Facile synthesis and emission enhancement in NaLuF4 upconversion nano/micro-crystals via Y3+ doping

A series of Y3+-absent/doped NaLuF4:Yb3+, Tm3+ nano/micro-crystals were prepared via a hydrothermal process with the assistance of citric acid. Cubic nanospheres, hexagonal microdisks, and hexagonal microprisms can be achieved by simply adjusting the reaction temperature. The effect of Y3+ doping on the morphology and upconversion (UC) emission of the as-prepared samples were systematically investigated. Compared to their Y3+-free counterpart, the integrated spectral intensities in the range of 445–495 nm from α-, β-, and α/β-mixed NaLuF4:Yb3+, Tm3+ crystals with 40 mol% Y3+ doping are increased by 9.7, 4.4, and 24.3 times, respectively; red UC luminescence intensities in the range of 630–725 nm are enhanced by 4.6, 2.4, and 24.9 times, respectively. It is proposed that the increased UC emission intensity is mainly ascribed to the deformation of crystal lattice, due to the electron cloud distortion in host lattice after Y3+ doping. This paper provides a facile route to achieve nano/micro-structures with intense UC luminescence, which may have potential applications in optoelectronic devices.

Optical upconversion (UC) is an anti-Stokes process that two or more low-energy photons can be converted into a single high-energy photon 1 . Rare-earth (RE) doped UC materials show many advantages, including high photochemical stability, low toxicity and long luminescence lifetimes [2][3][4][5][6] , which may have great potential applications in fields such as biological imaging, multi-dimensional displays, optical temperature sensors and solar cells [7][8][9][10] . However, compared to downconversion materials, the main shortcoming of UC materials is their low luminescence efficiency. Thus, an effective strategy to enhance the UC luminescence intensity is urgently needed. In recent years, many kinds of methods have been used to achieve efficient UC luminescence. For instance, Zhao et al. reported the enhanced red UC emission in Mn 2+ doped NaYF 4 : Yb/Er nanoparticles, due to the efficient energy transfer between Er 3+ and Mn 2+ 11 . Tan et al. demonstrated NaYbF 4 :Tm 3+ and NaYbF 4 :Er 3+ nanocrystals with the enhanced red UC luminescence, which is attributed to the cross relaxation effect among the activators at high activator content 12 . As is known, the UC emission of RE doped materials is remarkably affected by the crystal field symmetry around activators 13 , and the asymmetric environment of activators can result in the emission enhancement. For instance, Zhao's group reported Li + doped GdF 3 :Yb 3+ , Er 3+ nanocrystals with the enhanced red UC luminescence, which was caused by the decrease of local crystal field symmetry around activators after Li + doping 14 . Rai et al. demonstrated the enhanced green UC emission in Li + doped Y 2 O 3 :Yb 3+ /Er 3+ nanocrystals 15 . Yin et al. reported Mo 3+ doped NaYF 4 : Yb/Er nanocrystals with 6 and 8 times enhancement of green and red UC emissions, due to the lattice distortion after Mo 3+ doping 16 . In order to obtain efficient UC emission, the selection of excellent host material is essential. With the similar crystalline plane, NaYF 4 and NaLuF 4 have been considered as the outstanding host matrix for UC processes, due to their high thermal stability, low phonon energy and high refractive index [17][18][19][20][21] . As is known, the ionic radius of Y 3+ (0.89 Å) is larger than that of Lu 3+ (0.85 Å), thus Y 3+ doping may cause the expansion of NaLuF 4 host lattice, leading to the distortion of local symmetry around activators. Consequently, Y 3+ doping is an effective approach for enhancing the UC emission intensity in NaLuF 4 -based system. In addition, due to the small difference in ionic radius between Y 3+ and Lu 3+ , the phase transformation does not occur during introducing Y 3+ in NaLuF 4 crystals, which would be favorable to maintain the stability of crystal structure. However, there is no report on the increase of UC luminescence intensity in NaLuF 4 -based system via Y 3+ doping.
