Structural and Spectroscopic Effects of Li + Substitution for Na + in Li x Na 1 − x CaLa 0.5 Er 0.05 Yb 0.45 (MoO 4 ) 3 Upconversion Scheelite-Type Phosphors

: New triple molybdates Li x Na 1 − x CaLa 0.5 (MoO 4 ) 3 :Er 3+ 0.05/Yb 3+ 0.45 (x = 0, 0.05, 0.1, 0.2, 0.3) were manufactured successfully using the microwave-assisted sol-gel-based technique (MAS). Their room-temperature crystal structures were determined in space group I 4 1 / a by Rietveld analysis. The compounds were found to have a scheelite-type structure. In Li-substituted samples, the sites of big cations were occupied by a mixture of (Li, Na, La, Er, Yb) ions, which provided a linear cell volume decrease with the Li content increase. The increased upconversion (UC) efﬁciency and Raman spectroscopic properties of the phosphors were discussed in detail. The mechanism of optimization of upconversion luminescence upon Li content variation was shown to be due to the control of excitation/energy transfer channel, while the control of luminescence channels played a minor role. The UC luminescence maximized at lithium content x = 0.05. The mechanism of UC optimization was shown to be due to the control of excitation/energy transfer channel, while the control of luminescence channels played a minor role. Over the whole spectral range, the Raman


Introduction
Complex molybdates are among the most widely studied oxide materials in modern electronics and photonics owing to their excellent performances [1][2][3][4][5][6]. It was noted that over the past few decades that special research attention was paid to the use of molybdate hosts in the design of new optical and electronic materials [7][8][9][10][11][12], and the syntheses of efficient rare-earth-containing molybdate-based phosphor materials working in the visible range became a hot topic of solid-state technology because of the possible wide-range cation substitution and related tuning of physical properties [6,8,[11][12][13][14][15][16].
Recently, the compounds with a scheelite-type structure were proposed as efficient phosphor hosts [3,[10][11][12][17][18][19][20]. As it can be reasonably assumed, in the disordered scheelite-type structure, the trivalent rare-earth ions could be substituted, at least partially, by different Ln 3+ ions, and the crystal lattice efficiently could incorporate the ions due to similar radii of Ln 3+ ions, and that would increase the limit at the appropriate doping level. Among rare-earth ions, the Er 3+ ion can be used for the infrared to visible light conversion through the frequency up-conversion (UC) process according to its appropriate electronic energy level configuration. The Yb 3+ ions co-doping can remarkably enhance the UC yield, and it is provided by the efficient Yb 3+ → Er 3+ energy transfer. In this cation pair, the Er 3+ ion activator is the luminescence center in the UC particles, and the Yb 3+ sensitizer increases the UC luminescence efficiency [21][22][23][24][25].
The simple and double molybdate crystals are widely applied in photonic and laser technologies because of their specific structural, thermal and electronic characteristics, high chemical stability and excellent spectroscopic properties [3,4,10,[26][27][28][29][30][31][32][33]. However, ternary molybdates are less studied, and only several ternary compounds, including scheelite-type ones, have been recently considered [12,18,[34][35][36][37][38][39][40]. For the practical application of UC photoluminescence in devices and systems, such as three-dimensional displays, lasers, light-emitting elements and biological detectors, the characteristics, such as the homogeneous UC particle size distribution and uniform morphology, need to be reached. Usually, complex molybdates are prepared by a multistep solid-state reaction method, and it requires high temperatures, long heating process and subsequent grinding. The stages may occasionally result in a loss of the emission intensity. In comparison, the sol-gel-based process has some advantages, including high starting atom intermixing, lower calcination temperature, small particle size and a narrow particle size distribution, which are promising for good phosphor characteristics. However, a disadvantage of long gelation time is common for sol-gel process. In comparison to the traditional methods, a very short reaction time, small-sized particles, narrow particle size distribution and high purity of the final polycrystalline product are the characteristics of microwave synthesis. Microwave heating is delivered to the material surface through the convection and/or radiant heating which is transferred to the material bulk via conduction [3,12,14,[41][42][43][44]. Thus, the microwave sol-gel process is an efficient method that yields high-homogeneity powder products, and it is emerging as a viable alternative approach to the quick synthesis of high-quality crystalline luminescent materials.
In the present study, the Li x Na 1−x CaLa 0.5 Er 0.05 Yb 0.45 (MoO 4 ) 3 compounds with the fixed Er 3+ and Yb 3+ contents and x = 0-0.3 were synthesized by the microwave sol-gel method, and the structural and spectroscopic properties were evaluated to determine the effects of Li + substitution for Na + . Previously, the scheelite-type NaCaLa(MoO 4 ) 3 :Er,Yb solid solutions were observed and the rare earth element ratio (La 0.5 Er 0.05 Yb 0. 45 ) was determined as optimal to reach strong UC emission [45]. Thus, in the present work, only this rare earth element ratio was selected, and a wide variation of the Li/Na ratio was implemented to elucidate the solid solution range and related spectroscopic effects. The synthesized products were characterized by scanning electron microscopy (SEM), X-ray diffraction (XRD) analysis and energy-dispersive X-ray spectroscopy (EDS). The spectroscopic properties were comparatively evaluated using Raman spectroscopy and photoluminescence (PL) measurements. O was prepared by the addition of 100 mL distilled water. Subsequently, the individual two chemical solutions in each sequence for (a)-(e) were co-mixed slowly in a 450 mL Pyrex glass. The co-mixed solutions were vigorously stirred and adjusted to pH = 7-8 using CA and 8M NH 4 OH. At this time, the molar ratio of CA to the full employed cation metal ions (CM) would be recommended to reach at 2:1. The solutions were slowly heated up to 80-100 • C before microwave treatments. At this stage, the final solutions of about 20-30 mL in volume appeared to be highly transparent sol formations in 450 mL Pyrex glasses. Then, the resultant solutions were moved into a microwave oven (Samsung Company, Republic of Korea, frequency of 2.45 GHz and maximum power of 1250 W). The parameters of the microwave process applied in the present experiment can be found in previous studies [36,38,45]. After the microwave treatment, the black gels dried at 120 • C were ground and annealed at 850 • C

