Defective TiO 2 Core-Shell Magnetic Photocatalyst Modiﬁed with Plasmonic Nanoparticles for Visible Light-Induced Photocatalytic Activity

: In the presented work, for the ﬁrst time, the metal-modiﬁed defective titanium(IV) oxide nanoparticles with well-deﬁned titanium vacancies, was successfully obtained. Introducing platinum and copper nanoparticles (NPs) as surface modiﬁers of defective d-TiO 2 signiﬁcantly increased the photocatalytic activity in both UV-Vis and Vis light ranges. Moreover, metal NPs deposition on the magnetic core allowed for the e ﬀ ective separation and reuse of the nanometer-sized photocatalyst from the suspension after the treatment process. The obtained Fe 3 O 4 @SiO 2 / d-TiO 2 -Pt / Cu photocatalysts were characterized by X-ray di ﬀ ractometry (XRD) and speciﬁc surface area (BET) measurements, UV-Vis di ﬀ use reﬂectance spectroscopy (DR-UV / Vis), X-ray photoelectron spectroscopy (XPS) and transmission electron microscopy (TEM). Further, the mechanism of phenol degradation and the role of four oxidative species (h + , e − , • OH, and • O 2 − ) in the studied photocatalytic process were investigated.


Introduction
In recent years, among wastewater treatment and environmental remediation technologies, photocatalysis has gained attention as a promising technique for the degradation of persistent organic pollutants at ambient temperature and pressure [1][2][3][4]. Pilot scale-installations for water treatment using photocatalysis are more and more popular among the world [5,6]. The structural and surface properties of photocatalysts significantly influence their physicochemical and photocatalytic properties. In this regard, one of the most important issues in the photocatalytic process is the preparation of well-characterized and highly active photocatalytic material.
Titanium(IV) oxide (TiO 2 ), the most widely used semiconductor in photocatalysis, is extensively exploited to obtain highly photoactive in UV-Vis range semiconductor material. The TiO 2 nanoparticles differing in size and surface area. Nonetheless, despite different morphology and polymorphic composition, all pristine titanium(IV) oxide particles own wide bandgap energy (Eg), which differs in the range of 3.0-3.2 eV, for rutile and anatase, respectively [7]. In this regard, TiO 2 photoexcitation is possible only with UV irradiation (λ < 388 nm), and therefore the application of solar radiation is highly limited. The general physicochemical and photocatalytic characteristics of the obtained defective d-TiO2-Pt/Cu and Fe3O4@SiO2/d-TiO2-Pt/Cu samples, i.e., BET surface area, pore volume, calculated bandgap (Eg) and phenol degradation efficiency in UV-Vis and Vis light are presented in Tables 1 and 2.  The general physicochemical and photocatalytic characteristics of the obtained defective d-TiO 2 -Pt/Cu and Fe 3 O 4 @SiO 2 /d-TiO 2 -Pt/Cu samples, i.e., BET surface area, pore volume, calculated bandgap (Eg) and phenol degradation efficiency in UV-Vis and Vis light are presented in Tables 1 and 2.  26 11 The BET surface area of pure TiO 2 obtained from Titanium(IV) butoxide (TBT) hydrolysis in water and defective TiO 2 samples was similar and ranged from 167 to 172 m 2 ·g −1 . The specific surface area of the metal-modified d-TiO 2 samples fluctuated from 166 to 101 m 2 ·g −1 and depended on the type and amount of metallic species deposited on d-TiO 2 surface. The samples modified with Pt NPs revealed a higher BET surface area of about 148 m 2 ·g −1 compared to d-TiO 2 modified with copper oxide (101 m 2 ·g −1 ), and bimetallic Pt/Cu NPs (152 m 2 ·g −1 ). The relations between photoactivity in UV-Vis and Vis light range versus BET surface area are also shown in Tables 1 and 2. The obtained results indicated that not so much the surface area but rather the presence of Ti defects and modification with metal nanoparticles caused the enhanced photoactivity of the obtained photocatalysts. Moreover, as shown in Table 2, the addition of surface-modifying metal nanoparticles, as well as further deposition of d-TiO 2 -Pt/Cu on magnetic matrice, did not affect the magnitude order of the BET surface area, which remained in the range of 101 to 172 m 2 ·g −1 for d-TiO 2 _20-Cu0.1 and d-TiO 2 _20, respectively.
The energy bandgaps for all samples were calculated from the plot of (Kubelka-Munk·E) 0.5 versus E, where E is energy equal to hv, and summarized in Table 1. The samples consist of defective TiO 2 exhibited narrower bandgap of 2.7-2.9 eV compared to TiO 2 and TiO 2 -Pt0.05 photocatalysts. Moreover, for all metal-modified defective photocatalysts, the bandgap value, calculated from Kubelka-Munk, transformation did not change, compared to d-TiO 2 matrice, indicating surface modification than doping [11].
