Spray pyrolysis synthesis and characterization of Mg 1-x Sr x MoO 4 heterostructure with white light emission

Molybdates are inorganic materials with great potential in white phosphors application, being an alternative to traditional lighting sources. In this study, we report the synthesis and characterization of Mg 1-x Sr x MoO 4 (x ¼ 0, 0.25, 0.50, 0.75, and 1) powders with white light-emitting properties. Using X-ray diffraction, the formation of the monoclinic b -MgMoO 4 phase was observed for x ¼ 0 and the formation of the tetragonal scheelite phase of SrMoO 4 was observed for x ¼ 1. The formation of a heterostructure composed of both phases was found for compositions with x ¼ 0.25, 0.50 and 0.75. Scanning and tras-mission electron microscopy images showed that the Mg 1-x Sr x MoO 4 particles exhibit a spherical morphology formed by several primary nanoparticles. Raman scattering spectroscopy enabled the accurate identi ﬁ cation of the Raman modes for different compositions and their assignment to either the SrMoO 4 or b -MgMoO 4 modes. The bandgap energies were determined to ﬂ uctuate between 4.25eV and 4.44 eV, being in ﬂ uenced by the degree of structural disorder. The photoluminescence emission spectra of the nanoparticles showed neutral-and cool-white emission with high-quality white light (CRI > 80%). The samples synthesized with x (cid:2) 0.50 are potential materials for the application in LED lamps (6500 K) and pure white-light sources (5500 K).


Introduction
Since its discovery, light-emitting diodes (LEDs) have been gradually replacing traditional lighting sources, such as incandescent and fluorescent lamps. LED enhancement studies have led to a considerable increase in luminous efficiency and also to discover materials with a better lifetime, lower energy consumption, greater stability, and which are environmentally friendly [1e4].
Generally, white light is achieved by using the combination of an InGaN chip (blue emission) and a YAG:Ce 3þ phosphor (yellow emission), but this methodology results in high color temperature correlation (CCT) and low color rendering index (CRI) [5,6]. One of the alternatives used to obtain white light sources is the combination of phosphors doped with rare-earth elements that emit red, green, and blue light. These phosphors belong to several classes of materials, such as rare-earth oxides, molybdates, tungstates, silicates, phosphates, or niobates [7e9]. Unfortunately, the use of rare-earth elements in phosphors implies higher costs and increases the toxicological risk [10]. Thus, as an alternative technology, the synthesis of several materials by different routes using morphological and structural modifiers and forming heterostructures has been carried out in order to obtain white light emission phosphors without the use of rare-earth elements [11e14].
The ultrasonic spray pyrolysis (USP) method had shown to be an interesting route to obtain white-emitting photoluminescent materials because of the structural and morphological alterations achieved during the process [15e18]. Also, the USP method has other advantages such as the simplicity of equipment, a quick and single-step process, a non-complex precursor solution, and the formation of powders with high homogeneous and purity [19e22]. Recently, the USP method has been used to obtain different molybdates such as NiMoO 4 [23,24], SrMoO 4 [25], La 2 Mo 2 O 9 [26], Na 2 Mo 2 O 7 [27], and CoMoO 4 [28].
Molybdates are inorganic materials that have great potential in different applications such as phosphors, supercapacitors, scintillators, and photocatalysis [29e35]. Among the molybdates, magnesium molybdate (MgMoO 4 ) stands out because it has an efficient matrix to host other elements. The monoclinic structure of its bphase is formed by two octahedral [MgO 6 ] and two tetrahedral [ MoO 4 ] sites [36,37]. On the other hand, strontium molybdate (SrMoO 4 ) has a tetragonal scheelite-type structure formed by one dodecahedral [SrO 8 ] and one tetrahedral [ MoO 4 ], resulting in a material with excellent thermal and chemical stability [38,39]. The photoluminescent properties of MgMoO 4 and SrMoO 4 have been recurrently studied, showing variations in: synthesis routes, such as microwave-assisted hydrothermal, solid-state reaction, combustion; and/or inserting different elements in the structure, such as Ba 2þ , Tm 3þ , Dy 3þ , Yb 3þ , Pr 3þ , Eu 3þ , Bi 3þ , Li þ , Na þ or K þ [37,40e47]. However, studies about the structural and photoluminescence properties of the solid-solution formed between MgMoO 4 and SrMoO 4 structures are not found in the literature.