Characterization
The structural properties of the synthesized powder products LixNa1-xCaLa0.5Er0.05Yb0.45(MoO4)3 were determined using an X-ray (D/MAX 2200, Rigaku, Tokyo, Japan) diffractometer equipped with the Cu Kα radiation (λ = 1.5406 Å) source. The scans were carried out over the diffraction angle range of 2θ = 5-90° at room temperature. The 2θ size step was 0.02°, and the counting time was as long as 5 s per step. For the Rietveld analysis, the TOPAS 4.2 software package was used [46]. The micromorphology of the synthesized particles was observed using scanning electron microscopy (SEM) (JSM-5600, JEOL, Tokyo, Japan). The photoluminescence spectra were acquired at room temperature using a spectrophotometer (Perkin Elmer LS55, Beaconsfield, UK). In the measurements, the samples were excited at 980 nm. The Raman scattering measurements were implemented using a LabRam Aramis (Horiba Jobin-Yvon, Palaiseau, France) device with the spectral resolution of 2 cm −1 . The 514.5 nm line of an Ar ion laser was exploited as an excitation source and the power on the sample surface was kept at the 0.5 mW level to avoid the possible sample decomposition.

Characterization
The structural properties of the synthesized powder products Li x Na 1−x CaLa 0.5 Er 0.05 Yb 0.45 -(MoO 4 ) 3 were determined using an X-ray (D/MAX 2200, Rigaku, Tokyo, Japan) diffractometer equipped with the Cu Kα radiation (λ = 1.5406 Å) source. The scans were carried out over the diffraction angle range of 2θ = 5-90 • at room temperature. The 2θ size step was 0.02 • , and the counting time was as long as 5 s per step. For the Rietveld analysis, the TOPAS 4.2 software package was used [46]. The micromorphology of the synthesized particles was observed using scanning electron microscopy (SEM) (JSM-5600, JEOL, Tokyo, Japan). The photoluminescence spectra were acquired at room temperature using a spectrophotometer (Perkin Elmer LS55, Beaconsfield, UK). In the measurements, the samples were excited at 980 nm. The Raman scattering measurements were implemented using a LabRam Aramis (Horiba Jobin-Yvon, Palaiseau, France) device with the spectral resolution of 2 cm −1 . The 514.5 nm line of an Ar ion laser was exploited as an excitation source and the power on the sample surface was kept at the 0.5 mW level to avoid the possible sample decomposition.