In this paper, in order to obtain different structures of NaLuF 4 nano/micro-crystals before Y 3+ doping, the influence of reaction temperature on the phase of Y 3+ -absent NaLuF 4 crystals is studied. It is found that cubic nanospheres, hexagonal microdisks and hexagonal microprisms can be achieved with the higher temperature. α-, β-, and α/β-mixed NaLuF 4 :Yb 3+ , Tm 3+ crystals with Y 3+ doping show the significant enhancement of UC emissions relative to Y 3+ -absent samples under 980 nm excitation at room temperature. The proposed mechanisms of UC emission enhancement and shape evolution through introducing Y 3+ are presented.

Results and Discussion
Phase and morphology. First, in order to obtain diverse structures of NaLuF 4 nano/micro-crystals before Y 3+ doping, the influence of reaction temperature on the crystal structure of Y 3+ -absent NaLuF 4 crystals is studied. The XRD patterns and the corresponding SEM images of Y 3+ -absent NaLuF 4 :Yb 3+ , Tm 3+ nano/micro-crystals prepared at different reaction temperatures for 12 h are displayed in Figs 1 and 2, respectively. As can be seen from Fig. 1, pure α-NaLuF 4 (JCPDS 27-0725) is formed at 110 °C. The related SEM image (Fig. 2a) shows that the sample is composed of a large number of small cubic nanospheres with an average diameter of 17 nm. At higher reaction temperature of 130 °C, α/β-mixed NaLuF 4 appears in the XRD pattern, indicating that the crystals partially change from α to β phase. Correspondingly, the SEM image of Fig. 2b exhibits two obvious particle morphologies containing small α-NaLuF 4 nanospheres and large β-NaLuF 4 microdisks with a mean diameter of 7.63 μm. After being treated at 150 °C, the corresponding XRD result demonstrates that pure β-NaLuF 4 (JCPDS 27-0726) can be obtained. The corresponding sample is composed of a large amount of hexagonal microdisks with regularity and smooth surfaces, and the small cubic nanoparticles completely disappear, as presented in From the above analysis, it can be concluded that higher reaction temperature favors the formation of NaLuF 4 crystals with hexagonal phase, which is ascribed to the fact that higher temperature favors the nucleation and the crystal growth 25 . The L/D ratio of β-NaLuF 4 microcrystals is enhanced as the temperature increases from 150 °C to 200 °C. As is known, β-NaLuF 4 has a high anisotropic structure 26 . The growth rate along [10ī0] direction is lower than that along [0001] direction at higher temperature due to Cit 3− absorbs onto the { 1 10 0} facets more strongly than the {0001} facets, thus results in the increase of L/D ratio and the shape evolution from disks to prisms.
In order to reveal the effect of Y 3+ doping on the morphology and UC emission of NaLuF 4 crystals, a series of Y 3+ doped α-, βand α/β-mixed NaLuF 4 :Yb 3+ , Tm 3+ nano/micro-crystals were synthesized. Figure 3(a and b) show the XRD patterns of α-NaLuF 4 :Yb 3+ , Tm 3+ nanocrystals and β-NaLuF 4 :Yb 3+ , Tm 3+ microcrystals introduced with different Y 3+ contents prepared at 110 °C and 200 °C for 12 h, respectively. As can be seen, pure cubic phase (Fig. 3a) and pure hexagonal phase (Fig. 3b) can be obtained even Y 3+ content increases up to 79 mol% (the Y 3+ -free samples have been shown in Fig. 1). No extra peaks can be observed, which indicates that Y 3+ doping has no influence on the crystal structure of cubic-phase nanocrystals and hexagonal-phase microcrystals. As demonstrated in the insets of Fig. 3(a and b), with the Y 3+ content increases from 0 to 79 mol%, the main diffraction peaks of α and β phases move to lower angles. According to Bragg's law 2d sinθ = nλ, where d represents the interplanar distance, θ represents the diffraction angle, and λ represents the diffraction wavelength. When Y 3+ doped into the lattice, Lu 3+ can be substituted by the relatively large Y 3+ , resulting in the expansion of NaLuF 4 host lattice (Fig. 3c), thus the interplanar distance increases and diffraction angle decreases. The values of the lattice constants and unit-cell volumes of α-NaLuF 4 :20%Yb 3+ , 1%Tm 3+ doped with different concentrations    Table 1, the higher unit-cell volumes are caused by the larger ionic radius of Y 3+ substituting Lu 3+ . Importantly, the lattice expansion may cause the distortion of local symmetry around Tm 3+ , which would break the forbidden transition of Tm 3+ , and consequently enhancing the UC emission intensity 27 . The above XRD results are well consistent with the corresponding SEM images.
