MoS2/Cu/TiO2 nanoparticles: synthesis, characterization and effect on photocatalytic decomposition of methylene blue in water under visible light

ABSTRACT
Photodegradation processes are of great interest in a range of applications, one of which is the photodecomposition of pollutants. For this reason, analysing nanoparticles that improve the efficiency of these processes under solar radiation are very necessary. Thus, in this study, TiO2 was doped with Mo and Cu using low-temperature hydrolysis as the method of synthesis. Pure TiO2 and x%MoS2/Cu/TiO2 nanoparticles were prepared, where x is the theoretical quantity of MoS2 added (0.0%, 1.0%, 5.5%, 10.0%), setting the nominal quantity of Cu at 0.5 wt.%. The samples obtained were characterised by X-Ray Diffraction, Raman spectroscopy, x-ray electron spectroscopy (XPS) and UV-Vis spectroscopy in diffuse reflectance mode (DR-UV-Vis). The results suggest that the TiO2 structure was doped with the Mo6þ and Cu2þ ions in the position of the Ti4þ. The x%MoS2/Cu/TiO2 samples presented lower band gap energy values and greater optical absorption in the visible region than the pure TiO2 sample. Lastly, the photocatalytic activity of the samples was assessed by means of the photodegradation of methylene blue under visible light. The results show that when the quantity of Mo in the co-doped samples increased (x%MoS2/Cu/TiO2) there were significant increases of up to 93% in the photocatalytic activity.

INTRODUCTION
A great deal of research has been performed in the last few decades into TiO2 thanks to its magnificent electronic andphotocatalytic properties, low toxicity, low cost and high chemical stability (Szczepanik ). Photocatalytic effi- ciency depends on a range of factors including crystallinephases, morphology, specific surface, chemical surface, defects, band gap energy, type of dopant and the level of doping (Vattikuti et al. ; Fu et al. ; Munir et al.). However, the studies performed have raised contro-versy regarding the influence of these properties on photocatalytic activity.The electron-gap pairs generated in the bulk material as a consequence of the light absorbed must reach the surface of the photocatalyst to react and begin the photodegrada- tion. The valence band establishes the electron-gap pairs available and the conductivity of the gaps, while theconduction band determines the mobility of the electrons. Consequently, the density of these states and their energy will determine the efficiency so that charged particlesreach active centres on the surface of the material (Wojtaszek et al. ). Therefore, for a photocatalyst to be really efficient it is essential to minimize the recombinationof photogenerated charge carriers.Studies exist reporting that doping with molybdenum is an effective way of improving photoactivity thanks to thecreation of low conduction band states that act as electron donors (Stengl & Bakardjieva ). Doping with copper has also been shown to enhance photocatalytic activityunder visible light due to the creation of new states in the valence band that lead to a smaller band gap (Navas et al.).