Quantum-Chemical Calculations
Quantum-chemical calculations were carried out within the framework of density functional theory (DFT) [47][48][49] using the CASTEP (Cambridge Serial Total Energy Package) code [50]. The meta-generalized gradient approximation (meta-GGA) and RSCAN [51] (improved version of SCAN [52]) functional were chosen. The on-the-fly generated normconserving pseudopotentials [53] with a cutoff energy equal to 800 eV were used for all investigated compounds. The Brillouin zone (BZ) was sampled according to the Monkhorst-Pack scheme [54] with the 4 × 4 × 5 k-point mesh that provided separation of k-points equal to 0.05 Å −1 . Considering that the structural site of Ca ion was occupied by a mixture of Na, Li, Ca, La, Er and Yb ions with the total occupancy equal to 1, the mixed ion was simulated within the virtual crystal approximation [55]. The tolerance for ground-state wavefunctions founding was chosen to be 2 × 10 −6 eV. Pulay density mixing scheme was chosen for electronic minimization.

Particle Morphology and Phase Composition
The SEM patterns of the synthesized powder samples are shown in Figures 1 and S1 (see Supplementary Materials). In general, the particle morphologies are similar for all samples. The particles are partly coalescent due to the material diffusion between the grains, and this is typical of oxides subjected to a high temperature annealing [9,56,57]. In all SEM patterns, the biggest grains were~5-7 µm in size. For comparison, in Figure 1, the SEM patterns are shown for pure NCLM:EY and LiNCLM:EY-0.3 with the highest Li content. In both samples, big grains were mixed with low-sized grains. However, in Figure 1b, the presence of partly faceted big grains is evident and the quantity of low-sized particles was lower. Thus, the presence of Li in the intermediate gels stimulated the grain growth and faceting.

Quantum-Chemical Calculations
Quantum-chemical calculations were carried out within the framework of density functional theory (DFT) [47][48][49] using the CASTEP (Cambridge Serial Total Energy Package) code [50]. The meta-generalized gradient approximation (meta-GGA) and RSCAN [51] (improved version of SCAN [52]) functional were chosen. The on-the-fly generated norm-conserving pseudopotentials [53] with a cutoff energy equal to 800 eV were used for all investigated compounds. The Brillouin zone (BZ) was sampled according to the Monkhorst-Pack scheme [54] with the 4 × 4 × 5 k-point mesh that provided separation of k-points equal to 0.05 Å −1 . Considering that the structural site of Ca ion was occupied by a mixture of Na, Li, Ca, La, Er and Yb ions with the total occupancy equal to 1, the mixed ion was simulated within the virtual crystal approximation [55]. The tolerance for groundstate wavefunctions founding was chosen to be 2 × 10 −6 eV. Pulay density mixing scheme was chosen for electronic minimization.