As shown in Fig. 4(a-f), the Y 3+ doped α-NaLuF 4 nanoparticles are composed of a great deal of small cubic nanospheres (the Y 3+ -absent sample has been shown in Fig. 2a). The full width at half maximum (FWHM) was gradually narrowed with the Y 3+ concentration increases up to 79 mol%, as presented in Fig. 5. The average crystalline sizes can be calculated based on Scherrer's equation: D = 0.89λ/(βcosθ), where D is the crystallite size, λ represents the wavelength of the X-ray, β stands for the corrected half width of the diffraction peak, and θ is the diffraction angle. The factor 0.89 is the characteristic of a spherical particle. Thus, the mean diameters ( Table 2) of the spheres were calculated to be about 17 nm, 17 nm, 18 nm, 19 nm, 22 nm, and 24 nm, respectively. From the above results, it can be seen that the replacement of Lu 3+ by larger Y 3+ may lead to the increasing size of cubic-phase nanospheres.
The SEM images of Y 3+ doped β-NaLuF 4 microparticles are displayed in Fig. 6(a-f). As exhibited in Fig. 6a, the Y 3+ -free sample has been shown in Fig. 2e. As the Y 3+ concentration increases from 10 to 20 mol%, short hexagonal microprisms with regularity and uniformity are obtained, as presented in Fig. 6(b and c). On average, the prisms have a length of 3.01 μm and 4.81 μm; a diameter of 6.72 μm and 7.42 μm, respectively. When the Y 3+ concentration increases to 40 mol%, irregular hexagonal microprisms with coarse surfaces are shown in Fig. 6d. The average length of the prisms is 14.08 μm, and the average diameter is 11.02 μm. With the Y 3+ content further increases to 60 and 79 mol% [ Fig. 6(e and f)], the corresponding samples consist of hexagonal microprisms with scrappy ends and concave centers on the top/bottom surfaces. The prisms have a mean size of 7.78 μm and 7.71 μm in length; 5.98 μm and 5.10 μm in diameter, respectively. The L/D ratios are calculated to be about 0.45, 0.65, 1.28, 1.30, and 1.51 when the Y 3+ content is 10, 20, 40, 60, and 79 mol%. Thus, the L/D ratio of hexagonal microprisms is increased as the Y 3+ content increases from 10 to 79 mol%. Under our experimental condition, the chelated Lu 3+ -Cit 3− complex and Y 3+ -Cit 3− complex were formed. As is known, both β-NaLuF 4 and β-NaYF 4 have high anisotropic structures. From Fig. 6a (Lu 3+ = 79 mol%, Y 3+ = 0 mol%) and Fig. 6f (Lu 3+ = 0 mol%, Y 3+ = 79 mol%), it can be clearly seen that the L/D ratio of β-NaYF 4 is larger than that of β-NaLuF 4 . Thus, the v 1 / v 2 ratio of β-NaYF 4 is higher than that of β-NaLuF 4 under the same experimental conditions (v 1 is the growth rate along [0001] direction, v 2 is the growth rate along [ 1 10 0] direction), leading to the enhancement of L/D ratio and the morphology evolution from short hexagonal microprisms to long hexagonal microprisms when the Y 3+ concentration increases from 10 to 79 mol%. According to Liu et al. 's report about the density functional theory calculation on Gd 3+ doped NaYF 4 :Yb 3+ , Er 3+ nanoparticles, the electron charge density in host lattice changes after Y 3+ is substituted by Gd 3+ in the crystal lattice 28 . Under our synthesis conditions, the replacement of Lu 3+ by larger Y 3+ is similar to the substitution of Y 3+ by larger Gd 3+ . Thus, it is creditable that Y 3+ doped into NaLuF 4 host lattice may change the electron charge density, leading to the electron cloud distortion in crystal lattice, which would cause the deformation of crystal lattice. The change in crystal lattice may result in the formation of irregular and distorted hexagonal microprisms with coarse surfaces when the Y 3+ content is 40 mol%. Figure 7 shows the XRD patterns (a) and the main diffraction peak (b) of different Y 3+ doped α/β-mixed NaLuF 4 :Yb 3+ , Tm 3+ nano/micro-crystals