However, there is not many studies of photocatalyticapplications using Mo and Cu as dopants. Chaudharyet al. (Chaudhary et al. ) reported the effect of co-dopingof TiO2 with Mo and Cu using different synthesis methods and proportions of the dopants and evaluated the photoac- tivity of TiO2 under UV light. Thus, it is reasonable to think that co-doping with Cu and Mo will have a synergic effect on photocatalytic efficiency under irradiation with vis- ible light. This study presents the effect of co-doping with copper and molybdenum on the properties of TiO2 using MoS2 and CuCl2·2H2O as precursors. More specifically, a study was performed into the influence of the percentage of co-doping on the optical and structural properties of the photocatalysts synthesized. Finally, the effect of these prop- erties on the enhanced photocatalytic efficiency was assessed in terms of the photodegradation of methylene blue (MB) under visible radiation.The reagents were from commercial sources and used with- out further purification. Titanium(IV) n-butoxide (TNB, 97%) and MoS2 powder were supplied by Sigma-Aldrich; nitric acid (HNO3, 65%) was from Merck; and CuCl2·2H2O (purity 98%) and methylene blue (MB, purity 82%) were supplied by Panreac.The MoS2/Cu/TiO2 samples were synthesized using the low temperature hydrolysis method described in previous studies (Alcantara et al. ). Titanium n-butoxide, MoS2and CuCl2·2H2O were used as precursors of TiO2, Mo andCu respectively. Also, the cost of reagents is low, and the procedure is not expensive.The stoichiometric amount of MoS2 was added to obtain nominal proportions of MoS2 of 1.0%, 5.5%, 10% wt. And a stoichiometric amount of CuCl2·2H2O was added to obtain nominal proportions of Cu/TiO2 of 0.5%. This Cu concen- tration makes it possible to produce a decrease in the band gap energy of TiO2, which leads to enhanced light har- vesting in the visible range. This concentration is also low enough so as not to generate distortions in the anatase phase of TiO2. Distortions in the crystalline phase of Cu-doped TiO2 have been reported previously at higher levels of doping (Navas et al. ). Finally, as-prepared MoS2/ Cu/TiO2 samples were annealed in air for 1 h at 973 K.Since the theoretical quantity of Cu was identical in all thesamples synthesized (0.5% wt.) and only the amount of the source of Mo added was modified, from here on the samples will be named x%MoS2/Cu/TiO2, where x is the theoretical quantity of MoS2 added (0.0%, 1.0%, 5.5%, 10.0%). Also, pure TiO2 was synthesized using the same procedure to compare the results obtained.Pure TiO2 and x%MoS2/Cu/TiO2 samples were character- ised to understand how CuCl2·2H2O and MoS2 affected their photocatalytic properties. Several instrumental tech- niques were used in order to obtain the proportion of Mo and Cu, their crystalline phases and band gap energy.Plasma atomic emission spectroscopy (ICP-AES) was used to study the composition of the x%MoS2/Cu/TiO2 samples using an Iris Intrepid spectrometer, supplied by Thermo Elemental©.

To identify the different crystalline phases and estimate the crystallite size of synthetized samples X-ray diffraction (XRD), model D8Advanced dif- fractometer supplied by Bruker©, was used. The XRD spectra were recorded conditions with 2θ in a range from20 to 70◦ with a resolution of 0.02◦, 40 kV and 30 mA.The structural characterization was accomplished by Raman spectroscopy using a Jobin Yvon U1000 double monochromator equipped with a Hamamatsu R-943 photo- multiplier, using a DPSS 532 nm laser supplied by CNI©, model MSL-III-532 nm–50 mW. On the other hand, to analysis the oxidation states of the samples X-ray photo- electron spectroscopy (XPS) experiments were performed using Kratos Axis UltraDLD spectrometer. Furthermore, the optical properties of the samples and the band gap energy were determined by means of UV-Vis spectroscopy in diffuse reflectance mode (DR-UV-Vis). The equipment, assembled in our laboratory, was composed of an ASB-XE-175 Xenon lamp supplied by Spectral Products©, as the illu- mination source; a USB2000þ spectrometer supplied by Ocean Optics©; and an integrating sphere from SpectraTech©. Finally, the photocatalytic acti ity of the MoS2/Cu/ TiO2 samples was analysed. The photodegradation of MB was assessed using an equipment assembled in ourlaboratory composed of a chamber (750 × 400 × 300 mm) wherein stable temperature (20 ◦C) are achieved during all the experiments. Inside of the chamber, at the top, a halogenlamp (model SL500R, 230 V/50 Hz, 500 W max.) used as the visible irradiation source is located. A manual Lab lifting platform was used to control the distance between the reac- tor and the irradiation source. The initial concentration of the aqueous solution of MB (purity 82%, Panreac) was5·10—5 M, and the amount of photocatalyst was 0.3 gL—1.The reaction time was 6 h, and the mixture of photocatalyst and the MB solution was kept in darkness for 3 h before the visible irradiation to reach the adsorption-desorption equilibrium. The photodegradation of MB was studied by absorbance measurements. A calibration curve was used to determine the evolution of the MB concentration and the kinetics of the photodegradation. The absorbance was measured using a spectrometer (Ocean Optics,USB2000þ) with a UV–Vis–NIR light source (OceanOptics, DH-2000-BAL).

