Insight into Potassium Vanadates as Visible-Light-Driven Photocatalysts: Synthesis of V(IV)-Rich Nano/Microstructures for the Photodegradation of Methylene Blue

Photocatalysis is regarded as a promising tool for wastewater remediation. In recent years, many studies have focused on investigating novel photocatalysts driven by visible light. In this study, K2V6O16·nH2O nanobelts and KV3O8 microplatelets were synthesized and investigated as photocatalysts. Samples were obtained via the facile method based on liquid-phase exfoliation with ion exchange. By changing the synthesis temperature (20–80 °C), different compositions, morphologies, and V4+/V5+ ratios were obtained and investigated as photocatalysts for organic dye degradation. Potassium vanadates’ structural, morphological, and optical properties were characterized using X-ray diffraction(XRD), Fourier transform infrared spectroscopy (FTIR), X-ray photoelectron spectroscopy (XPS), Physical Property Measurement System (PPMS), thermogravimetric analysis (TGA) with mass spectrometry (MS), N2 adsorption, scanning electron microscopy (SEM), photoluminescence (PL), and UV–vis diffuse reflectance spectroscopy (DRS). Synthesized K2V6O16·nH2O and KV3O8 showed an efficient absorption in the visible wavelength region with a narrow band gap energy of 1.80 and 1.91 eV, respectively. Their photocatalytic activity was evaluated by the degradation of methylene blue (MB) under simulated solar light illumination. The KV3O8 microplatelets exhibited the greatest photocatalytic activity, resulting in more than 90% degradation of the dye within the first 30 min. It is suggested that the observed excellent photocatalytic performance is attributed to the high content of V4+ species. Furthermore, the influence of active species was investigated, and the mechanism responsible for the photodegradation of the MB dye was discussed for the first time for potassium vanadates.


■ INTRODUCTION
The search for new, efficient, and low-cost photocatalysts has attracted increasing attention due to their perfect utilization of clean and renewable solar energy for treating wastewaters. In the past years, various materials have been investigated and tested for water purification. 1−4 Examples include metal 5,6 and metal oxide nanoparticles, 7 sulfides, 8,9 nitrides, 10 metal− organic frameworks, 11,12 or carbon-based nanostructures. 13 Recent studies demonstrate that metal vanadates exhibit promising visible-light photocatalytic reactivity in decomposing pollutants and water splitting. 14−17 A series of compounds can be mentioned here, such as Ag 3 VO 4 , 18 24 and BiVO 4 25,26 as the most known representatives from the vanadate family. The conducted research is mainly devoted to the nanostructured vanadates, which can appear in various forms such as nanoparticles, 27,28 nanobelts, 29,30 or nanotubes. 31 It is well known that the shape and size play a significant role, as broadly described in the literature. 32 −35 The photocatalytic properties are also strongly dependent on the crystalline structure. Generally, the monoclinic scheelite BiVO 4 (m-BiVO 4 ) is considered to be more active than the other two tetragonal phases. 36−38 In addition, the face-dependent photocatalytic behavior was also shown with the beneficial effect of (040) facets on contaminant degradation. 39−42 Another critical factor in the photocatalytic reaction is crystal defects, for example, oxygen vacancies, which trap photogenerated pairs and reduce their recombination rate. 43 The presence of oxygen vacancies can also broaden the activity range of a semiconductor oxide from UV to NIR light. 44 As a result, new pathways to synthesize photocatalysts with controllable properties are sought.
In this study, we present a new synthesis method of potassium vanadates as efficient photocatalysts for methylene blue (MB) degradation. Potassium vanadates were prepared by the facile LPE-IonEx method (liquid-phase exfoliation with ion exchange), which was recently proposed by our group. 63 LPE-IonEx is a straightforward, low-temperature, one-pot approach for the synthesis for transition metal oxide bronzes with controlled structural and morphological properties. Notably, the proposed method uses water as a solvent, making it ecofriendly. We demonstrated that, depending on the synthesis temperature (20−80°C), the hydrated single-phase K 2 V 6 O 16 · nH 2 O to nonhydrated KV 3 O 8 can be obtained through their mixture. In addition, an increase in temperature leads to increased V 4+ concentration. The photocatalytic activity of prepared materials was evaluated by the degradation of an organic dyemethylene bluein water under simulated solar light illumination. So far, KV 3 O 8 has been considered as a precursor for V 3 O 7 nanobelt synthesis for photocatalytic water splitting applications. 45 To the best of our knowledge, the photocatalytic properties of nonhydrated phase KV 3 O 8 have never been investigated. Furthermore, the mechanism and kinetics of photocatalytic degradation were studied for the first time.