prepared at 130 °C for 12 h. As shown in Fig. 7a, all samples are composed of a mixture of cubic and hexagonal phases (the Y 3+ -free sample has been shown in Fig. 1). Figure 7b displays the main diffraction peak of cubic phase shifts towards lower angles as the Y 3+ content increases from 0 to 79 mol%, which is mainly attributed to the expansion of crystal lattice after Lu 3+ is replaced by the relatively large Y 3+ . The shifting peak reveals that Y 3+ can be doped into the host lattice. The corresponding SEM images [ Fig. 8(a-f)] present two distinct particle morphologies including large microdisks (hexagonal phase) and small nanoparticles (cubic phase). It can be obviously seen that numerous spherical nanoparticles are attached on the surfaces of microdisks. The corresponding diameters of the disks are 7.63 μm, 5.64 μm, 4.79 μm, 3.50 μm, 2.66 μm, and 2.33 μm, respectively. The reduced diameter of the disks can be ascribed to the fact that β-NaYF 4 has higher v 1 /v 2 ratio than β-NaLuF 4 under the same experimental conditions.
The above results demonstrate that reaction temperature has a significant effect on the crystal structure of the products, and Y 3+ doping may cause the size-tuning and shape evolution of the crystals. Figure 9 summarizes the formation processes of Y 3+ -absent/doped NaLuF 4 :Yb 3+ , Tm 3+ nano/micro-crystals synthesized under different experimental conditions. UC photoluminescence properties. Figure 10(a-c)  in Fig. 11. For 450 nm emission, the Tm 3+1 D 2 level is populated by the ET1+ET2+CR processes (ET = energy transfer, CR = cross relaxation). For 477 nm and 649 nm emissions, the Tm 3+ 1 G 4 level is populated by the ET1+ET2+ET3 processes. For 696 nm emission, the Tm 3+3 F 3 level is populated by the ET1+ET2 processes. As can be seen from Fig. 10(a-c), the blue and red UC emission intensities are distinctly enhanced as the Y 3+ content increases from 0 to 40 mol%, and then declined at the content of 40-79 mol%. Thus, the strongest UC luminescence intensities are observed in the samples with 40 mol% Y 3+ doping. Compared to their Y 3+ -free samples, the integrated spectral intensities in the range of 445-495 nm from α-, β-, and α/β-mixed      30 , when the Li + content is below 7 mol%, Li + substitutes Na + , causing the shrinking of host lattice; however, as the Li + content increases from 7 to 15 mol%, Li + begins to occupy interstitial site, leading to  the expansion of crystal lattice; thus the sample with 7 mol% Li + doping has the highest UC emission intensity, owing to the lowest crystal field symmetry around activators. Besides, Y 3+ doping causes the electron cloud distortion in host lattice, resulting in the tunable size of the as-prepared samples. As is known, as for larger-size crystals, the nonradiative energy transfer processes of Tm 3+ would decrease due to their fewer surface quenching  sites 28 , which is in favor of UC emission. Thus, as for Y 3+ doped β-NaLuF 4 :20%Yb 3+ , 1%Tm 3+ microcrystals, the larger-size (relative to Y 3+ -absent samples) of the samples with 40 mol% Y 3+ doping may have a small contribution to the enhancement of UC luminescence intensity. Figure 12 presents the decay curves of (a) 1 G 4 → 3 H 6 and (b) 1 G 4 → 3 F 4 transitions of Tm 3+ in α-NaLuF 4 :20%Yb 3+ , 1%Tm 3+ nanocrystals doped with 0, 40 and 79 mol% Y 3+ . Based on the function: τ = ∫I(t) dt/I max , where I(t) represents the emission intensity at time t, and I max represents the peak intensity in the decay curve. The calculation results (Table 3)   is equal to the sum (A r+nr = A r + A nr ) of radiative (A r ) and nonradiative (A nr ) transition probability. Thus, the lowest luminescence lifetime in the sample with 40 mol% Y 3+ doping is mainly caused by the maximum emission intensity.

