formation; in contrast, its calcination at elevated temperatures of 550 and, most evidently, at 600 °C resulted in an activity decrease, which is explained by the rutile phase formation [97].Furthermore, an addition of tungsten extends the optimal activity to 550 °C (2 mol% W).Except for the activity increase upon low cation doping of a few mol% (followed by the activity decrease) taking place in many cation-doped TiO2-based materials, an evident second greatest enhancement in activity appearing around the concentration of 15 mol% clearly showed the added value of the mixed-oxide material.The photocatalytic activity of the prepared W-TiO 2 materials was additionally evaluated with natural sunlight illumination.In order to soften the influence of the daily sunlight intensity variations, these experiments were carried out simultaneously at the same time for the complete series of W-TiO 2 nanopowders with different compositions.The results of the MB decomposition kinetics are shown in Figure 10a for the material calcined at 550 • C. The data followed first-order kinetics (Figure S3 of Supplementary Materials), and the rate constants are presented in Figure 10b as a function of the composition x of the synthetized compounds.
The comparison of the W-TiO 2 photocatalysts' activities under sunlight (Figure 10b) and UV-A (Figure 8b) illuminations confirms the existence of two maxima, with a tungsten content of W 1 = 2 mol% and W 2 = 15 mol%.At the same time, the relative peak values of the corresponding rate constants are dependent on the light source.Indeed, while with UV-A illumination the rate constant k UV (W 1 ) = 2.6 × 10 −3 min −1 was significantly higher than k UV (W 2 ) = 7.8 × 10 −4 min −1 , the materials' activities with sunlight illumination were both enhanced and close to each other: k SL (W 1 ) = 6.1 × 10 −3 min −1 versus k SL (W 2 ) = 5.1 × 10 −3 min −1 .Because, as previously discussed, W 2 composition corresponds to the peak adsorption, the peak in the photocatalytic activity might be related to the direct photolysis of the adsorbed MB molecules (~97% according to Figure 6) by sunlight.However, we reject this possibility because of the negligible MB decomposition in the blank photocatalytic tests.As a result, we attribute k SL (W 2 ) to the heterogeneous photocatalytic decomposition of MB by W x=0.15 -TiO 2 nanopowders.
sunlight intensity variations, these experiments were carried out simultaneously at the same time for the complete series of W-TiO2 nanopowders with different compositions.The results of the MB decomposition kinetics are shown in Figure 10a for the material calcined at 550 °C.The data followed first-order kinetics (Figure S3 of Supplementary Information), and the rate constants are presented in Figure 10b as a function of the composition x of the synthetized compounds.The comparison of the W-TiO2 photocatalysts' activities under sunlight (Figure 10b) and UV-A (Figure 8b) illuminations confirms the existence of two maxima, with a tungsten content of W1 = 2 mol% and W2 = 15 mol%.At the same time, the relative peak values of the corresponding rate constants are dependent on the light source.Indeed, while with UV-A illumination the rate constant kUV(W1) = 2.6 × 10 −3 min −1 was significantly higher than kUV(W2) = 7.8 × 10 −4 min −1 , the materials' activities with sunlight illumination were both enhanced and close to each other: kSL(W1) = 6.1 × 10 −3 min −1 versus kSL(W2) = 5.1 × 10 −3 min −1 .Because, as previously discussed, W2 composition corresponds to the peak adsorption, the peak in the photocatalytic activity might be related to the direct photolysis of the adsorbed MB molecules (~97% according to Figure 6) by sunlight.However, we reject this possibility because of the negligible MB decomposition in the blank photocatalytic tests.As a result, we a ribute kSL(W2) to the heterogeneous photocatalytic decomposition of MB by Wx=0.15-TiO2 nanopowders.

Synergy of Adsorption and Photocatalysis
The heterogeneous photocatalytic process basically includes two steps of (1) physical adsorption of pollutant molecules on the material surface and (2) surface reactions including the release of reaction products.Since, in common experiments, the material activation by light is applied after a aining adsorption equilibrium, the contributions of