■ EXPERIMENTAL SECTION Synthesis. Potassium formate (99%, Sigma Aldrich) and V 2 O 5 (99.2%, Alfa Aesar) without further purification were used as reagents. Milli-Q deionized water was used (resistivity >19 MΩ·cm). The samples were prepared via the LPE-IonEx method, where the procedure was as follows: 500 mg of V 2 O 5 was added to 50 mL of a 1 M solution of potassium formate in deionized water. The mixture was vigorously stirred for 72 h, and the synthesis was conducted in four temperatures: 20, 40, 60, and 80°C. The samples were labeled according to the reaction temperature as KVO-20, KVO-40, KVO-60, and KVO-80, respectively. From the obtained solution, rusty red to burnt orange to orange precipitations were collected by centrifugation. After washing several times with deionized water, products were dried overnight at 40°C under reduced pressure (0.01 bar).
Physicochemical Characterization. The X-ray diffractograms (XRD) were collected on a BrukerD2 Phaser diffractometer with Cu Kα radiation (λ = 1.5404 Å) in the range 2θ of 5−70°at room temperature. The samples' morphology was studied at room temperature by an FEI Quanta FEG 250 scanning electron microscope (SEM) in secondary electron mode using an Everhart-Thornley detector (ETD). Fourier transform infrared (FTIR) spectra were measured on a Perkin Elmer Frontier spectrophotometer. The KBr pellet method was used, and transmittance spectra were recorded from 4000 to 500 cm −1 with a resolution of 4 cm −1 . The highresolution X-ray photoelectron spectroscopy (XPS) analysis was performed using an Escalab 250Xi device (ThermoFisher Scientific, USA) equipped with a monochromatic Al Kα source. Measurements were carried out at 25 eV pass energy with 0.05 eV energy step size. The X-ray spot size was 250 μm. The calibration of the XPS spectrum was done using the characteristic peak of adventitious carbon C 1s at 284.6 eV. 46 Thermogravimetric analysis (TGA) was carried out under an argon atmosphere with a flow rate 60 mL min −1 in the temperature range of 40−400°C (with a heating rate of 5°C min −1 ) using Netzsch STA449 F1. A constant sample mass (20 ± 0.5 mg) was used. The thermal behavior has also been studied by EGA−MS (evolved gas analysis−MS). The gases that come out from the sample during heating were monitored by the quadruple mass spectrometer Netzsch QMS 403 Aeölos. The differential scanning calorimetry (DSC) measurement was performed under an argon atmosphere with a flow rate 60 mL min −1 in the temperature range of 35−450°C (with a heating rate of 15°C min −1 ) using a NETZSCH DSC 204 F1 Phoenix calorimeter.
Nitrogen adsorption−desorption isotherms were measured on a surface area analyzer (NOVAtouch 2, Quantachrome Instruments) at 77 K. Before the measurements, samples were degassed under a dynamic vacuum at 40°C for 12 h. The specific surface area was calculated using the Brunauer−Emmett−Teller (BET) linear equation in the relative pressure range (p/p 0 ) from 0.1 to 0.3. The correlation coefficient of the linear regression was not less than 0.99.