The XRD patterns for selected d-TiO 2 -Pt/Cu and Fe 3 O 4 @SiO 2 /d-TiO 2 -Pt/Cu samples are presented in Figures 1 and 2, with a detailed phase composition and crystalline sizes for all photocatalysts being listed in Tables 3 and 4. Peaks marked "A", "R", and "B" corresponds to anatase, rutile, and brookite phases, respectively. Both crystalline structures (anatase and brookite) appeared for pure TiO 2 prepared by the sol-gel method. For Pt-modified TiO 2 anatase was the major phase, whereas brookite existed as the minor phase. The average crystallite size of anatase was 5-6 nm. The preparation of d-TiO 2 photocatalysts proceeded in the oxidative environment. The introduction into the crystal structure of various types of defects promotes the transformation of anatase to rutile at lower temperatures. Therefore, for the samples obtained in the presence of HIO 3 as the oxidizing agent, after the annealing process the percentage of anatase (the most intense peak at 25 • 2θ, with the (101) plane diffraction, ICDD's card No. 7206075) was decreased in favor of (110) rutile, with the peak at 31 • 2θ (ICDD's card No. 9004141), even below the anatase to rutile phase transformation temperature [35][36][37]. For the samples d-TiO 2 _75 and d-TiO 2 _75-Pt0.05, the dominant phase was rutile with a crystallite size of about 6 nm. Further, the surface modification with plasmonic platinum and semi-noble copper did not cause changes in anatase crystallite size, remaining about 5-6 nm size. The percentage of the brookite phase increased to 8.5% and 13% for d-TiO 2 _20-Pt0.1/Cu0.1, and d-TiO 2 _20-Pt0.1 samples, respectively. It resulted from the additional thermal treatment after metal nanoparticles deposition on the photocatalyst surface. Moreover, Pt and Cu modification of TiO 2 did not cause the shift of the peaks in the XRD pattern. The presence of platinum and copper deposited on TiO 2 was not approved (no peaks for platinum or copper) due to low content (0.05-0.1 mol%) and nanometric size.  The XRD analysis of Fe3O4@SiO2/d-TiO2-Pt/Cu confirmed the formation of a magnetic composite, and, as observed in Figure 3 and Table 4, there was no significant difference between the diffraction patterns of the obtained magnetic photocatalysts modified with Pt/Cu NPs. The presence of pure magnetite, with diffraction peaks at 30.2°, 35.6°, 43.3°, 57.3°, and 62.9° 2θ corresponding to (220), (311), (400), (511), and (440) cubic inverse spinel planes (ICDD's card No. 9005813) was confirmed for all Fe3O4@SiO2/d-TiO2-Pt/Cu magnetic photocatalysts. The decrease in Fe3O4 peaks intensity was caused by the formation of tight non-magnetic shell on the core surface, which was previously described by Zielińska-Jurek et al. [26]. The broad peak at 15-25° 2θ corresponds to amorphous silica [38,39]. The content of the magnetite crystalline phase varied from 21% to 28% for Fe3O4@SiO2/d-TiO2_20-Pt0.05 and Fe3O4@SiO2/d-TiO2_20-Cu0.1, respectively. At the same time, TiO2 crystallite size and anatase to rutile phase content ratio remained unchanged for Fe3O4@SiO2/d-TiO2_20 and Fe3O4@SiO2/d-TiO2_20-Pt/Cu samples. No other crystalline phases were identified in the XRD patterns, which indicated the crystal purity of the obtained composites.   The XRD analysis of Fe 3 O 4 @SiO 2 /d-TiO 2 -Pt/Cu confirmed the formation of a magnetic composite, and, as observed in Figure 3 and Table 4, there was no significant difference between the diffraction patterns of the obtained magnetic photocatalysts modified with Pt/Cu NPs. The presence of pure magnetite, with diffraction peaks at 30.2 • , 35.6 • , 43.3 • , 57.3 • , and 62.9 • 2θ corresponding to (220), (311), (400), (511), and (440) cubic inverse spinel planes (ICDD's card No. 9005813) was confirmed for all Fe 3 O 4 @SiO 2 /d-TiO 2 -Pt/Cu magnetic photocatalysts. The decrease in Fe 3 O 4 peaks intensity was caused by the formation of tight non-magnetic shell on the core surface, which was previously described by Zielińska-Jurek et al. [26]. The broad peak at 15-25 • 2θ corresponds to amorphous silica [38,39]. The content of the magnetite crystalline phase varied from 21% to 28% for Fe 3 O 4 @SiO 2 /d-TiO 2 _20-Pt0.05 and Fe 3 O 4 @SiO 2 /d-TiO 2 _20-Cu0.1, respectively. At the same time, TiO 2 crystallite size and anatase to rutile phase content ratio remained unchanged for Fe 3 O 4 @SiO 2 /d-TiO 2 _20 and Fe 3 O 4 @SiO 2 /d-TiO 2 _20-Pt/Cu samples. No other crystalline phases were identified in the XRD patterns, which indicated the crystal purity of the obtained composites.  The photoabsorption properties of metal-modified defective d-TiO2 samples were studied by diffuse reflectance spectroscopy, and exemplary data are shown in Figure 4. Comparing to pure TiO2 photocatalyst, introducing platinum as a surface modifier caused an increase of absorption in the visible light region, however, without shifting a maximum, as presented for sample TiO2-Pt0.05. Modification of defective d-TiO2 with Pt and Cu was associated with a further increase of Vis light absorbance and proportional to the amount of the deposited metal. Moreover, the deposition of Pt caused a more significant absorbance increment than the same modification with Cu species.