RESULTS AND DISCUSSION
ICP-AES was used to determine the amount to Mo and Cu incorporated into the structure of the samples. The theoreti- cal percentage of Cu by weight added to the samples was 0.5% wt., while the theoretical quantities of MoS2 by weight added were 0.0%, 1.0%, 5.5% and 10.0% wt. Herein- after, the nomenclature used for the samples will be x% MoS2-Cu-TiO2, where x is the theoretical quantity of MoS2 added (0.0%, 1.0%, 5.5%, 10.0%).The percentages of Cu and Mo obtained in the ICP-AES analysis of the samples synthesized are shown in Table 1. The results show that between 80–100% of the precursor of Cu and between 37–59% of the sources of Mo were successfully incorporated into the TiO2 structure. As the percentage of MoS2 added increased there was a slight decrease in the per- centage of Cu incorporated, as Figure 1 shows.On the other hand, the percentage of Mo incorporated was seen to increase when the percentage of MoS2 added increased, but only up to a point, a slight decrease being observed in the sample with the highest concentration (10.0%MoS2/Cu/TiO2).Regarding the phases present, in the pure TiO2 sample, a mixture of anatase-rutile phase is present, rutile phase being predominant. In the TiO2 0.0%MoS2/Cu/TiO2 samples (containing 0.5%Cu and 0.0%MoS2) the presence of Cu is seen to increase the proportion of rutile with regard to the pure sample; in other words, it promotes the transformation of anatase phase to rutile. In the remaining samples (x%MoS /Cu/TiO ) the quantity of anatase increases as thesample, similar percentages of anatase and rutile to the unmodified TiO2 sample were obtained. The same trend isamount of MoS2 added increases, which suggests that the Mo inhibits the transition of anatase to rutile, anatase being the predominant phase. However, sample 10% MoS2/Cu/TiO2 does not follow this trend, showing instead a diffraction pattern similar to the pure TiO2 sample but with a shift in the peaks. This difference suggests the struc- tures are co-doped with Mo and Cu.

However, as the quantity of Mo increases it tends to diffuse towards the sur- face in the case of the sample with the highest concentration (10%MoS2/Cu/TiO2) producing a structural reorganisation in which the internal part presents higher levels of Cu doping, which would justify the appearance of rutile, while the part nearer the surface is enriched with Mo.From the XRD patterns, a semi-quantitative assessment was performed of several crystalline properties, such as the percentage of anatase (A) and rutile (R) phases, the average crystallite size (t) and the unit cell volume (V ).The average crystallite size (t) was calculated according to the Scherrer equation (Landmann et al. ): t ¼ (0.9·λ)/ β·cosθ, where λ is the wavelength of the x-ray radiation(1.5406 Å) and β is the full width at half-maximum height of the most intense peak of the sample. To calculate the pro- portion of anatase (WA) and rutile (WR) phase, given in mass fraction, the relative intensity of (101) and (110) peaks ofanatase and rutile, respectively, was used. The relative con- tent of anatase were estimated by the equation WA ¼ 1/ (1 þ 1.26·IR/IA) (Landmann et al. ) and relative content of rutile as WR ¼ 100- WA; that is, only anatase and rutile phases were considered. Furthermore, the volume of theunit cell was calculated from the values of the lattice con- stant a, b and c corresponding to the maximum intensity peak as V ¼ a2·c (a ¼ b ≠ c, for anatase and rutile phases).Table 2 shows that the percentage of anatase increasedin the MoS2/Cu/TiO2 samples with regard to the pure TiO2 sample, except in the case of 10.0%MoS2/Cu/TiO2, in which a higher percentage of rutile was formed. In thisobserved in the lattice parameters with slight modifications in a and c and a very similar cell unit volume for the pure TiO2 10.0%MoS2/Cu/TiO2 samples.