The UV−vis reflectance spectra of the selected materials were measured with a UV−vis spectrophotometer (Lambda 35, Perkin-Elmer) equipped with a diffuse reflectance accessory. The spectra were registered in a range of 300−900 nm with a scanning speed of 120 nm min −1 . Band gap energy values were determined as the intercept of the tangent of the plot of transformation of the Kubelka− Munk function. To determine the energy band gap (E bg ) of the chosen powders, the Kubelka−Munk function (eq 1) was applied: where R is the reflectance. The band gap was estimated by extrapolation of the linear region of (f(KM) hν) n vs hν to y = 0, where the power ″n″ depends on the electron transition (n = 1/3, indirect forbidden (i.f.); n = 0.5, indirect allowed (i.a.); n = 2/3, direct forbidden (d.f.); and n = 2, direct allowed (d.a.)). 47 The photoluminescence spectra were recorded using 0.3 m Czerny−Turner spectrograph (SR303i, Andor) equipped with an ICCD camera (DH740, Andor). Powders were excited with UV LED (365 nm, 9 nm FWHM, 350 mW).
Magnetic Properties. Quantum Design Physical Property Measurement System (PPMS) with a vibrating sample magnetometer function was used to measure the temperature-dependent magnetic susceptibility (defined as χ = M/H, where M is the magnetization and H is the applied field). The temperature dependence of the zero-field cooled (ZFC) magnetization was measured in the applied field of 3 T. Magnetic measurements were performed on a sample of arbitrary shape with a mass of about 10 mg.
Photocatalytic Activity. To evaluate the powders' photocatalytic activity, the degradation rate of methylene blue (MB) dye was monitored. Twenty milligrams of catalysts and 50 mL of an aqueous solution of MB (C 0 = 10 −5 M) were transferred to the beaker. Before irradiation, the suspension was magnetically stirred in the dark for 30 min to achieve adsorption/desorption equilibrium. Then, the suspension was irradiated with sunlight irradiation (a high-pressure 150 W xenon lamp, LOT−Quantum Design GmbH equipped with an AM1.5G filter) with constant magnetic stirring (150 rpm). The intensity of the light that reached the solution's surface was equal to 100 mW cm −2 . The changes of the concentration of MB (C) during decomposition were monitored using a UV−vis spectrophotometer at a wavelength of 665 nm. The quantitative evaluation of the role of the chemical individuals responsible for dye decomposition was performed using appropriate scavengers (1 mM of tert-butyl alcohol, benzoquinone, and ammonium oxalate).
The MB products after photocatalytic degradation using KVO-20 and KVO-80 samples were analyzed by ultra-performance liquid chromatography quadrupole time-of-flight mass spectrometry (UPLC−QTOF-MS/MS). The analysis of solutions obtained after photocatalytic degradation was performed on an Agilent 1290 Infinity liquid chromatograph and Agilent 6550 iFunnel Q-TOF LC/MS System. The mobile phase consisted of phase A: 0. 1% aqueous formic acid with 0.3% acetonitrile and phase B: 0.1% formic acid in acetonitrile. The analysis was recorded with a gradient elution from 0 to 100% B during 10 min with an injection volume of 5 μL. The Agilent ZORBAX RRHD Eclipse Plus C18 column (95 Å, 2.1 × 50 mm, 1.8 μm) was used for the separation and detection of resulting products with a column temperature of 24°C and flow rate 0.3 mL/ min. The mass spectra were obtained in positive modes using the following operating parameters: capillary voltage 3.5 kV, nozzle voltage 2 kV, and fragmentor voltage 175 V. The gas temperature was 250°C with a flow rate of 12 L/min and nebulizer pressure of 40 psi; the sheath gas temperature was 300°C with a flow rate 11 L/min. Inorganic Chemistry pubs.acs.org/IC Article The mass spectra were monitored in the range from 50 to 1000 m/z with a scan range of one spectrum per second. The obtained UPLC− QTOF-MS/MS data were analyzed using the Mass Hunter Qualitative Analysis software. Electrochemical Characterization. The electrochemical measurements were performed using an Ivium Vertex potentiostat/ galvanostat in a three-electrode cell using Pt mesh as the counter electrode and Ag/AgCl (3 M KCl) as the reference electrode. The tested powder was deposited onto degreased FTO (fluorine-doped tin oxide) using a dip-coating method according to the procedure described previously. 48 Measurements were performed in deaerated aqueous 0.2 M K 2 SO 4 . The Mott−Schottky analysis was performed to determine the flat band potential. The impedance spectra were recorded at a potential range from 0.4 to 0.9 V vs Ag/AgCl (3 M KCl). The potential was held before each spectrum registration to achieve a steady-state condition. The potential range for analysis was determined on the basis of a cyclic voltammetry (CV) curve. The space charge capacitance was determined from 1000 Hz frequency according to the formula C sc = 1/ωZ″.