In this study we report the synthesis of Mg 1-x Sr x MoO 4 (x ¼ 0, 0.25, 0.5, 0.75, and 1) powder samples, prepared by the USP method, as well the characterization of their structural, morphological, vibrational, and optical properties at room conditions by means of x-ray diffraction (XRD), field-emission scanning electron microscopy (FESEM), transmission electron microscopy (TEM) and high-resolution transmission electron microscopy (HRTEM), Raman scattering (RS), photoluminescence (PL) and ultravioletevisible (UVeVis) reflectance measurements.  [16,48]. The precursor solution for MgMoO 4 was prepared by dissolving 3.75 mmol of magnesium nitrate, 3.75 mmol of molybdic acid, and 0.75 mmol of guanidine hydrochloride in 150 ml of distilled water under magnetic stirring at 25 C. The precursor solution was atomized using an ultrasonic nebulizer (frequency of 2.4 MHz) with an airflow of 3 L min À1 and the heating temperatures were from 600 C for zone 1 and 800 C for zone 2. More details on the equipment and implemented technique are listed in Ref. [15]. The other precursor solutions were made according to the stoichiometry shown in Table 1.

Characterization of Mg 1-x Sr x MoO 4 powders
Powder XRD patterns of Mg 1-x Sr x MoO 4 were collected within the 10 to 120 angular range with step speed of 0.02 min À1 using an XRD 7000 Shimadzu diffractometer and monochromatic Cu K a (l ¼ 1.5406 Å) radiation. The diffraction patterns were analyzed by means of the General Structure Analysis System (GSAS) with graphic interface EXPGUI [49] in order to perform the Rietveld refinement [50] and to analyze the possible structural modifications. The following parameters were refined: scaling factor and phase fraction; background (displaced Chebyshev polynomial function); peak shape (Thomson-Cox-Hasting pseudo-Voigt); change in the network constants; fractional atomic coordinates; and isotropic thermal parameters.
The powder morphology was examined using a FESEM (Carl Zeiss, model Supra 35-VP) operating at 6 kV. Transmission electron microscopy (TEM) and high-resolution transmission electron microscopy (HRTEM) images were obtained with an FEI Tecnai G2 F20 microscope operating at 200 kV. The sample was initially prepared by using 20 mg of MS50MO dispersed in 3 mL of acetone via ultrasonication. The copper TEM grids with carbon film (3 nm) were rapidly submerged in this dispersion and then dried in a vacuum for 12 h.
Non-polarized RS measurements were carried out using a Lab-RAM HR UV coupled to a Peltier cooled CCD camera, using a 632.8 nm HeNe laser excitation line with a power smaller than 10 mW and a spectral resolution better than 2 cm À1 . Samples were checked before and after each RS measurement to ensure that no heating effects occurred during the measurements. RS spectra peaks were analyzed by fitting them to Voigt profiles where the spectrometer resolution is taken as the fixed Gaussian width.
The UVeVis reflectance spectrum was obtained using a UVeVis spectrometer (Shimadzu, model UV-2600) and PL spectra were measured using a Thermal Jarrell-Ash Monospec 27 monochromator and a Hamamatsu R446 photomultiplier. The excitation source used was a krypton laser with a wavelength of 325 nm (Coherent Innova) with an output of approximately 13.3 mW.

Structural characterization
Powder XRD patterns of Mg 1-x Sr x MoO 4 (x ¼ 0, 0.25, 0.50, 0.75, Table 1 Codes for Mg 1-x Sr x MoO 4 powder samples with nominal concentration, x ¼ 0, 0.25, 0.5, 0.75 and 1. x (Mg 1-x Sr x MoO 4 ) Code and (e) SMO. * is referent to JCPDS 21e096 and # is referent to JCPDS 85e0809. It is noted the tetragonal phase increases with increasing the value of x due to the increase in the substitution of Mg 2þ cations by Sr 2þ cations. and 1) samples are shown in Fig. 1 and Fig. S1 (see Supporting Information). The MMO sample was indexed in a unit cell with a monoclinic structure (b-MgMoO 4 polymorph) with space group C2/ m (number 12) in agreement with JCPDS 21e0961 and the literature [51,52]. The SMO sample was indexed in a unit cell with a tetragonal scheelite-type structure with space group I4 1 /a (number 88) in agreement with JCPDS 85e0809 and the literature [53,54]. The MS25MO, MS50MO, and MS75MO samples were indexed with JCPDS 21e096 and JCPDS 85e0809 cards, being observed the coexistence of all Bragg peaks of each structure. Note that the fraction of the tetragonal phase increases with increasing the value of x since there is an increase in the substitution of Mg 2þ cations by Sr 2þ cations in the heterostructure. Fig. 2 and Table 2 show the results obtained from the Rietveld refinement of XRD patterns of the different samples. The XRD pattern of the MMO sample was well matched to ICSD 20418 (b-      Small changes in the structural parameters have been observed between samples, such as in the unit-cell parameter, crystallite size, and displacement of the atomic positions within the crystalline structure. These phenomena may be related to defects caused during the USP process which, generate oxygen vacancies, possible substitutions of Mg 2þ ions for Sr 2þ ions in the monoclinic structure and vice versa, distortions in the [MoO 4 ] 2clusters, and in the interface itself between the two heterostructure phases.    (Fig. 4a) exhibit slight distortion together with more apparent surface roughness. On the other hand, the microspheres of SMO (Fig. 4e) have a smoother surface with small distortions in the circumference of the microsphere. Thus, the MS25MO, MS50MO and MS75MO microspheres (Fig. 4bed) show the variations in morphology between MMO and SMO. The variations in the morphology are related to the utilized precursor reagents due to the solubility in water and decomposition temperature. Several physicochemical phenomena occur simultaneously in the USP process, such as evaporation of the solvent on the surface of the droplet, the diffusion of solvent vapors away from the droplet in the gas phase, and the melting of metal nitrates. The metal nitrates tend to melt before the solvent is completely removed forming a molten salt and this molten salt prevents the solvent from being removed, thus giving rise to the porous or hollow microspheres [55].

Morphological characterization
Since metal nitrates were used for Mg 1-x Sr x MoO 4 synthesis, and the result is the formation of porous and hollow microspheres with holes. In addition, metal nitrates generally tend to have a high degree of supersaturation and reagents with this characteristic tend to form a high number of nanosized crystallites that form the microspheres [55]. It is noted that the microspheres are formed by small particles of nanometric size, which joined together during the USP process. As the precursor reagents used were magnesium nitrate and strontium nitrate and present high solubility in water (0.71 g/ml of H 2 O and 0.80 g/ml H 2 O, respectively [56]), this favors the formation of nanocrystals which composes the microspheres that are observed.
The TEM and HRTEM were utilized for investing the nanocrystals composing the microspheres. The MS50MO particles are formed by microspheres composed of several nanoparticles, as seen in Fig. 5a. Moreover, the microsphere has the (220) plane of the MgMoO 4 and the (112) plane of the SrMoO 4 in the same particle (Fig. 5b). Thus, the presence of these planes and the interface between phases confirm the formation of the heterostructure. Fig. 6 shows the RS spectra measured in the different samples. We found that the spectra gradually change from the typical RS spectrum of monoclinic b-MgMoO 4 (bottom trace) to that expected in tetragonal scheelite-type SrMoO 4 (top trace). For intermediate compositions, we found the coexistence of the monoclinic and tetragonal end-member compounds. The intensity of the Raman modes gradually loses intensity as the content of Sr increases in the sample, gradually becoming the dominant Raman modes of SrMoO 4 , as expected. In Table 3 we report the frequencies of the Raman modes observed in each sample. Group theory analysis predicts that monoclinic MgMoO 4 has 36 Raman-active modes (19A g þ 17B g ) and tetragonal scheelite-type SrMoO 4 has 13 Ramanactive modes (3 A g þ 5B g þ 5E g ).

Vibrational characterization
In previous works, more than 25 Raman-active modes have been reported for MMO [57,58]. In our case, we have observed 25 modes, in which 20 of our modes agree within uncertainties with the frequencies reported in the literature (see Table 3). There are several weak modes reported by Miller [57] and by Coelho et al. [58] which we did not observe, and there are three weak modes that we found at 188, 236, and 889 cm À1 which have not been previously reported. The comparison with the works of Miller [57] and by Coelho et al. [58] indicates that all the modes we measured in MMO can be assigned to b-MgMoO 4 . In total, 32 out of 36 Raman-active modes of MMO can be identified between ours and previous measurements.
In the case of SMO, the agreement with previous studies [59,60] is excellent (Table 3). For intermediate compositions, we found that the samples have the Raman modes of the two compounds, as expected for the prepared heterostructures. The mode frequencies are only slightly affected by the composition. In MS25MO, the presence of scheelite-type Raman modes of SMO in the highfrequency region is evident in Fig. 6e. Notice how these peaks and other new peaks gradually win intensity in Fig. 6, while the original MMO modes gradually lose intensity as the Sr content increases. However, the MMO Raman modes are still present in MS75MO, as can be clearly seen in the high-frequency region of Fig. 6.