Synergy of Adsorption and Photocatalysis
The heterogeneous photocatalytic process basically includes two steps of (1) physical adsorption of pollutant molecules on the material surface and (2) surface reactions including the release of reaction products.Since, in common experiments, the material activation by light is applied after attaining adsorption equilibrium, the contributions of these steps can be, respectively, characterized by the adsorbed amount of material C (in ppm units) and the apparent reaction rate constant k (in min −1 units).
In order to inspect the reaction constant of the materials with different compositions, TRMC measurements were performed.The decay curves of the photoinduced charges in pure and W-doped TiO 2 nanopowders are shown in Figure 11.The observed signal has been previously attributed to the liberation of charges from shallow traps with binding energy E * following a power law decay I = At α with parameter α = k B T/E * , where k B is the Boltzmann constant and T is temperature [23].The shallow states have attracted much attention over the last decade because of their key influence on the photocatalysts' activity [98,99].According to our data, the depths of the shallow electron traps increased from 169 meV (0 mol% W) to 183 meV (1 mol% W), while they decreased afterwards to 95 meV (4 mol% W) and 73 meV (15 mol% W).In nanopowders with x = 100 mol% (W-only material), the signal disappeared very rapidly to the noise level.Although no quantitative relation between the charges' lifetime and photocatalytic rate has been proposed in previous studies, the present TRMC measurements confirm the observed tendency in the activity of the prepared powders, showing that photoinduced charges do not effectively trigger the reactions when they disappear fast.
A comparison with previous results summarized in Table 1 showed a general agreement about the enhancement factor of 2.5 ± 0.5 after the insertion of ~2 mol% W into TiO 2 , with both UV-A and sunlight illuminations.According to the TRMC measurements, the intrinsic material reactivity can be responsible for the maximum activity W 1 (Figures 8 and 10) but cannot be the only reason explaining the W 2 maximum activity.Below, we discuss this issue in more detail.
(4 mol% W) and 73 meV (15 mol% W).In nanopowders with x = 100 mol% (W-only material), the signal disappeared very rapidly to the noise level.Although no quantitative relation between the charges' lifetime and photocatalytic rate has been proposed in previous studies, the present TRMC measurements confirm the observed tendency in the activity of the prepared powders, showing that photoinduced charges do not effectively trigger the reactions when they disappear fast.A comparison with previous results summarized in Table 1 showed a general agreement about the enhancement factor of 2.5 ± 0.5 after the insertion of ~2 mol% W into TiO2, with both UV-A and sunlight illuminations.According to the TRMC measurements, the intrinsic material reactivity can be responsible for the maximum activity W1 (Figures 8 and 10) but cannot be the only reason explaining the W2 maximum activity.Below, we discuss this issue in more detail.
The material activity in Figures 8 and 10 generally reflects the combined action of photocatalytic and adsorption processes.In order to distinguish these two contributions, we normalized the decomposition rate constants (in min −1 units) for the adsorbed component (in ppb units) and plo ed kNORM (in min −1 ppb −1 units) versus the tungsten content in the W-TiO2 nanopowders in Figure 12a.This rate constant kNORM apparently reflects a pure photocatalytic response of the material to the adsorbed MB pollutant.Two particularities can be noticed:  The material activity in Figures 8 and 10 generally reflects the combined action of photocatalytic and adsorption processes.In order to distinguish these two contributions, we normalized the decomposition rate constants (in min −1 units) for the adsorbed component (in ppb units) and plotted k NORM (in min −1 ppb −1 units) versus the tungsten content in the W-TiO 2 nanopowders in Figure 12a.This rate constant k NORM apparently reflects a pure photocatalytic response of the material to the adsorbed MB pollutant.Two particularities can be noticed: k NORM was greatly different for 0 ≤ C W /(C W + C Ti ) ≤ 0.02 and 0.04 ≤ C W /(C W + C Ti ) ≤ 1; -k NORM was stronger under sunlight illumination compared to UV-A, attaining the peak value at 15 mol% W.
As we can see, the insertion of W into the TiO2 matrix does not enhance the intrinsic response of the mixed oxide material and the photocatalytic activity is most effective on pure anatase TiO2.Furthermore, the material with a W content of 4 mol% and higher has a greatly reduced kNORM, which drops down by about 30 times to a lower level.
At the same time, the pollutant decomposition takes place at the material surface and the decomposition kinetics can be enhanced by increasing the number of pollutant molecules at the photocatalyst surface.A strong increase in the material adsorption due to the compositional adjustment may therefore promote synergy between the two processes (adsorption and reaction), moving toward the highest total effectiveness of the pollutant removal process.Taking into account a much stronger (by almost 100 times) initial elimination rate of MB via adsorption (Figure 6) compared to photocatalysis (Figures 8 and 10), one can suggest preferential use of 15 mol% W-doped TiO2 nanopowders in the environmental process, especially in emergency cases.Indeed, a large amount of a pollutant can be rapidly bound to the photocatalyst surface in the first few minutes after an accident, while its decomposition can be readily achieved within a longer timescale.This would significantly reduce the negative impact of an accident on nature.
The reaction inhibition in W-TiO2 nanopowders (see in Figure 12a), which probably is due to faster recombination of photoinduced electron-hole pairs (as confirmed by the TRMC data in Figure 11), can be largely compensated for by the material's adsorption rate and capacity.The synergy factor (R) earlier proposed for the two-component composite TiO2-AC (AC = activated carbon) materials [20] cannot be directly applied to the one-component system, in which both reaction and absorption rates are tuned via elemental composition.However, taking it as a ratio of the pollutant removal rates R = k(W-TiO2)/k(TiO2), one can obtain R~10 at the beginning of the process, converging to R~2 after longer durations of the stationary operation under illumination, when adsorpequilibrium is a ained.This shows the high potential of this material for depollution use in emergency cases of an occasional pollutant leak.The very high concentration of adsorbed pollutant molecules can largely compensate for the reduced reaction constant and results in effective pollutant removal.As a result, the Wx=0.15-TiO2nanocatalyst could be highly promising for wastewater purification in industrial plants.As we can see, the insertion of W into the TiO 2 matrix does not enhance the intrinsic response of the mixed oxide material and the photocatalytic activity is most effective on pure anatase TiO 2 .Furthermore, the material with a W content of 4 mol% and higher has a greatly reduced k NORM , which drops down by about 30 times to a lower level.
At the same time, the pollutant decomposition takes place at the material surface and the decomposition kinetics can be enhanced by increasing the number of pollutant molecules at the photocatalyst surface.A strong increase in the material adsorption due to the compositional adjustment may therefore promote synergy between the two processes (adsorption and reaction), moving toward the highest total effectiveness of the pollutant removal process.Taking into account a much stronger (by almost 100 times) initial elimination rate of MB via adsorption (Figure 6) compared to photocatalysis (Figures 8 and 10), one can suggest preferential use of 15 mol% W-doped TiO 2 nanopowders in the environmental process, especially in emergency cases.Indeed, a large amount of a pollutant can be rapidly bound to the photocatalyst surface in the first few minutes after an accident, while its decomposition can be readily achieved within a longer timescale.This would significantly reduce the negative impact of an accident on nature.
The reaction inhibition in W-TiO 2 nanopowders (see in Figure 12a), which probably is due to faster recombination of photoinduced electron-hole pairs (as confirmed by the TRMC data in Figure 11), can be largely compensated for by the material's adsorption rate and capacity.The synergy factor (R) earlier proposed for the two-component composite TiO 2 -AC (AC = activated carbon) materials [20] cannot be directly applied to the one-component system, in which both reaction and absorption rates are tuned via elemental composition.However, taking it as a ratio of the pollutant removal rates R = k(W-TiO 2 )/k(TiO 2 ), one can obtain R~10 at the beginning of the process, converging to R~2 after longer durations of the stationary operation under illumination, when adsorption equilibrium is attained.This shows the high potential of this material for depollution use in emergency cases of an occasional pollutant leak.The very high concentration of adsorbed pollutant molecules can largely compensate for the reduced reaction constant and results in effective pollutant removal.As a result, the W x=0.15 -TiO 2 nanocatalyst could be highly promising for wastewater purification in industrial plants.
The observed difference in the photocatalytic effectiveness of the W-TiO 2 materials under UV-A and sunlight illuminations (Figure 12b) seems to be a general feature (see, e.g., V-TiO 2 photocatalyst [22]).We noticed that a great enhancement in the MB decomposition rate by anatase TiO 2 (0 mol% W), k sunlight /k UV-A ≈ 1.7 (Figure 12b), can be explained by tge greater intensity of the UV-A component in sunlight as compared to the laboratory lamp.However, the W insertion further increased this ratio up to 6.5 for 15 mol% W, which can be a sign of photocatalyst activation by the visible light.Since WO 3 does not crystallize in the material at W ≤ 15 mol%, this activation is believed to involve intraband impurity states.The activation of the W x=0.15 -TiO 2 nanocatalyst by sunlight can be an additional argument supporting its use in environmental catalysis.