Defective d-TiO2-Pt/Cu deposited on Fe3O4@SiO2 core were characterized by extended light absorption ranged to 700 nm. It could be observed that the described absorption properties in the Vis light for metal-modified TiO2 and absorption properties of final composites have been preserved. 15   The photoabsorption properties of metal-modified defective d-TiO 2 samples were studied by diffuse reflectance spectroscopy, and exemplary data are shown in Figure 4. Comparing to pure TiO 2 photocatalyst, introducing platinum as a surface modifier caused an increase of absorption in the visible light region, however, without shifting a maximum, as presented for sample TiO 2 -Pt0.05. Modification of defective d-TiO 2 with Pt and Cu was associated with a further increase of Vis light absorbance and proportional to the amount of the deposited metal. Moreover, the deposition of Pt caused a more significant absorbance increment than the same modification with Cu species.
Modification of defective d-TiO2 with Pt and Cu was associated with a further increase of Vis light absorbance and proportional to the amount of the deposited metal. Moreover, the deposition of Pt caused a more significant absorbance increment than the same modification with Cu species.
Defective d-TiO2-Pt/Cu deposited on Fe3O4@SiO2 core were characterized by extended light absorption ranged to 700 nm. It could be observed that the described absorption properties in the Vis light for metal-modified TiO2 and absorption properties of final composites have been preserved.  The presence of Localized Surface Plasmon Resonance (LSPR) peaks for Pt and Cu were confirmed based on DR-UV/Vis spectra measurements with pure TiO2 as a reference (see in Figure  4c). Platinum surface plasmon resonance was observed at the wavelength of about 410-420 nm [33,40]. Electron transfer between Cu(II) and valence band of titanium(IV) oxide could be confirmed by absorption increment from 400 to 450 nm. The typical LSPR signal for zero valent copper at 500-580 nm was not observed, suggesting that Cu is mainly present in its oxidized forms [41,42].
To confirm the presence of noble metal and semi-noble metal NPs on defective TiO2 surface, the XPS analyses for the selected photocatalysts and deconvolution of Pt 4f and Cu 2p were performed,  Defective d-TiO 2 -Pt/Cu deposited on Fe 3 O 4 @SiO 2 core were characterized by extended light absorption ranged to 700 nm. It could be observed that the described absorption properties in the Vis light for metal-modified TiO 2 and absorption properties of final composites have been preserved.
The presence of Localized Surface Plasmon Resonance (LSPR) peaks for Pt and Cu were confirmed based on DR-UV/Vis spectra measurements with pure TiO 2 as a reference (see in Figure 4c). Platinum surface plasmon resonance was observed at the wavelength of about 410-420 nm [33,40]. Electron transfer between Cu(II) and valence band of titanium(IV) oxide could be confirmed by absorption increment from 400 to 450 nm. The typical LSPR signal for zero valent copper at 500-580 nm was not observed, suggesting that Cu is mainly present in its oxidized forms [41,42].
To confirm the presence of noble metal and semi-noble metal NPs on defective TiO 2 surface, the XPS analyses for the selected photocatalysts and deconvolution of Pt 4f and Cu 2p were performed, and the results are presented in Figure 5. Platinum species deposited on the titania surface were designated by deconvolution of Pt 4f peak into two components: Pt 4f 7/2 and Pt 4f 5/2 . According to the literature, Pt 4f 7/2 peak, with binding energies in the range of 74.2 to 75.0 eV, refers to the Pt 0 , while Pt 4f 5/2 peak, appearing at 77.5-77.9 eV is assigned to Pt 4+ [11]. The main peaks for Cu 2p appeared as Cu 2p 3/2 and Cu 2p 1/2 at 934 eV and 952 eV. Both of those peaks are commonly attributed to Cu + and Cu 2+ ions [13,43,44]. Obtained data indicated that both Pt and Cu species were successfully deposited on the titania surface.  Moreover, the presence of Pt NPs at the surface of the magnetic nanocomposites was also confirmed by microscopy analysis. As presented in Figure 6, the formation of SiO2/TiO2 shell, with a thickness of about 20 nm, tightly covering magnetite nanoparticles was observed. Platinum nanoparticles with a diameter of about 10-20 nm were evenly distributed on the d-TiO2 layer.  Moreover, the presence of Pt NPs at the surface of the magnetic nanocomposites was also confirmed by microscopy analysis. As presented in Figure 6, the formation of SiO 2 /TiO 2 shell, with a thickness of about 20 nm, tightly covering magnetite nanoparticles was observed. Platinum nanoparticles with a diameter of about 10-20 nm were evenly distributed on the d-TiO 2 layer.