There are a few interesting facts that we would like to remark on here. One is the decrease in the number of Raman modes in SrMoO 4 in comparison with b-MgMoO 4 ; a consequence of the increase in the symmetry of the tetragonal scheelite with respect to the monoclinic structure as well as in the decrease of the number of formula units in the primitive unit cell (Z ¼ 1 in SrMoO 4 and Z ¼ 2 in MgMoO 4 ). This is quite noticeable in the high-frequency region where only three modes (one doubly degenerated) are expected in SMO associated to the internal stretching vibrations of the MoO 4 tetrahedron. In contrast, eight internal modes are expected in the high-frequency region of MMO due to the presence of two different distorted MoO 4 tetrahedra. Another qualitative difference between the RS spectra of SMO and MMO is that the Raman modes in the high-frequency region corresponding to the stretching modes of the MoO 4 tetrahedra cover a much smaller frequency region (between 790 and 900 cm À1 ) in SMO than in MMO, where highfrequency modes are found between 740 and 980 cm À1 . This is due to the same four MoeO distances (~1.88 Å) in the regular MoO 4 tetrahedron of SMO in comparison to those of MMO where the two irregular MoO 4 tetrahedra show MoeO distances in the range between 1.73 and 1.85 Å. In addition, the high-frequency Raman modes of SMO are much more intense than in MMO, with the high intensity of the SMO Raman modes being typical of scheelite-type oxides. These qualitative differences between the Raman spectra of both compounds enable accurate identification of the Raman modes of each phase in Mg 1-x Sr x MoO 4 samples and their assignment to either SMO or MMO modes. Fig. 7 presents the estimated values for the band-gap energy of Mg 1-x Sr x MoO 4 samples. UVeVis spectroscopy was utilized for obtaining the diffuse reflectance data of samples and the gap energy values (Egap) were estimated using the Kubelka-Munk function [61] for converting the reflectance data to absorbance followed by the Wood and Tauc method [62]. The optical band-gap energy is assumed by ahv f (hv-E gap ) 1/k [62], where a is the absorbance, h is the Planck constant, v is the frequency, and k is indicated for the different kind of transitions. In the literature, molybdates are typically reported with permitted direct electronic transition [63e67], so the results from Mg 1-x Sr x MoO 4 samples were analyzed assuming k ¼ 1/2, which is the expected value for such transitions.

UVevisible spectroscopy
The estimated values for samples are according to the literature. Silva et al. [68] and Zhu et al. [45] synthesized SrMoO 4 and obtained band gap values between 4.17 and 4.36 eV (agglomerates particles by complexation) and between 3.81 and 2.71 eV (nanoparticles by thermal decomposition), respectively. Wannapop et al. [69] and Zhang et al. [70] produced MgMoO 4 and estimated the band gap value of 5.15 eV (fibers via electrospinning) and between 3.72 and 3.78 eV (nano-flowers and nanosheets via self-assembly), respectively.
The molybdates have wide-bandgap values due to the low number of intermediary levels between the valence band and the conduction band. The band-gap is mainly determined by Mo 4d states and oxygen 2p states. This is the reason why (as occurs in tungstates) the bandgap only changes 0.2 eV from MgMoO 4 to SrMoO 4 [71]. In addition, the band-gap energy could be strongly influenced by the degree of order and structural disorder [72,73]. The greater the structural disorder, the lower the band-gap of molybdate will be due to the increase of intermediate levels.
Sczancoski et al. [74] calculated the bandgap of BaMoO 4 and related the decrease in bandgap with the increase of defects due to increases of intermediary energy levels, being 4.93 eV for the material free of structural defects and 4.10 eV for the material with distorted structure. Longo et al. [75] calculated the bandgap of CaMoO 4 and demonstrated than units cells ordered exhibit larger bandgap comparing to cells disordered (4.57 eVe2.57 eV, respectively). It is also reported that other factors may influence the bandgap of molybdates, such as distortion of [MoO 4 ] À2 clusters, the formation of oxygen vacancies and growth mechanisms which may promote an increase of intermediate levels and decrease the bandgap [74,76,77].
The MMO sample (Fig. 7a) shows 4.25 eV, while SMO (Fig. 7e) presents 4.44 eV. However, the MS25MO, MS50MO and MS75MO samples have intermediate gap bands, but with greater proximity to the band-gap shown in the SMO. Thus, the SMO phase has a greater influence on the band-gap of the formed heterostructure than the MMO, even when the SMO presents a smaller phase amount. This result suggests the presence of a large band-gap bowing parameter in this solid solution.