Due to the effect of the Cu, the predominant phase in the 0.0%MoS2/Cu/TiO2 sample was rutile. Thus, this likeness suggests that the Cu- doping of the crystal may be responsible for the higher pro- portion of rutile generated. And as discussed above, at high concentration of Mo, it tends to diffuse towards the surface. This diffusion process leads to a reorganisation where the inner part of the crystal shows an enrichment in Cu doping. This can justify the presence of a high percentage of rutile in this phase. Moreover, the outer part of the crys- tallites shows an enrichment in Mo, which is of great interest for photocatalytic processes as is discussed below.Furthermore, analysing the evolution of the average crystallite size (t) in accordance with the percentage of Mo and Cu in the structure of the TiO2, t is seen to decrease with an increased percentage of co-doping. This smaller crystallite size may be justified by the incorporation of Cu and Mo into the structure. This distorts the structure and breaks the crystal continuity and resulting in smaller average crystallite sizes, as reported for others dopants (Xu et al.; Aguilar et al. ). In the samples with the highest pro-portions of Mo (10.0%MoS2/Cu/TiO2), the increase in the crystallite size is due to the appearance of rutile. The decrease in the crystal particle size observed in the co-doped samples helps to enhance photocatalytic activity (Zhang et al. ), possibly thanks to their higher specific surface area which leads to an increase of adsorptioncentres which favours the adsorption of species onto the sur- face, as is shown below from XPS results.Rutile was the predominant phase in the pure TiO2, 0.0%MoS2/Cu/TiO2 and 10.0% MoS2/Cu/TiO2 samples, but anatase was the predominant phase in the remaining MoS2/Cu/TiO2 doped samples.Furthermore, as the quantity of MoS2 added to the co- doped samples increases, the bands widen suggesting that the TiO2 structure becomes distorted because of the co- doping with Cu and Mo. In addition, the Raman spectra obtained reveal a blue-shift in the band of the Eg(1) vibrational mode of the anatase phase. The substitution of the Ti ions by Mo generates tensions and distortions in thecrystalline lattice and these distortions lead to the shift in the Raman bands (Kang ; Zhan et al. ). Thus, the results obtained from the Raman spectra are in agreementwith those from XRD given above.The oxidation states of the samples were determined using XPS spectroscopy. As an example, the spectra are shown of the pure TiO2, 0.0%MoS2/Cu/TiO2 and 5.5%MoS2/Cu/ TiO2 samples (Figure 4). In the spectra obtained for the Ti 2p signal (Figure 4(a)), two peaks can be seen at 464 and 459 eV corresponding to Ti 2p1/2 and Ti 2p3/2, consistentwith the Ti4þ values in the TiO2 structure (Wang et al. ). The typical values reported for the Ti 2p3/2 signal with a þ4 oxidation state is 458.66 eV, while the Ti signal with þ3, þ2,0 oxidation state appears at 457.12 eV, 455.34 eV and453.86 eV (Biesinger et al. ).

In addition, Figure 4(b) shows the spectrum of the Mo 3d signal. The peaks of Mo 3d5/2 and Mo 3d3/2 are located at 233.17 and 236.17 eV, which are assigned to Mo6þ (Houng et al. ; Chaudharyet al. ; Huang et al. ). The typical values reported for the Mo 3d5/2 signal with þ4, þ5, þ6 oxidation states appear at 229.7 eV, 231.4 eV and 232.5 eV, respectively (Nguyen et al. ; Erdogan et al. ). No signal can befound related with the presence of Moþ4 and Moþ5, indicat-ing that the main oxidation state of molybdenum is þ6. Inturn, Figure 4(c) shows the Cu 2p signal of the 0.0%MoS2/ Cu/TiO2 and 5.5%MoS2/Cu/TiO2 samples. Two signals appear at 932 and 953 eV attributed to Cu 2p3/2 and Cu 2p1/2, respectively. These signals may correspond to aCu2þ oxidation state, which is coherent for the procedureused (Duke et al. ).Furthermore, in the O 1 s signal of the pure TiO2 sample (Figure 4(d)) an asymmetric peak can be seen that could be composed of the peak at 529.0 eV, which is usually assigned to the oxygen in the TiO2 lattice, together with another peakin the 530–532 eV region, which is indexed to absorbed oxygen species (Luo et al. ). The 0.0% MoS2/Cu/TiO2 sample shows a slight shift of the O signal of the latticewith regard to the pure TiO2 sample due to the surroundings of the Ti being modified because of the doping with Cu. It is also possible to observe an increase in the signal belonging to absorbed species after Cu doping (Figure 4(d)), possible due to the increase in the absorption centres produced tocompensate for the oxygen vacancies created when the Ti4þ is replaced by Cu2þ (Yang et al. ). In turn, in the5.5%MoS2/Cu/TiO2 sample, the peak centred at 529.0 eV shifts to 530.43 eV, indicating that the surface oxygen of the lattice is bonded to Mo (Lu et al. ; Li et al. ). Fur-thermore, as Figure 4(e)–4(f) show, the signal correspondingto absorbed species is more intense in the co-doped sample than in the one doped only with copper. The results suggest that substituting Ti4þ with Mo6þ generates an O deficiency. This may be compensated for with more oxygen adsorbedonto the surface. Thus, Cu-Mo co-doping promotes the adsorption of species onto the surface that may capture the electrons necessary to form H2O2, HO—2 and O—2 speciesthat enhance photocatalytic activity (Simonsen et al. ;Wang et al. ).The diffuse reflectance spectra of the pure TiO2 sample and the x%MoS2/Cu/TiO2 samples were recorded to analyse the absorption of light in the samples synthesized (Figure 5(a)).