Conclusions
Photocatalytic W-TiO 2 nanopowders with a large variation in the composition x = C W /(C W + C Ti ), including end members TiO 2 and WO 3 , were successfully synthesized by using the sol-gel method, using TTIP and WCl 6 precursors and a low hydrolysis ratio h = C H2O /(C Ti + C W ) = 1.25, following calcination at 450, 500, 550 and 600 • C to form a crystalline product.The materials preserved the anatase crystalline structure, while the crystalline size decreased with an increase in W content up to 0.15 mol%; the nucleation of WO 3 crystallites was observed with the greater W loadings.The structural characterizations suggest that the solubility limit of W into anatase TiO 2 is 15 mol% W. The adsorption and photocatalytic kinetics of the prepared materials were evaluated based on MB removal in aqueous solutions.Both adsorption kinetics and adsorption capacity strongly increased in the material with 15 mol% W, in which the maximum adsorption capacity reached 0.25 mM/g, which is almost 50 times higher compared to that of pure anatase TiO 2 .Overall, the rates of MB adsorption and photocatalytic decomposition were stronger in, respectively, W-rich and Ti-rich compositions.A comparison of the photocatalytic activities under UV-A and sunlight illuminations revealed two optimal compositions with tungsten contents of W 1 = 2 mol% and W 2 = 15 mol%.Our analysis permitted us to attribute them to the adsorption/photocatalytic decomposition of MB molecules at the W-TiO 2 photocatalyst surface.The normalization of decomposition rates for the adsorbed amount of MB showed (i) the conservation of the intrinsic TiO 2 material's activity at low doping 0 ≤ W ≤ 0.02 and (ii) a decrease in the activity with greater W insertion.The compensation for the activity decrease by the adsorption increase was observed in the material's pollutant removal process.
The synergy of adsorption and photocatalysis in the water purification process was shown, which revealed the material with 15 mol% W to be the most effective in terms of pollutant removal.Furthermore, an additional activation of the W-TiO 2 photocatalyst by sunlight compared to a UV-A lamp was shown.To summarize, the mixed oxide W x=0.15 -TiO 2 nanocatalyst appeared to be promising for applications in the environmental process of pollutant removal in aqueous media, especially in emergency cases involving a strong peak in the pollutant concentration.