Photoluminescence (PL)
The PL emission spectra of the Mg 1-x Sr x MoO 4 (x ¼ 0, 0.25, 0.5, 0.75, and 1) powders obtained by the USP method were obtained at room temperature (see Fig. 8). The PL emission spectra of the samples exhibit a broadband spectrum. This phenomenon occurs when the relaxation involves the participation of many states in several places within the band-gap of the material, and this is also known as a multiphonon and multilevel process [78].
The PL emission of molybdates is fundamentally related to the crystalline structure and the distortion of their polyhedral units. Intermediate levels appear within the molybdate band-gap when a molybdate is excited and there are distortions in its clusters [MoO 4 ] because of the energy absorption of low-energy states (O 2p orbitals) and promotion of high energy states (Mo 4d orbitals) [79]. The PL emission colors are due to electronic transitions associated with the structural arrangements and cluster-to-cluster transfer charge, with the emission in the yellow-red region being associated with deep defects, structural disorder and oxygen vacancies in [MoO 4 ] clusters, while the emission in the blue-green region associated with shallow defects and structural ordering, being the color emission association proven in literature by the thermoluminescence spectrum and EPR analysis [73,80e83]. Thus, the deconvolution of the Gaussian curves of Mg 1-x Sr x MoO 4 samples was performed using the PeakFit 4.12 software to estimate the contribution of each color in the PL emission of the samples, as can be seen in Fig. 9.
The deconvolutions propose that all samples have a predominance of PL emission in the blue region. The MMO sample shows a small difference between the proportions of each color. Thus, it is believed that the color emission is generated by cluster-to-cluster transfer charge and the ratio between ordered [ Meanwhile, the SMO sample shows a higher amount of blue and orange colors compared to green and red colors. It is believed that the emission of the colors follows the same line of thought as the MMO, but the emission occurs due to the ratio between ordered On the other hand, if the value of x increases, the spectrum tends to change from the MMO spectrum to the SMO spectrum. This occurs due to the proportion of each phase within the heterostructure, as well as defects generated during the formation and charge transfer at the interface between the phases. Thus, it is believed that the color emission occurs due to the ratio between ordered [ The CIE chromatic coordinates were calculated by determining the x and y coordinates to verify the color emitted by the spectrum of samples. The CIE is an important factor for checking the performance of matches and is calculated by integrating the X, Y and Z values. The CIE (x, y) coordinates of Mg 1-x Sr x MoO 4 samples (x ¼ 0, 0.25, 0.5, 0.75 and 1) are shown in Fig. 10, while Table 4 lists the CIE coordinates values, correlation color temperature (CCT), color rendering indices (CRI) and the color emitted from each sample.
White LEDs which show light generation with CCT between 4000 and 8000 K have neutral-and cool-white emission, requiring high CRI values (>80) to achieve a high-quality white light source [84,85]. All Mg 1-x Sr x MoO 4 samples demonstrated emission in the white region and showed CRI>80, so these materials can be considered as promising high-quality white light sources. Several studies have used the doping of rare earth elements in molybdates to generate white light. Adya et al. [86] obtained the SrMoO 4 :(0.005 mol Eu, 0.01 mol Tm) by the coprecipitation method and the CCT response was 6863 K. In using the solid-state method, Cao et al. [87] obtained Ca 0.95 MoO 4 :0.05Sm 3þ and presented CCT response of 5608 K and CRI of 92.1%. Ham et al. [88] synthesized Ca 0.8 La 0.2-x MoO 4 :xDy 3þ with x ¼ 0.01 and 0.05 by the hydrothermal method, in which the white light emission had CCT of 6429 K and 5555 K, respectively. Laguna et al. [89] used the precipitation method to obtain CaMoO 4 :(0.35% Dy, 0.15% Eu) and had a response in white light with CCT of 5407 K. The toxicological significance of rare earth elements still needs to be explored, but studies show that prolonged exposure to RE dust can induce bronchiolar, alveolar and interstitial histological reactions in the lungs [90,91]. Thus, obtaining phosphors with white light emission that do not use RE have mainly been required for biological reasons. Consequently, the ultrasonic spray pyrolysis method has been shown to be an interesting route to obtain phosphors with white light emission without the use of rare earths materials, introducing simplicity and fast synthesis [15e17]. Therefore, the Mg 1-x Sr x MoO 4 samples synthesized by the USP method have shown to be interesting in the application of LEDs, mainly for the MMO and MS50MO samples because they are close to the pure white CCT value (5500 K), and the MS25MO sample because of the CCT value being close to that of LED commercial lamps (6500 K).