The x%MoS2/Cu/TiO2 samples produce absorbed 40% more radiation in the visible range than the pure TiO2. These enhanced optical properties in the x%MoS2/Cu/ TiO2 samples is one of the factors affecting their improved photocatalytic activity, offering more valence electronsthat can easily be converted into free electrons as a result of photonic absorption (Erdogan et al. ).Furthermore, the band gap energy was calculated fromthe Kubelka-Munk (f(R)) function and Tauc plot (Murphy). For TiO2, the Tauc plot satisfies the equation (Serpone et al. ): [ f (R)·hν]1/2 ¼ K·(hν-Eg), where hν isthe photon energy, Eg is the band gap energy, and K is a characteristic constant of each semiconductor. The Tauc plots for the samples analysed in this study are shown in Figure 5(b). The linear segment of the graph of [ f (R)·hν]1/2versus hν was extrapolated to intersect the hν axis to obtain the indirect band gap value of the samples under study (Ser- pone et al. ; Sasca & Popa ). Table 3 shows the bandgap energy values obtained.The evolution of the band gap shown in Table 3 indi- cates that a red shift takes place in the x%MoS2/Cu/TiO2 samples with regard to the pure TiO2 sample. However, as the quantity of MoS2 added to the %MoS2/Cu/TiO2 samples increases, there is a slight increase in the band gap value. The tendency for the band gap value to increase with the increase in Mo doping may be explained by the well-known Burstein-Moss effect (Moss ), whereby thelower states of the conduction band are blocked and tran- sitions can only take place at higher energy levels in the conduction band. This effect is often observed in degener- ated semiconductors in which the Fermi level coincides with or exceeds one of the edges of the permitted bands, something that often occurs when there is a high concen- tration of impurities. Thus, the shift observed towards ahigher band gap suggests an increase in the carrier density when the Mo content increases (Bharti et al. ; Munir et al. ).

Nevertheless, in all the %MoS2/Cu/TiO2 samples the band gap value was lower than that of thepure TiO2 sample, which benefits photocatalytic activity under visible light.Tests were performed into the photodegradation of MB in order to study the effect of the Mo-Cu co-doped TiO2 on photocatalytic activity. A reference experiment was per- formed without a catalyst and no significant degradation of the MB was found. Furthermore, the samples of MB with photocatalyst were kept in the dark for 3 hours to ensure they reached the adsorption-desorption equilibrium.The concentration of MB used for the tests was 5·10—5 M.Figure 6 shows the time evolution of the degradation of MB using the different samples studied as a photocatalyst.Studies about photodegradation routes of MB have been reported in the literature (Houas et al. ; Sinha & Ahmar-MB, generating different products. These products can be generated following several routes. One of these routes isthe oxidation of S following by N-demethylation, hydroxy- lation and mineralization (Bakre et al. ; Ray et al.). Other possible route is the attack of •OH radical tothe group C─Sþ ¼ C of MB. In this stage, O atom is joined to S atom (S ¼ O) and H atom is joined to N atom (N-H). After, two methyl groups joined to N can be lost. Also,these products can be degradated generating several com- pounds with simple rings. Thus, the breakup of the bonds C-S and N-C by •OH radical can generate several structures with single rings, which can finally lead to mineralization inNO3—, SO2—, Cl—, NHþ, CO and H O.In accordance with the results obtained (Figure 6), the photocatalytic degradation of MB, using the samples studiedas a photocatalyst, is a pseudo-first order reaction and and its kinetics can be described by ln(C0/C) ¼ —kKt ¼ kappt, where C0 is the initial concentration of MB, C is the concen-tration of MB at the time of irradiation t, k is the reaction velocity constant, K is the adsorption constant of the reac- tant and kapp is the apparent constant rate (Hwang et al.).