■ RESULTS AND DISCUSSION
Structural Analysis. The X-ray diffractograms were recorded to confirm the phase purity and crystallinity of obtained samples (see Figure1a). For the diffractogram for the sample synthesized at 20°C (KVO-20), all indices can be indexed within hydrated potassium vanadate phase K 2 V 6 O 16 · 1.5H 2 O, JCPDS card no. 00-051-0379, which corresponds to a monoclinic structure with lattice parameter values of a = 12.29 Å, b = 3.59 Å, and c = 16.01 Å. The main diffraction peak, located at approximately 11°, corresponds to the diffraction from the (002) crystallographic plane of this phase. Moreover, no signals of other phases were detected, indicating the high purity of the sample. For the samples synthesized at higher temperatures, additional peaks can be observed, indicating the presence of the secondary phase KV 3 O 8 , JCPDS card no. 00-086-24-95, which corresponds to a monoclinic structure with lattice parameter values of a = 4.97 Å, b = 8.38 Å, and c = 7.64 Å. The presence of KV 3 O 8 is revealed by its characteristic diffraction peak located at approximately 11.5°, which corresponds to the (001) crystallographic plane (see inset in Figure 1a). With increasing synthesis temperature, the intensity of the diffraction peaks ascribed to the KV 3 O 8 phase increases, and the first peak of KVO-40 and KVO-60 is split. This suggests the existence of two phases in these samples. The XRD pattern of sample KVO-80 shows indexed reflection only for the KV 3 O 8 phase. The characteristic peak for the hydrated potassium vanadate phase is no more visible. FTIR spectra of the samples are shown in Figure 1b and summarized in Table  S1. The bands at ∼1005 and ∼970 cm −1 can be assigned to VO vibrations, while those at 525 and 590 cm −1 can be assigned to the symmetric and asymmetric stretching of V− O−V. 49 The band at 730 cm −1 corresponds to bridging V−O··· K stretching. 50 In all spectra, the splitting of the VO bands can be observed. Such phenomenon implies the distortions in the vanadium oxide layers and the existence of VO 5 and VO 6 polyhedra, which are typical for hexavanadates and mixedvalence vanadium compounds. 51,52 Significant changes were   56 Thermogravimetric measurements were performed to further investigate the weight percentage of crystalline water in samples (Figure 1c). According to the TG and DTG curves, the decomposition of the material occurred stepwise. Especially for single-phase, KVO-20 is seen on the DTG curve that a broad peak between 40 and 180 is asymmetric (which indicates two different decomposition kinetics) and overlaps with another broad one. Simultaneously, the molecular weight channels 17 (corre-sponding to OH) and 18 (H 2 O) on the EGA−MS curves showed a broad peak ion current corresponding to the shape of the peak on the DTG curve (Supporting Information Figure  S1a). The ratio of the integrated ionic current of m/z = 17 and 18 is around 0.2 in the whole temperature range, which corresponds to a well-known signature of water molecules. 57 Thus, the weight loss between 40 and 100°C is attributed to the evaporation of physisorbed water and that between 100 and 400°C to crystalline water. 58 The weight loss of crystalline water for a single-phase hydrated sample (KVO-20) is 1.75 wt %, which corresponds to 0.65 molecule of water per K 2 V 6 O 16 formula unit. The weight loss for KVO-40 and KVO-60 samples is smaller and equal to 1.48 and 0.72%, respectively. According to the XRD and FTIR results, these samples are composed of two phases: the hydrated and nonhydrated one. Therefore, upon comparison, the weight loss of the crystalline water in these samples will be smaller. The KVO-80 sample is thermally stable up to 350°C, which is confirmed by DSC results where no thermal effects are visible (Supporting Information Figure S1b). The structural and thermal measurements confirmed that this sample consists of only a nonhydrated KV 3 O 8 phase. X-ray photoelectron spectroscopy (XPS) was used to assess the chemical composition and the charge state of vanadium ions in the studied samples. The high-resolution XPS spectra recorded in O 1s and V 2p binding energy regions (Figure 2a− d) show a complex shape of not less than three peaks located around 530 eV attributed to O 1s, as well as 517 and 524 eV attributed to the V 2p 3/2 and V 2p 1/2 peak doublet. 