Conclusion
In summary, cubic nanospheres, hexagonal microdisks, and hexagonal microprisms can be achieved by simply adjusting the reaction temperature. It is found that higher temperature favors the nucleation and the crystal growth. The effect of Y 3+ doping on the morphology and UC emission of the as-prepared samples were systematically investigated. The results demonstrate that Y 3+ doping may cause the size-tuning and shape evolution of the crystals. Compared to their Y 3+ -free samples, the integrated spectral intensities in the range of 445-495 nm from α-, β-, and α/β-mixed NaLuF 4 :20%Yb 3+ , 1%Tm 3+ crystals with 40 mol% Y 3+ doping are increased by 9.7, 4.4, and 24.3 times, respectively; red UC luminescence intensities in the range of 630-725 nm are enhanced by 4.6, 2.4, and 24.9 times, respectively. It is proposed that the increased UC emission intensity is mainly ascribed to the deformation of crystal lattice, due to the electron cloud distortion in host lattice after Y 3+ doping. Besides, as for Y 3+ doped β-NaLuF 4 :20%Yb 3+ , 1%Tm 3+ microcrystals, the larger-size (relative to Y 3+ -absent samples) of the samples with 40 mol% Y 3+ doping may have a small contribution to the enhancement of UC luminescence intensity. As a result of their intense UC emission, these phosphors may be suitable for optoelectronic devices. Preparation. All samples were prepared based on our previously reported procedures [22][23][24] . As for the synthesis of Y 3+ -absent α-NaLuF 4 :20%Yb 3+ , 1%Tm 3+ nanocrystals, 3 mmol of citric acid (2 M, 1.5 mL), 5 mmol of NaOH (4 M, 1.25 mL) and 10 mL of deionized water were mixed and stirred for 10 min. Then 1 mmol of RE(NO 3 ) 3 (0.79 mmol of Lu(NO 3 ) 3 (1M, 0.79 mL), 0.2 mmol of Yb(NO 3 ) 3 (0.5 M, 0.4 mL), and 0.01 mmol of Tm(NO 3 ) 3 (0.1 M, 0.1 mL)) were added to above mixture and then stirred for 30 min to form the RE-Cit 3− complex. Subsequently, 16 mL of aqueous solution containing 9 mmol of NaF (1 M, 9 mL) and 7 mL of deionized water were added into the chelated RE-Cit 3− complex to form a colloidal suspension and kept stirring for another 30 min. Finally, the suspension was transferred into a 50 ml-Teflon vessel, sealed in autoclave and maintained at 110 °C for 12 h. After the autoclave was cooled to room temperature naturally, the final products separated by centrifugation, washed with ethanol and deionized water several times, and then dried in air at 60 °C for 12 h. Other samples were prepared by a similar process only by tuning the reaction temperature (110-200 °C) and Y 3+ content (0-79 mol%).  Characterization. The crystal structure of the as-prepared samples was confirmed by powder X-ray diffraction (XRD) patterns using the D-Max 2200VPC XRD from Rigaku Company (Cu-Kα radiation, λ = 1.5418 Å). The morphology was observed by Oxford Quanta 400 F Thermal Field Emission environmental Scanning Electronic Microscope (SEM). UC photoluminescence spectra were carried out on an Edinburgh Instrument Company FLS980 combined fluorescence lifetime and steady-state fluorescence spectrometer equipped with a 1 W 980 nm laser diode.