The kapp values and the calculation of the percentageof degradation are shown in Table 4. The constant ratevalues show that the co-doped x%MoS2/Cu/TiO2 samples were between 3.7–10.4 times faster than the pure TiO2 samples. If we compare the co-doped x%MoS2/Cu/TiO2 samples with the Cu-doped sample (0.0%MoS2/Cu/TiO2) or with the commercial sample (P25 TiO2), the rate can be seen to be 1.4 times higher, the 10.0%MoS2/Cu/TiO2 sample presenting the fastest rate (up to 3.7 times faster).Regarding the percentage of photodegradation of MB, the P25 TiO2 (51.5%) and 0.0%MoS2/Cu/TiO2 (52.6%)samples presented higher degradation than the pure TiO2 sample (22.3%). In turn, as the percentage of Mo increased in the co-doped samples, so did the percentage of degra- dation, reaching 92.9% in the 10.0%MoS2/Cu/TiO2 sample. Photocatalytic activity depends on many of the catalyst’s properties including its band gap energy, crystallite size, sur- face structure, extent of the crystallinity and the structure of the material. The band gap values obtained for the all the co- doped samples were lower than that of the pure TiO2 sample. However, as the percentage of doping with Mo increased the band gap value rose slightly; thus, although the band gap values have an influence, they alone cannot explain the enhanced photoactivity observed. In turn, ana- lysing the differences in the size of the crystallite and in the proportion of anatase-rutile obtained for the pure TiO2 sample and the co-doped samples (Table 2), the MoS2/Cu/ TiO2 samples are seen to have a smaller crystallite size than the pure sample. Also, the amount of O speciescapture electrons from the conduction band and gaps from the valence band. The captured electrons and gaps may betransferred to the surface of the photocatalyst and promote the photocatalytic reaction (Gupta & Tripathi ). In addition, as reported by Khan et al. (Khan & Berk ), Mo6þ may act as an electron trap and promote charge sep- aration and consequently decrease the recombination ofphotogenerated charge carriers. Thus, dopant enrichment close to the surface of the photocatalyst along with increased movement of the charge carriers would explain the increase in photocatalytic activity observed in the co- doped samples with the highest percentage of dopant.

CONCLUSIONS
The results from XRD and Raman spectroscopy suggest that the co-doping of the TiO2 with Mo and Cu using MoS2 and CuCl2·2H2O as the precursor of the dopant took place internally. This is supported by the data obtained using XPS, which also indicate the possible presence of oxygen adsorbed onto the surface of the catalyst to compensate for the internal O deficiencies. In turn, the percentage of Cu added favours the transition from anatase to rutile. How- ever, increasing amounts of Mo have the opposite effect, inhibiting the transformation of anatase to rutile. Further-more, in every case the band gap of the x%MoS /Cu/TiO adsorbed increased in doped, as it is deduced from XPS 22 results, which favours photocatalysis. In addition, regarding the phases, although all the samples presented a mixture of anatase and rutile phase, the co-doped samples generated a higher proportion of anatase phase, which presents greaterphotoactivity (Scanlon et al. ; Chimupala et al. ).However, the 10.0%MoS2/Cu/TiO2 sample had very similar values in terms of the crystallite size and the proportion of anatase-rutile phase to the pure sample but was the most effi- cient sample.Typically, anatase phase shows higher photoactivity than rutile phase, however Luttrell et al. (Luttrell et al.) reported a study on photocatalytic activity as a func-tion of film thickness for anatase and rutile and showedsamples was lower than that of the pure TiO2 sample. How- ever, the increase in Mo in the co-doped samples generated a slight shift towards higher values that can be explained by the Burstein-Moss effect. Finally, the x%MoS2/Cu/TiO2 samples presented a significant improvement in photocata- lysis processes compared with pure TiO2 samples. Of the co-doped samples studied, the 10.0%MoS2/Cu/TiO2 sample was the most efficient photocatalyst with a percen- tage of degradation above 92% after 6 hours of irradiation under visible light. The results indicate that this is a promising material for use in Methylene Blue photodegradation reactions under conditions of solar irradiation.