59 The V 2p peaks for all samples show an asymmetric shape with a very weak shoulder line shifted toward negative binding energies  and O−V 5+ bonds, respectively, whose ratios are in good agreement with the V 5+ /V 4+ ratios estimated from the deconvolution of V 2p 3/2 . The third peak centered at 532.5 eV, observable for all samples, is ascribed to adsorbed OH − ions on the surface. 61 The above XPS analysis confirms the mixed-valence character of the obtained samples and the existence of V(IV) and V(V) in their structure. However, it should be borne in mind that XPS is a surface-sensitive technique and that the obtained results indicate a high content of V 4+ on the surface (which is important in the view of catalytic action) and not in the bulk. Thus, to further investigate the valence of vanadium in the materials, static magnetic susceptibility was measured for samples obtained at the lowest (KVO-20) and highest temperature (KVO-80).
The temperature dependence of the magnetic susceptibility for KVO-20 and KVO-80 measured in an applied field of 3 T is shown in Figure 3. It can be clearly seen that for both samples above about 50 K, χ(T) is weakly temperature-dependent and its magnitude is small. At low temperatures, the pronounced tail is observed, which can be attributed to the presence of a small amount of uncompensated V 4+ ions. 62 To determine the effective magnetic moment (p eff ), the experimental data were fitted by the modified Curie−Weiss law, χ = χ 0 + C/(T − θ P ), where χ 0 , C, and θ P are the temperature-independent susceptibility, the Curie constant, and the paramagnetic Curie temperature, respectively. The fit gave C = 0.00712(3) emuK for KVO-20 and C = 0.00178(2) emuK for  Assuming that the magnetic moment originates from V 4+ only, the effective magnetic moment per V can be obtained using the relation p eff = (3Ck B /μ B 2 N A ) 1/2 , where k B is the Boltzmann constant, μ B is the Bohr magneton, and N A is Avogadro's number. The resulting effective magnetic moment p eff is 0.24 and 0.12 μ B /V for KVO-20 and KVO-80, respectively. The calculated effective magnetic moments are much smaller than the expected value of p eff = 1.73 μ B for the free V 4+ ion. Thus, obtained results suggest that the samples consist of a vast majority of nonmagnetic V(V) species and only a small fraction of V(IV) located on the sample surface or between crystalline grains, and in case of bulk samples, the V 4+ /V 5+ ratio is smaller than that on the surface.
The morphologies of the samples undergo significant temperature evolution. The sample KVO-20 (single-phase K 2 V 6 O 16 ·0.65H 2 O according to XRD, FTIR, and TG results) consists of belt-like structures with a width between 50 and 200 nm, a thickness of 10−40 nm, and a length of a few micrometers (Figure 4a,b). With the increase of the synthesis temperature, the K 2 V 6 O 16 ·0.   Inorganic Chemistry pubs.acs.org/IC Article surface areas of samples were determined through N 2 adsorption and BET analysis ( Figure 5). As expected, the highest surface area (30.9 m 2 /g) has been noted for KVO-20 with a nanostructural morphology and the lowest (4.6 m 2 /g) for KVO-80 with microplatelets. The schematic illustration of the evolution of potassium vanadate morphology is presented in Figure 6. The synthesis reaction is based on the intercalation of the solvent molecule into layer spacing and its further exchange with alkali metal ions. The reaction scheme of KVO by the LPE-IonEx synthesis method is similar to the ammonium metavanadate synthesis reported previously. 63 The solid vanadium oxide underwent delamination by the insertion of water molecules into the interlayer space. The H 2 O molecules are partially trapped during potassium ion intercalation, and hydrated potassium vanadate (K 2 V 6 O 16 ·mH 2 O) is formed. During the synthesis at room temperature, the water stoichiometry is stabilized and elongated nanobelts (K 2 V 6 O 16 ·nH 2 O) are obtained. The KV 3 O 8 microcrystals are formed via the dehydration and recrystallization of K 2 V 6 O 16 ·mH 2 O at higher synthesis temperatures because the exchange between solvent molecule and potassium ion is more efficient, and a nonhydrated phase is obtained: KV 3 O 8 . Throughout the dehydration process, the c axis of the crystal unit cell is decreased by almost half. During this process, the creation of V 4+ species is probable. Therefore, the obtained KV 3 O 8 phase possesses a rich vanadium-defective structure.