Particle Morphology and Phase Composition
The SEM patterns of the synthesized powder samples are shown in Figures 1 and S1 (see Supplementary Materials). In general, the particle morphologies are similar for all samples. The particles are partly coalescent due to the material diffusion between the grains, and this is typical of oxides subjected to a high temperature annealing [9,56,57]. In all SEM patterns, the biggest grains were ~5-7 μm in size. For comparison, in Figure 1, the SEM patterns are shown for pure NCLM:EY and LiNCLM:EY-0.3 with the highest Li content. In both samples, big grains were mixed with low-sized grains. However, in Figure  1b, the presence of partly faceted big grains is evident and the quantity of low-sized particles was lower. Thus, the presence of Li in the intermediate gels stimulated the grain growth and faceting. The XRD patterns recorded for the samples are presented in Figures 2 and S2. All peaks were well indexed by the tetragonal cell (space group I41/a) with the cell parameters close to CaMoO4 (COD 9009632) [58]. There were no alien peaks detected in the patterns, and it meant that all samples were in the pure phase state. Therefore, the CaMoO4 crystal structure was taken as a starting model for the Rietveld refinement. The Ca ion site was considered as that occupied by Na, Li, Ca, La, Er and Yb ions ( Figure 3) with fixed occupancies according to nominal compositions. The refinement was stable and gives low Rfactors ( Table 2, Figures 2 and S2). The coordinates of atoms and main bond lengths are listed in Tables S1 and S2, respectively. The linear increase in cell volume per average ion radii IR (Na/Li/Ca/La/Er/Yb) proves the suggested chemical compositions (Figure 2c). The XRD patterns recorded for the samples are presented in Figures 2 and S2. All peaks were well indexed by the tetragonal cell (space group I4 1 /a) with the cell parameters close to CaMoO 4 (COD 9009632) [58]. There were no alien peaks detected in the patterns, and it meant that all samples were in the pure phase state. Therefore, the CaMoO 4 crystal structure was taken as a starting model for the Rietveld refinement. The Ca ion site was considered as that occupied by Na, Li, Ca, La, Er and Yb ions (Figure 3) with fixed occupancies according to nominal compositions. The refinement was stable and gives low R-factors (Table 2, Figures 2 and S2). The coordinates of atoms and main bond lengths are      Further details of the crystal structures may be obtained from Fachinformationszentrum Karlsruhe, 76344 Eggenstein-Leopoldshafen, Germany (fax: (+49)7247-808-666; E-mail: crystdata@fiz-karlsruhe.de; http://www.fiz-karlsruhe.de/request_for_deposited_data.html, accessed on 5 February 2023) on quoting the deposition numbers: CSD 2117700-2117704.
The dependence of the unit cell volume on the average ion radius of the cations mixed in the Ca 2+ site of the Li x Na 1−x CaLa 0.5 Er 0.05 Yb 0. 45 Figure 4. For comparison, the cell volume values in CaMoO 4 [58], Na 0.5 La 0.5 MoO 4 [59] and Na 1/3 Ca 1/3 La (1−x−y)/3 Er x/3 Yb y/3 MoO 4 [45] scheelites are also shown. As can be seen, the previously obtained points of CaMoO 4 , Na 1/3 Ca 1/3 La (1−x−y)/3 Er x/3 Yb y/3 MoO 4 and Na 0.5 La 0.5 MoO 4 were well fitted by the linear function general for the scheelite-type structures. Additionally, the points related to the Na 1/3 Ca 1/3 La 0.5/3 Er 0.05/3 Yb 0.45/3 MoO 4 composition, as obtained in [45] and in this study, practically coincided, and that verifies high reproducibility of sol-gel microwave synthesis in the application to complex molybdates. However, the points related to Li-containing compositions Li x/3 Na (1−x)/3 CaLa 0.5/3 Er 0.05/3 Yb 0.45/3 MoO 4 form a new branch going strongly away from the straight line general for Li-free scheelites.

(MoO 4 ) 3 solid solutions is shown in
For the first time, this structural effect was detected and considered in detail in [39], and it was attributed to a very small radius of the Li + ion in reference to the radius of Na + ion. Thus, in the present work, this unusual structural effect was observed in one more molybdate system. ously obtained points of CaMoO4, Na1/3Ca1/3La(1-x-y)/3Erx/3Yby/3MoO4 and Na0.5La0.5MoO4 were well fitted by the linear function general for the scheelite-type structures. Additionally, the points related to the Na1/3Ca1/3La0.5/3Er0.05/3Yb0.45/3MoO4 composition, as obtained in [45] and in this study, practically coincided, and that verifies high reproducibility of solgel microwave synthesis in the application to complex molybdates. However, the points related to Li-containing compositions Lix/3Na(1-x)/3CaLa0.5/3Er0.05/3Yb0.45/3MoO4 form a new branch going strongly away from the straight line general for Li-free scheelites. For the first time, this structural effect was detected and considered in detail in [39], and it was attributed to a very small radius of the Li + ion in reference to the radius of Na + ion. Thus, in the present work, this unusual structural effect was observed in one more molybdate system.