Moreover, the presence of Pt NPs at the surface of the magnetic nanocomposites was also confirmed by microscopy analysis. As presented in Figure 6, the formation of SiO2/TiO2 shell, with a thickness of about 20 nm, tightly covering magnetite nanoparticles was observed. Platinum nanoparticles with a diameter of about 10-20 nm were evenly distributed on the d-TiO2 layer.  Among analyzed metal-modified photocatalysts, TiO 2 -Pt0.05 revealed the highest phenol degradation in UV-Vis light. After 60 min of irradiation, about 76% of phenol was degraded. After introducing plasmonic platinum and semi-noble copper species as a surface modifiers, UV-Vis photoactivity of defective d-TiO 2 samples increased to 59%. The degradation rate constant k increased to 1.47 × 10 −2 min −1 compared to d-TiO 2 _20 (0.79 × 10 −2 min −1 ), and d-TiO 2 _75 (0.52 × 10 −2 min −1 ) photocatalysts. Nonetheless, the most significant changes were observed during the photocatalytic process in visible light (λ > 420 nm). Modifying with 0.05 mol% of Pt, the surface of almost inactive in Vis light d-TiO 2 _75 resulted in three-times higher photocatalytic activity under visible light. Therefore, a highly positive effect of metal surface modification of defective d-TiO 2 photocatalyst surface was noticed. It resulted from better charge carriers' separation and decreasing the electron-hole recombination rate. Moreover, the narrower bandgap of the defective d-TiO 2 (in comparison with pure TiO 2 ) and modification with Pt possessing surface plasmon resonance properties, could also enhance visible light absorption and consequently led to photocatalytic activity increase.

Photocatalytic Activity of d-TiO2-Pt/Cu and Fe3O4@SiO2/d-TiO2-Pt/Cu Photocatalysts
The effect of Pt and Cu presence on the properties of defective d-TiO2 photocatalysts was evaluated in reaction of phenol degradation under UV-Vis and Vis light irradiation. The results, presented as the efficiency of phenol degradation as well as phenol degradation rate constant k, are given in Figures 7 and 8. Additionally, the effect of the electron (e − ), hole (h + ), hydroxyl radical ( • OH), and superoxide radical ( • O2 − ) scavengers were investigated and presented in Figure 9.
Among analyzed metal-modified photocatalysts, TiO2-Pt0.05 revealed the highest phenol degradation in UV-Vis light. After 60 min of irradiation, about 76% of phenol was degraded. After introducing plasmonic platinum and semi-noble copper species as a surface modifiers, UV-Vis photoactivity of defective d-TiO2 samples increased to 59%. The degradation rate constant k increased to 1.47 × 10 −2 min −1 compared to d-TiO2_20 (0.79 × 10 −2 min −1 ), and d-TiO2_75 (0.52 × 10 −2 min −1 ) photocatalysts. Nonetheless, the most significant changes were observed during the photocatalytic process in visible light (λ > 420 nm). Modifying with 0.05 mol% of Pt, the surface of almost inactive in Vis light d-TiO2_75 resulted in three-times higher photocatalytic activity under visible light. Therefore, a highly positive effect of metal surface modification of defective d-TiO2 photocatalyst surface was noticed. It resulted from better charge carriers' separation and decreasing the electronhole recombination rate. Moreover, the narrower bandgap of the defective d-TiO2 (in comparison with pure TiO2) and modification with Pt possessing surface plasmon resonance properties, could also enhance visible light absorption and consequently led to photocatalytic activity increase. Pure magnetite, coated with inert silica, did not affect the photocatalytic process. Furthermore, the Fe3O4@SiO2/d-TiO2 composite modified with Pt NPs, and bimetallic Pt/Cu NPs revealed the highest photocatalytic activity in Vis light range. The phenol degradation rate constant in Vis light was 2-times higher for Fe3O4@SiO2/d-TiO2-Pt/Cu compared to Fe3O4@SiO2/d-TiO2 sample. However, Pure magnetite, coated with inert silica, did not affect the photocatalytic process. Furthermore, the Fe 3 O 4 @SiO 2 /d-TiO 2 composite modified with Pt NPs, and bimetallic Pt/Cu NPs revealed the highest photocatalytic activity in Vis light range. The phenol degradation rate constant in Vis light was 2-times higher for Fe 3 O 4 @SiO 2 /d-TiO 2 -Pt/Cu compared to Fe 3 O 4 @SiO 2 /d-TiO 2 sample. However, the obtained magnetic photocatalysts had similar photocatalytic activity in UV-Vis light, almost regardless of the surface modification of d-TiO 2 with noble metals. It probably resulted from larger Pt particles (~20 nm) deposition at the surface of Fe 3 O 4 @SiO 2 /d-TiO 2 composite