Photocatalytic Properties. The photocatalytic performance of investigated powders is presented as a C to C 0 ratio in Figure 7a. Additionally, the effect of the photolysis of MB is added for comparison (blank). Before the irradiation, methylene blue was adsorbed on the surface of the tested powders at a similar, considerable level of 30%. Adsorption is an essential process in photocatalysis, which enriches the pollutants around the catalyst surface, and often, strong adsorption results in increased photocatalytic performance. 64 All samples acted as efficient photocatalysts for dye decomposition; however, the sample obtained at 80°C exhibited the highest photoactivity. After about 30 min of exposure, over 90% of MB was degraded. In the case of the rest of the samples (obtained at 20, 40, and 60°C), the observed photoactivity was similar and the time required for MB degradation was longer (approximately 90% of degradation after 90 min). To the best of our knowledge, there is only one     72 Yu and co-workers presented CaV 2 O 6 nanorods that exhibited improved photocatalytic activity in MB degradation due to the coexistence of V 4+ /V 5+ in the lattice. CaV 2 O 6 shows an indirect allowed electronic transition with a band gap energy of 2.56 eV and hydroxyl radicals as the major active species. 73 Saputera et al. obtained BiVO 4 with a different vanadium vacancy content by changing the calcination temperature after sol−gel synthesis. Obtained samples were then evaluated by the degradation of palm oil mill effluent waste under UV−visible light. Results revealed that the higher the V 4+ content was, the better were the degradation rates and reaction rate constant. It was also observed that the increase in V 4+ content resulted in a slight increase of band gap energy (2.44 to 2.50 eV). 74 To investigate the kinetics of the process, the ln(C/C 0 ) vs t plot was prepared (see Supporting Information Figure S2). The most probable is the pseudo-first-order kinetics according to the Langmuir−Hinshelwood model; 75 however, the mechanism of the photocatalytic degradation can be more complex. Thus, the additional measurements for single-phase samples K 2 V 6 O 16 ·0.65H 2 O (KVO-20) and KV  in the presence of appropriate scavengers were performed (see Figure 7b,c). tert-Butyl alcohol (TBA), benzoquinone (BQ), and ammonium oxalate (AO) act as a hydroxyl radical, superoxide radical, and hole scavenger, respectively. 76 The decrease of the decomposition efficiency in a solution containing a scavenger suggests the role of the relevant species in the mechanism of the degradation mechanism. Thus, in both cases, the role of superoxide radicals is negligible. On the other hand, degradation efficiency in the presence of a hydroxyl radical scavenger is significantly diminished in comparison with the control measurement. It is indirect nanobelts, which were, respectively, the most and the least photoactive material, was further examined using UV−vis spectroscopy in the reflectance mode. The spectra of both samples are presented in Figure 8a. As can be seen, synthesized materials absorb a significant part of the light in the visible range. This feature is beneficial for the materials used in photocatalysis. Reflectance edges seen on both spectra are related to the energy band gap transition, so the obtained results allowed the energy band gap to be estimated. There is no information about the type of electron transition; thus, all possibilities were taken into account, as presented in Figure S3. The absorbance spectra of both powders with marked band gaps estimated from (f(KM)·hν) n vs hν plots are presented in Figure 8b. Since the energy band gap corresponds to the absorbed photons with the lowest energy, the direct energy band gap (d.a.) in the case of K 2 V 6 O 16 ·0.65H 2 O nanobelts is the most probable. On the other hand, the estimated value of the direct band gap for KV 3 O 8 microplatelets corresponds to a wavelength that is already within the absorption edge. Thus, it is very likely that the optical band gap of the KV 3 O 8 material is related to the allowed indirect transition (i.d.); nevertheless, the direct band gap can be taken into account as well. Thus, the determined direct energy band gaps for K 2 V 6 O 16 ·0.  Figure 5). It can be concluded that all the abovementioned parameters have a greater impact on photoactivity than UV−vis absorption ability.