Upconversion Properties
The UC emission spectra of Na 1−x Li x CaLa(MoO 4 ) 3 :Er,Yb (x = 0, 0,05, 0.1, 0.2 and 0.3) phosphors excited at 980 nm at room temperature are shown in Figure 5. Under the excitation at 980 nm, Yb 3+ ions were excited to the 2 F 5/2 state. A pair of excited Yb 3+ ions transferred their excitation to a single neighboring Er 3+ ion, thus exciting that Er 3+ ion to the 2 I 11/2 state. Furthermore, this excitation can be distributed to lower-lying 4 S 3/2 and 4 F 7/2 states. As a result, the upconversion emission of the samples under study exhibited the emission composed of two strong green bands and one weaker red band of the Er 3+ ion, namely the 2 I 11/2 → 4 I 15/2 , and 4 S 3/2 → 4 I 15/2 bands in green and 4 F 7/2 → 4 I 15/2 band in red. The variation of peak intensities of green and red bands, as well as the variation of integral intensity of UC luminescence over the whole spectrum, showed similar behavior on the variation of Li/Na ratio, as seen in Figure 6, all of these dependencies maximizing at comparatively low Li content x = 0.05. The Li + ion incorporation into the lattice of hosts instead of larger ions was efficient for manipulating the crystal field affecting the ions positioned within the second coordination sphere of the Li + ion or even onto more distant ions. This crystal field manipulation had led to the variation of luminescence of doping ions positioned in the second coordination sphere of the Li + ion (see, e.g., [39]. It is highly likely that Li + ions did not severely alter the local symmetry of the rare-earth ions in the lattice, but they enabled useful crystal field variations. According to Figure 2c, the cell volume of the solid solution under study monotonically decreases upon the Li + content. Thus, in the first approximation, we can expect the monotonical increase in crystal field strengths on Er 3+ and Yb 3+ ions. This increase can affect either oscillator strengths in the radiative recombination channels or the efficiency of upconversion excitation process. Hence, there are two ways to optimize the upconversion luminescence, namely (1) to control the full probability of energy transfer from Yb ions to Er ions in the excitation channel or (2) to control the oscillator strengths in the emission channel in order to increase radiative recombination rate against non-radiative one. The similarity of the concentration dependence of upconversion luminescence ( Figure 6) shows that the effect of Li + incorporation acts through the efficiency of excitation channel, while changes in the oscillator strengths at luminescence channels lead to a practically unobservable effect.
ions. This increase can affect either oscillator strengths in the radiative recombination channels or the efficiency of upconversion excitation process. Hence, there are two ways to optimize the upconversion luminescence, namely (1) to control the full probability of energy transfer from Yb ions to Er ions in the excitation channel or (2) to control the oscillator strengths in the emission channel in order to increase radiative recombination rate against non-radiative one. The similarity of the concentration dependence of upconversion luminescence ( Figure 6) shows that the effect of Li + incorporation acts through the efficiency of excitation channel, while changes in the oscillator strengths at luminescence channels lead to a practically unobservable effect.

Raman Spectroscopy
The spectral signals recorded from Na 1−x Li x CaLa(MoO 4 ) 3 :Er,Yb (x = 0, 0,05, 0.1, 0.2 and 0.3) samples under the excitation of 514.5 nm in the range of Raman-active vibrations are shown in Figure 7. In scheelite-type molybdates, the Raman bands are typically located from 200 to 900 cm −1 and they correspond to rotation (around 200 cm −1 ), bending (from 300 to 450 cm −1 ) and stretching (700-900 cm −1 ) vibrations of MoO 4 tetrahedra [3,12,14,18,22,36,38,39,45]. At the same time, the most intense Raman band of scheelite-type molybdates is associated with the symmetrical stretching of [MoO 4 ] 2− ions, and this band is commonly located around 880 cm −1 . As can be seen in Figure 7, only a very weak peak was observed in this spectral range in the Raman spectra of Na 1−x Li x CaLa(MoO 4 ) 3 :Er,Yb (x = 0, 0,05, 0.1, 0.2 and 0.3) compounds. Such behavior is associated with the fact that the Raman signal is almost totally superimposed by the 2 H 11/2 and 4 S 3/2 luminescence lines of Er 3+ ions. According to the spectra shown in Figure 7, the variation of peak intensities of 2 H 11/2 bands was different from the UC intensity dependence presented in Figures 5 and 6.