than for TiO 2 -Pt0.05 with particles size of about 2-3 nm. Previously, we have reported that the size of noble metal nanoparticles, especially platinum, deposited on the TiO 2 surface strictly depends on the semiconductor surface area, as well as its crystal lattice defects [33,45]. Fine metal particles are produced on the TiO 2 surface with a developed specific surface area with a high density of oxygen traps and nucleation sites, and the highest photocatalytic activity is noticed for Pt-modified photocatalyst, where the size of Pt is below 3 nm [33]. In the present study, Pt nanoparticles' average diameter was about 20 nm as a result of the deposition of Pt ions and their reduction on formed particles' defects. Therefore, the lower metal/semiconductor interface resulted in a decrease in photocatalytic activity under UV-Vis light irradiation.  For the final stability and reusability test, the most active defective photocatalyst was selected. For sample d-TiO2_20/Pt0.1/Cu0.1, three 1-h-long subsequent cycles of phenol degradation under UV-Vis light were performed. The obtained results are presented in Figure 9. There was no significant change in phenol degradation rate constant after the second and third cycles. Thus, the analyzed photocatalyst revealed good stability and reusability.
Furthermore, the reactive species were investigated to understand the photocatalytic reaction mechanism. Benzoquinone (BQ), silver nitrate (SN), ammonium oxalate (AO), and tert-butanol (t-BuOH) were used as superoxide radical anions (·O2 − ), electrons (e − ), holes (h + ), and hydroxyl radicals (·OH) scavengers, respectively. Obtained results, presented as phenol degradation rate constant k, in Fe3O4@SiO2/d-TiO2_20-Pt0.05 Fe3O4@SiO2/d-TiO2_20-Pt0.  For the final stability and reusability test, the most active defective photocatalyst was selected. For sample d-TiO 2 _20/Pt0.1/Cu0.1, three 1-h-long subsequent cycles of phenol degradation under UV-Vis light were performed. The obtained results are presented in Figure 9.  For the final stability and reusability test, the most active defective photocatalyst was selected. For sample d-TiO2_20/Pt0.1/Cu0.1, three 1-h-long subsequent cycles of phenol degradation under UV-Vis light were performed. The obtained results are presented in Figure 9. There was no significant change in phenol degradation rate constant after the second and third cycles. Thus, the analyzed photocatalyst revealed good stability and reusability.
The most significant impact on phenol degradation reaction in the presence of metal-modified d-TiO 2 was observed for superoxide radicals. After introducing to the photocatalyst suspension BQ solution, the phenol degradation efficiency was significantly inhibited. A slight decrease was also observed in the presence of SN as an electron trap. On the other hand, the addition of AO and t-BuOH did not decrease the phenol degradation rate.
Catalysts 2020, 10, x FOR PEER REVIEW 11 of 18 observed for superoxide radicals. After introducing to the photocatalyst suspension BQ solution, the phenol degradation efficiency was significantly inhibited. A slight decrease was also observed in the presence of SN as an electron trap. On the other hand, the addition of AO and t-BuOH did not decrease the phenol degradation rate. Modification of TiO2 resulted in the shift of the valence band as was revealed from the analysis of Mott-Schottky plot, where the relation between applied potential vs. Csc −2 is presented (see in Figure 11). According to the intersection with E axis the flat band potential was estimated. In the case of pure titania it equals to −1.2 V, whereas for d-TiO2_20-Pt0 the value of −1.13 was reached. In order to prepare energy diagram of both materials given in Figure 12, the values of bandgap energy was taken into account. As could be seen, for the modified material the position both the conduction and valence band are shifted. According to Monga et al. [46] the Schottky barrier formed at the metal-TiO2 interface affecting the efficiency of e-transfer. The lowering of the CB band edge is in accordance with the literature indicating that the work function of the metal prone decrease of the CB location. Then, the Schotky barrier is decreased at the metal/semiconductor heterojunction. As a result, the transfer of the photoexcited electron from metal NPs to titania is facilitated and plays important role in photocatalytic activity improvement. The introduction of titanium defects to the TiO2 crystal structure also resulted in narrowing the bandgap from 3.2 to 2.7 eV.