Additionally, the photoluminescence spectra have been recorded using UV excitation. Results are presented in Figure  8c. Generally, the emission spectrum of KVO-80 is characterized by a much higher intensity of photoluminescence in comparison with KVO-20. It could be simply related to the better crystallinity of the sample prepared at a higher temperature; 77 however, the shape of the emission bands is also different. Often, photocatalysts are examined with PL to compare emission intensity. In the case of photoactive materials, a higher PL intensity means a higher rate of recombination processes and therefore poorer photocatalytic performance. Thus, here PL spectra suggest that better photocatalytic activity should be observed for the KVO-20 sample (which is not true). However, such direct comparison can be done for samples with similar characteristics (composition, morphology, surface area, etc.) because PL intensity is affected by many parameters. 78 The PL intensity of the KVO-20 spectrum is very low, which makes the analysis very difficult to perform. The maximum can be found at about 527 nm (2.35 eV). Taking into account the absorbance spectrum, the recorded emission is not related with conduction to valence band extinction. In the case of the KVO-80 sample, the spectrum can be deconvoluted to 3 Gaussian peaks with the maximum at 467 nm (2.66 eV), 561 nm (2.21 eV), and 612 nm (2.03 eV). The PL emission at energies lower than the band gap (2.03 eV) is related to the presence of states within the energy band gap. Here, it can be associated with the presence of oxygen vacancies and V(IV) centers seen on XPS. A similar phenomenon was observed for partially reduced metal oxides (TiO 2 , SnO 2 , and In 2 O 3 ), while for fully oxidized materials, visible light photoluminescence was diminished. 79 The PL bands at energies lower than the energy band gap were also reported for V 2 O 5 and were described as an effect of the oxygen vacancies' presence. 80 The second band at the PL spectrum of KVO-80 has an energy (2.21 eV, 561 nm) close to Inorganic Chemistry pubs.acs.org/IC Article the band gap; thus, it probably originates from the direct band edge transition. The low-intensity band at 467 nm that can be seen in Figure 8c and a similar one were already reported for V 2 O 5 nanostructures. It was described as the recombination of UV excited electrons from a higher level in the V 3d orbital (in the conduction band) to the ground level. 81 Energy Band Position of KV 3 O 8 (KVO-80). Electrochemical characterization of the sample allowed the estimation of the energy band position for the stable nonhydrated KV 3 O 8 phase. The procedure of sample thin film preparation did not affect its crystal structure and the charge state of vanadium (see Supporting Information Figures S4 and S5). The flat band potential (E fb ) was determined using the Mott−Schottky (MS) analysis. To estimate the space charge region capacitance (C sc ), the impedance spectra were measured at dark conditions as a function of applied potential. The range of potential for MS analysis was chosen on the basis of the cyclic voltammetry (CV) study presented in the Figure 9a inset. The peaks observed at the CV curve are probably related to the V center oxidation/reduction with the simultaneous K + intercalation/ deintercalation process. The impedance spectra were measured between 0.4 and 0.9 V vs Ag/AgCl (3 M KCl), but the linear behavior of 1/C sc 2 vs E was found between 0.58 and 0.69 V (see Figure 8a). The positive slope of the Mott−Schottky plot confirmed that the obtained nonhydrated KV 3 O 8 phase is an ntype semiconductor and E fb equals to about 0.56 V vs Ag/AgCl (3 M KCl) (0.765 V vs NHE). The energy band diagram was prepared assuming that E fb lies just below the conduction band and the UV−vis determined indirect energy band gap of KV 3 O 8 is 1.91 eV. As presented in Figure 8b, the valence band (VB) location allows photoexcited holes to react with adsorbed water and form hydroxyl radicals. On the other hand, the potential of the photoexcited electrons from the conduction band (CB) is not enough to form superoxide radicals. These conclusions are consistent with the efficiency of the photocatalysts.