Calculation of Band Structure
The primitive cell of scheelite-type LixNa1-xCaLa0.5Er0.05Yb0.45(MoO pounds is shown in Figure 8a    The dominant part of the DFT simulations of scheelite-type structures is related to CaMoO 4 [60][61][62]. Additionally, the calculation of the band structure for CaMoO 4 doped with one ion can be found in several contributions [62,63]. In this work, we implemented the electronic band structure simulation of Li x Na 1−x CaLa 0. 5 Table 3. As is evident, the band gap energy only slightly increased on the Li content increase. An exception was observed for the LiNCLM:EY-0.2 sample. As for this fluctuation, it was revealed that the specific band gap energy value was induced by a slight variation of the O atom coordinates, as determined by the Rietveld refinement.

Calculation of Band Structure
The primitive cell of scheelite-type LixNa1-xCaLa0.5Er0.05Yb0.45(MoO4)3 (x = 0-0.3) compounds is shown in Figure 8a and the corresponding Brillouin zone is presented in Figure  8b  The dominant part of the DFT simulations of scheelite-type structures is related to CaMoO4 [60][61][62]. Additionally, the calculation of the band structure for CaMoO4 doped with one ion can be found in several contributions [62,63]. In this work, we implemented the electronic band structure simulation of LixNa1-xCaLa0.5Er0.05Yb0.45(MoO4)3 (x = 0-0.3) complex compounds using VCA (virtual crystal approximation). The result of the calculation for NCLM:EY is presented in Figure 9a. The band gap value for the direct (optical) electronic transition was 4.0 eV. It should be noted that the obtained values for direct and indirect band gaps differed in the second decimal place only. According to Figure 9b, the  Table 3. As is evident, the band gap energy only slightly increased on the Li content increase. An exception was observed for the LiNCLM:EY-0.2 sample. As for this fluctuation, it was revealed that the specific band gap energy value was induced by a slight variation of the O atom coordinates, as determined by the Rietveld refinement.

Conclusions
In the present study, the microwave sol-gel method was successfully employed for the preparation of the Li x Na 1−x CaLa 0.5 Er 0.05 Yb 0.45 (MoO 4 ) 3 , x = 0-0.3, solid solutions. The compounds were crystallized in tetragonal space group I4 1 /a, and they are typical representatives of the scheelite crystal family. However, the specific behavior of the structural parameters was observed as a function of the Li content. As the similar effect was previously observed in several related solid solutions with the partial substitution of Li for Na, it can be concluded that an unusual variation of the structural parameters is a general feature of complex Li-containing scheelite-type crystals. Evidently, it could be interesting to search for this effect in other complex scheelites containing big alkaline ions instead of Na.
The upconversion luminescence intensity of Li x Na 1−x CaLa 0. 5 3 , which exhibited its non-monotonic variation of the UC intensity upon the Li content, the material under study was featured by a single pronounced maximum of UCL at the Li content x = 0.05. The mechanism that controls the UC intensity was shown to be the variation of probability of energy transfer from Yb ions to Er ions, while the control of the emission channels played an insignificant role. Consequently, these Li-containing phosphors doped with Er 3+ /Yb 3+ lead to higher UC emitting efficiency with superior thermal and chemical stabilities overcoming the current limitations of traditional UC materials. It is emphasized that these new UC phosphors can be considered as potentially active optical switching and solar cell devices in practical optoelectronic fields.