Based on the presented results, a schematic mechanism of UV-Vis phenol degradation in the presence of metal-modified defective Fe3O4@SiO2/d-TiO2-Pt/Cu photocatalyst was proposed and shown in Figure 12. After irradiation of the photocatalyst surface with UV-Vis light, electrons from the Pt are injected to the conduction band of titania and then utilized in oxygen reduction to form reactive oxygen radicals. The path of phenol degradation led through several intermediates, such as benzoquinone, hydroquinone, catechol, resorcinol, oxalic acid, and finally, to complete mineralization to CO2 and H2O [47][48][49]. An analysis of possible charge carriers' impact revealed that for photoactivity of d-TiO2-Pt/Cu, they are responsible for mainly generated superoxide radicals. The phenol degradation mechanism proceeded by the generation of reactive oxygen species, e.g., • O2 − , which attacked the phenol ring, resulting in benzoquinone and hydroquinone formation confirmed by high-performance liquid chromatography (HPLC) analyses. Moreover, during the photoreaction, Modification of TiO 2 resulted in the shift of the valence band as was revealed from the analysis of Mott-Schottky plot, where the relation between applied potential vs. Csc −2 is presented (see in Figure 11). According to the intersection with E axis the flat band potential was estimated. In the case of pure titania it equals to −1.2 V, whereas for d-TiO 2 _20-Pt0 the value of −1.13 was reached. In order to prepare energy diagram of both materials given in Figure 12, the values of bandgap energy was taken into account. As could be seen, for the modified material the position both the conduction and valence band are shifted. According to Monga et al. [46] the Schottky barrier formed at the metal-TiO 2 interface affecting the efficiency of e-transfer. The lowering of the CB band edge is in accordance with the literature indicating that the work function of the metal prone decrease of the CB location. Then, the Schotky barrier is decreased at the metal/semiconductor heterojunction. As a result, the transfer of the photoexcited electron from metal NPs to titania is facilitated and plays important role in photocatalytic activity improvement. The introduction of titanium defects to the TiO 2 crystal structure also resulted in narrowing the bandgap from 3.2 to 2.7 eV.
Based on the presented results, a schematic mechanism of UV-Vis phenol degradation in the presence of metal-modified defective Fe 3 O 4 @SiO 2 /d-TiO 2 -Pt/Cu photocatalyst was proposed and shown in Figure 12. After irradiation of the photocatalyst surface with UV-Vis light, electrons from the Pt are injected to the conduction band of titania and then utilized in oxygen reduction to form reactive oxygen radicals. The path of phenol degradation led through several intermediates, such as benzoquinone, hydroquinone, catechol, resorcinol, oxalic acid, and finally, to complete mineralization to CO 2 and H 2 O [47][48][49]. An analysis of possible charge carriers' impact revealed that for photoactivity of d-TiO 2 -Pt/Cu, they are responsible for mainly generated superoxide radicals. The phenol degradation mechanism proceeded by the generation of reactive oxygen species, e.g., • O 2 − , which attacked the phenol ring, resulting in benzoquinone and hydroquinone formation confirmed by high-performance liquid chromatography (HPLC) analyses. Moreover, during the photoreaction, the concentration of formed intermediates decreased, which suggests mineralization of recalcitrant chemicals to simple organic compounds.
Catalysts 2020, 10, x FOR PEER REVIEW 12 of 18 Figure 11. The Mott-Schottky plot for the bare and Pt modified d-TiO2.

Preparation of Defective TiO 2 -Pt/Cu Photocatalysts
Defective TiO 2 (marked as d-TiO 2 ) was obtained by the hydrothermal method assisted with the annealing process. Titanium(IV) butoxide (TBT) and iodic acid (HIO 3 ) were used as a TiO 2 precursor and oxidizing environment for titanium vacancies formation, respectively. Briefly, the appropriate amount of HIO 3 (presented in Table 5) was dissolved in 80 cm 3 of distilled water. After that, 10 cm 3 of TBT was added dropwise, and the obtained suspension was stirred for 1 h with magnetic stirring at room temperature. In the next step, the suspension was transferred into a Teflon-lined autoclave for thermal treatment at 110 • C for 24 h. The resultant precipitate was centrifuged, dried at 70 • C and then calcined at 300 • C for 3 h. The obtained defective d-TiO 2 photocatalysts were modified using platinum and copper nanoparticles by the co-precipitation method. In this regard, d-TiO 2 was dispersed in 50 cm 3 of deionized water, and Pt/Cu precursor solutions (0.05 and 0.1 mol% of Pt and 0.1 mol% of Cu with respect to TiO 2 ) were added. After that, NaBH 4 solution was introduced to reduce the metals ions followed by their deposition on the titania surface. The mole ratio of metal ions to NaBH 4 was 1:3. After the reduction process, the photocatalyst suspension was mixed for 2 h, and the d-TiO 2 -Pt/Cu nanoparticles were separated, washed with deionized water, and dried at 80 • C to dry mass. The final step was calcination at 300 • C for 3 h.