Figure 1 .
Figure 1.Procedure for synthesis of W-TiO2 nanopowders.A and B are isopropanol solutions of TTIP and WCl6, respectively.

Figure 1 .
Figure 1.Procedure for synthesis of W-TiO 2 nanopowders.A and B are isopropanol solutions of TTIP and WCl 6 , respectively.

Figure 2 .
Figure 2. X-ray diffraction pa erns of TiO2 nanopowders with 0 mol% (a) and 1 mol% (b) W doping after calcination at temperatures between 450 and 600 °C.Vertical bars show positions and heights of anatase (a) and rutile (r) peaks.

Figure 2 .
Figure 2. X-ray diffraction patterns of TiO 2 nanopowders with 0 mol% (a) and 1 mol% (b) W doping after calcination at temperatures between 450 and 600 • C. Vertical bars show positions and heights of anatase (a) and rutile (r) peaks.

Figure 2 .
Figure 2. X-ray diffraction pa erns of TiO2 nanopowders with 0 mol% (a) and 1 mol% (b) W doping after calcination at temperatures between 450 and 600 °C.Vertical bars show positions and heights of anatase (a) and rutile (r) peaks.

Figure 4 .
Figure 4. Crystalline size of W-TiO2 powders versus W content after calcination at different temperatures (obtained from XRD pa erns by Scherrer equation).

Figure 5 .
Figure 5. Raman spectra of W-TiO2 nanopowders with different W contents after calcination at 550 °C.Positions of appearing WO3 bands at 30 mol% W are shown by (*). 0

Figure 4 .
Figure 4. Crystalline size of W-TiO2 powders versus W content after calcination at different temperatures (obtained from XRD pa erns by Scherrer equation).

Figure 5 .
Figure 5. Raman spectra of W-TiO2 nanopowders with different W contents after calcination at 550 °C.Positions of appearing WO3 bands at 30 mol% W are shown by (*). 0

Figure 12 .
Figure 12.Normalized decomposition rate constant kNORM for adsorbed MB by W-TiO2 photocatalyst calcined at 550 °C under UV-A and sunlight illuminations (a), and ratio of normalized rate constants under sunlight and UV-A illuminations (b).

Figure 12 .
Figure 12.Normalized decomposition rate constant k NORM for adsorbed MB by W-TiO 2 photocatalyst calcined at 550 • C under UV-A and sunlight illuminations (a), and ratio of normalized rate constants under sunlight and UV-A illuminations (b).

Table 1 .
Preparation method and photocatalytic activity of W-TiO 2 nanomaterials.
Raman spectra of W-TiO 2 nanopowders with different W contents after calcination at 550 • C. Positions of appearing WO 3 bands at 30 mol% W are shown by (*).