According to the above-described results, the possible mechanism of photocatalytic performance of KV 3 O 8 has been proposed. The major reaction steps responsible for the photocatalytic activity could be expressed as follows: (  Figure S6. On the basis of the received data after using KVO-80 as a catalyst, besides the nondecomposed MB, two additional degradation compounds were identified. The observed peaks on the chromatogram at 2.576 and 8.122 min correspond probably to 4-nitroaniline [C 6 Figure S7, Supporting Information), while the peak observed at 4.591 min is attributed to MB. The above photodegradation process generates about 8.34% of nondecomposed MB, 0.12% of aniline, and 0.45% of phenol. The possible degradation pathway where the aniline occurred as a product was described elsewhere. 82 In the case of analysis of MB degradation products using KVO-20 as the photocatalyst, the obtained results allowed for the identification of at least four additional compounds besides MB residues (4.586 min, see Figure S8). The observed peaks on the chromatogram at 2.604 and 8.116 min in the resulting products, as in the case of KVO-80 catalyst, can be assigned to 4-nitroaniline [C 6 H 6 N 2 O 2 + H] + and m/z 139.0508, and phenol [C 6 H 6 O + H] + and m/z 95.0494, respectively. Analysis of the sample after the use of KVO-20 as a catalyst also identifies two additional compounds. On the chromatogram at 3.12 min, the [C 7 H 10 N 2 O 4 S + H] + and m/z 219.0444 was identified, corresponding to 2-amino-5-(methylamino)-hydroxybenzenesulfonic acid, which was previously described in the photocatalytic MB degradation pathway, 83 which leads to H 2 O and CO 2 . In addition, it is worth mentioning that on the chromatogram peak presence at 5.219 min reveal to benzenesulfonic acid [C 6 H 6 O 3 S + H] + and m/z 159.0063, which also occur in the photocatalytic MB degradation pathway as previously described. 84 Comparing the efficiency of MB degradation through the use of the KVO-20 catalyst with the KVO-80 catalyst, the efficiency of the KVO-20 catalyst is significantly lower. The photodegradation process catalyzed by KVO-20 generates about 16.78% of nondecomposed MB, 36.27% of benzenesulfonic acid, and comparable amounts (0.40%) of phenol and the amino-5-(methylamino)-hydroxybenzenesulfonic acid, which is formed in approximately 0.46%. On the other hand, a negligible 0.07% amount of 4-nitroaniline is formed when the KVO-20 catalyst was applied, which indicates the efficiency and difference of the photodegradation mechanisms of MB using the described catalysts.

■ CONCLUSIONS
In this study, K 2 V 6 O 16 ·0.65H 2 O nanobelts and KV 3 O 8 microplatelets were synthesized using the facile LPE-IonEx method and investigated as catalysts for organic dye degradation. The obtained samples were characterized using XRD, FTIR, XPS,PPMS, UV−vis DRS, PL, TGA with MS, SEM, and N 2 adsorption isotherms. Controlling the reaction temperature (20−80°C) resulted in various phase compositions, morphologies, and surface areas. Moreover, different synthesis temperatures provided varied concentrations of V 4+ (from 9% up to such a high content as 62%). The photocatalytic degradation of methylene blue, under simulated solar light illumination, was examined to evaluate the photocatalytic performance of prepared samples. All samples acted as efficient photocatalysts, resulting in approximately 90% degradation of the dye within the first 30−90 min. The highest activity was observed for the sample obtained at 80°C. According to the scavenger test, OH radicals and holes are the main active species in the case of K 2 V 6 O 16 ·0.65H 2 O. The role of direct oxidation in the case of KV 3 O 8 is smaller but still significant. The results indicate that potassium vanadates are potential candidates for light-driven photocatalysts. The observed excellent photocatalytic performances result from the specific structural properties, which were provided by the proposed innovative LPE-IonEx method. We suggest that it can be attributed to the high content of V 4+ species in the samples, which traps electrons and facilitates charge separation.