Preparation of Magnetic Fe 3 O 4 @SiO 2 /d-TiO 2 -Pt/Cu Nanocomposites
Previously obtained d-TiO 2 -Pt/Cu nanoparticles were deposited on a magnetic substrate as a thin photocatalytic active shell. Magnetite (Fe 3 O 4 ) was selected as a core of the designed composite due to its excellent magnetic properties (high Ms value and low Hc), which enable us to separate obtained photocatalyst in the external magnetic field. Silica was used as an interlayer to isolate Fe 3 O 4 from TiO 2 and suppress possible electron transfer between them. The magnetic photocatalysts were obtained in the w/o microemulsion system based on changes in the particles surface charge as a function of pH, described in the previous study [26].
Firstly, commercially available Fe 3 O 4 nanoparticles with nominate particles diameter of 50 nm were dispersed in water at pH 10. The prepared suspension was then introduced to cyclohexane/isopropanol (100:6 volume ratio) mixture in the presence of cationic surfactant and cetyltrimethylammonium bromide (CTAB) creating stable w/o microemulsion system with water nanodroplets dispersed in the continuous oil phase. The molar ratio of water to surfactant was set at 30. After the microemulsion stabilization, the corresponding amount of tetraethyl orthosilicate (TEOS) was added, resulting in the formation of SiO 2 interlayer on Fe 3 O 4 core, after ammonia solution introduced into the microemulsion system. The molar ratio of TEOS to Fe 3 O 4 was 8:1, and NH 4 OH to TEOS was 16:1. The microemulsion was destabilized using acetone and obtained nanocomposite Fe 3 O 4 @SiO 2 was separated, washed with ethanol and water, dried at 70 • C to dry mass, and calcined at 400 • C for 2 h.
In the second step, previously obtained Fe 3 O 4 @SiO 2 particles were combined with d-TiO 2 -Pt/Cu in order to create photocatalytic active nanomaterial. The reversed-phased microemulsion system at pH 10 was used, and Fe 3 O 4 to the TiO 2 molar ratio was equaled to 1:4 [27]. The junction between magnetic/SiO 2 and photocatalytic layers was promoted by their opposite surface charges, provided by the presence of CTAB at the basic conditions. The as-obtained Fe 3 O 4 @SiO 2 /d-TiO 2 -Pt/Cu samples, after their separation and purification using water and ethanol, were dried at 70 • C to dry mass and calcined at 300 • C for 2 h.

Characterization of the Obtained Magnetic Photocatalysts
The XRD analyses were performed using the Rigaku Intelligent X-ray diffraction system SmartLab equipped with a sealed tube X-ray generator (a copper target; operated at 40 kV and 30 mA). Data was collected in the 2θ range of 5-80 • with a scan speed and scan step of 1 • ·min −1 and 0.01 • , respectively. The analyses were based on the International Centre for Diffraction Data (ICDD) databased. The crystallite size of the photocatalysts in the vertical direction to the corresponding lattice plane was determined using Scherrer's equation with Scherrer's constant equal to 0.891. Quantitative analysis, including phase composition with standard deviation, was calculated using the Reference Intensity Ratio (RIR) method from the most intensive independent peak of each phase.
Nitrogen adsorption-desorption isotherms (BET method for the specific surface area) were recorded using the Micromeritics Gemini V (model 2365) (Norcross, GA, USA) instrument at 77 K (liquid nitrogen temperature).
Light absorption properties were measured using diffuse reflectance (DR) spectroscopy in the range of 200-800 nm. The bandgap energy of obtained samples was calculated from (F(R)·E) 0.5 against E graph, where E is photon energy, and F(R) is Kubelka-Munk function, proportional to the radiation's absorption. The measurements were carried out using ThermoScientific Evolution 220 Spectrophotometer (Waltham, MA, USA) equipped with a PIN-757 integrating sphere. As a reference, BaSO 4 was used.
The morphology and distribution size for Fe 3 O 4 @SiO 2 /d-TiO 2 -Pt as a reference magnetic nanocomposite sample was further analyzed using HR-TEM imagining (Tecnai F20 X-Twin, FEI Europe) together with elements identification in nanometric scale by EDS mapping.
Electron paramagnetic resonance (EPR) spectroscopy was used for intrinsic defects formation confirmation.

Photocatalytic Activity Analysis
Photocatalytic activity of the obtained samples was evaluated in phenol degradation reaction, both in UV-Vis and Vis light irradiation, using 300 W Xenon lamp (LOT Oriel, Darmstadt, Germany). For the visible light measurements, a cut-off 420 nm filter (Optel, Opole, Poland) was used to obtain a settled irradiation interval. A 0.05 g (2 g·dm −3 ) of a photocatalyst, together with a 20 mg·dm −3 phenol solution, was added to a 25 cm 3 quartz photoreactor with an exposure layer thickness of 3 cm and obtained suspension was stirred in darkness for 30 min to provide adsorption-desorption equilibrium. After that, photocatalyst suspension was irradiated under continuous stirring and a power flux (irradiation intensity) of 30 mW·cm −2 for 60 min. The constant temperature of the aqueous phase was kept at 20 • C using a water bath. Every 20 min of irradiation, 1.0 cm 3 of suspension was collected and filtered through syringe filters (pore size = 0.2 µm) for the removal of photocatalysts particles. Phenol concentration, as well as a formation of degradation intermediates, were analyzed using reversed-phase high-performance liquid chromatography (HPLC) system, equipped with C18 chromatography column with bound residual silane groups (Phenomenex, model 00F-4435-E0) and a UV-Vis detector with a DAD photodiodes array (model SPD-M20A, Shimadzu). The tests were carried out at 45 • C and under isocratic flow conditions of 0.3 mL·min −1 and volume composition of the mobile phase of 70% acetonitrile, 29.5% water and 0.5% orthophosphoric acid. Qualitative and quantitative analysis was performed based on previously made measurements of relevant substance standards and using the method of an external calibration curve.
Rate constant k was determined from ln(C o /C n ) against t plot where C o and C n are phenol concentrations [mg·dm −3 ] and t is degradation time [min]. Rate constant k is equal to directional coefficient "a" of the plot.
In order to evaluate the stability of the obtained photocatalysts, three 1-h-long subsequent cycles of phenol degradation under UV-Vis light using the most active defective d-TiO 2 _20/Pt0.1/Cu0.1 sample were performed. After each cycle, photocatalyst was separated from the suspension and use in the next cycle without additional treatment.

Electrochemical Measurements
In order to prepare Mott-Schottky plot the fabricated titania powders were used to form the paste, deposited using doctor-blade technique onto the Pt support. The paste consist of 0.2 g of photocatalyst in 0.1 g of polyethylene glycol (PEG) and 1 cm 3 of deionized water. Finally the calcination was carried out at 400 • C for 5 h with a heating rate 1 • C·min-1 ensuring removal of the organic binder. The fabricated electrode material stayed as working electrode tested in three electrode arrangement where Ag/AgCl/0.1M KCl and Pt mesh were used as reference and counter electrode, respectively. The deaerated 0.5 M Na 2 SO 4 was applied as electrolyte. The electrochemical spectroscopy (EIS) impedance data was recorded from the anodic towards cathodic direction. Prior the tests, the investigated samples were not subjected to any preliminary treatment or measurement and their potential was held to reach a steady-state conditions. EIS data were recorded for the single frequency of 1000 Hz in the potential range from +0.1 to −1.2 V vs. Ag/AgCl/0.1 M KCl using a 10 mV amplitude of the AC signal. The capacitance of space charge layer was further calculated from the imaginary part of the measured impedance following the equation [50]: where f stands for the frequency of the AC signal and Z im for the imaginary part of impedance.

Conclusions
Surface modification of defective d-TiO 2 photocatalyst with platinum and copper nanoparticles resulted in a significant increase in its photocatalytic activity, both in UV-Vis and Vis range. The EPR analysis confirmed the presence of Ti defects in the structure of TiO 2 samples. The highest activity in Vis light was noticed for d-TiO 2 modified with Pt NPs. It resulted from surface plasmon resonance properties of Pt and narrowing the bandgap of the defective d-TiO 2 . Among magnetic photocatalysts, the highest activity in Vis light was observed for Pt-modified and Pt/Cu-modified defective d-TiO 2 deposited on Fe 3 O 4 @SiO 2 magnetic core. Analysis of phenol degradation mechanism revealed that superoxide radicals are mainly responsible for phenol oxidation and mineralization. However, the photocatalytic activity in reaction of phenol degradation in UV-Vis light in the presence of Pt-modified Fe 3 O 4 @SiO 2 /d-TiO 2 with the Pt particle size of about 20 nm was comparable with the activity of Fe 3 O 4 @SiO 2 /d-TiO 2 . It resulted from the deposition of Pt NPs in the place of titanium vacancies, and as a consequence formation of larger metal particles due to the seed-mediated growth mechanism on the TiO 2 . In this regard, a lower metal/semiconductor interface resulted in a decrease in photocatalytic activity in the UV-Vis spectrum range. Furthermore, the creation of a core-shell magnetic Fe 3 O 4 @SiO 2 /d-TiO 2 -Pt/Cu nanostructures allowed an effective separation of the obtained magnetic photocatalysts.