Low-temperature molten-salt synthesis and upconversion of novel hexagonal NaBiF₄:Er³⁺/Yb³⁺ micro-/nanocrystals

Abstract

A series of NaBiF₄:Er³⁺/Yb³⁺ micro-/nanocrystals were synthesized via the low-temperature molten-salt method in NH₄NO₃ flux. The influences of the amount of NaF and NH₄NO₃ flux used, the reaction time and reaction temperature on the crystal phase structure, morphology and size of the resultant samples were systematically investigated. The morphology, size, doped concentrations of Er³⁺ and Yb³⁺, pumping power, and temperature dependences of UC emissions were studied in detail under 980 nm irradiation. The NaBiF₄:3 mol% Er³⁺/20 mol% Yb³⁺ crystal exhibits the strongest green UC emission, which was observed under several mW 980 nm excitation at room temperature. The energy transfer between Bi³⁺ and Er³⁺ during UC emission process was also discussed at length.

---

1. Introduction

In the past decades, lanthanide-doped upconversion (UC) inorganic materials have evoked a tremendous amount of attention to the ongoing research. UC refers to a system sequentially absorbing two or more long-wavelength (low-energy) photons, combining their energies, and emitting one short-wavelength (high-energy) photon. Compared to semiconductor quantum dots, these UC materials exhibit some advantages of low toxicity, large Stokes shift, sharp emission band, high quantum yield, long lifetime, weak background luminescence, and good resistance to photo-bleaching, blinking and photochemical degradation.^1–3 Owing to their outstanding properties, they have been proposed for diverse potential applications in photovoltaic,^3–5 photocatalysis,^4,5 sensor and biological labels.^6–9

Er³⁺, as being an excellent activated ion, can emit different wavelength photons ranging from ultraviolet (UV) through visible (red, blue, green) to near infrared (NIR) light. Unfortunately, Er³⁺ suffers from inferior absorption near 980 nm (compared with Yb³⁺), and low Er³⁺ doping concentration (in the range 0.2–2%) to avoid significant concentration-quenching.^5–9 To enhance UC efficiency, a popular approach is adopted where a sensitizer with a reasonable cross section in NIR region is co-doped due to an efficient energy transfer between the two. Yb³⁺ has a relatively large absorption cross section (1.2 × 10⁻²⁰ cm²) compared with that of Er³⁺ (1.7 × 10⁻²¹ cm²) and broad absorption band matching with the emission wavelength of powerful commercial laser diodes, and its ²F_5/2 excited state resonates well with the ⁴I_11/2 excited state of Er³⁺, thus Yb³⁺ can significantly improve UC efficiency of Er³⁺.^10–13

It is generally accepted that Bi³⁺ ions possess some inherent features of non-toxicity, low cost, and large rare-earth ion admittance. From the point of optical physics, Bi³⁺ ions have 6s² electronic configuration (¹S₀ corresponds to the ground state, whereas ³P₀, ³P₁ and ³P₂ are due to three excited states) and a broad absorption and emission band,^14–16 which is usually used in downconversion and UC emission system to enhance luminescence.^16–21 For example, the Bi³⁺ ion in hexagonal NaYF₄ possesses a broad absorption near 300 nm and a emission band ranging from 400 to 500 nm.²² The introduction of the optimal Bi³⁺ can obviously enhance UC emission of hexagonal NaYF₄:Er³⁺/Yb³⁺ materials.²¹ In addition, fluoride compounds possess a host lattice of low phonon energies (less than 350 cm⁻¹) causing to decrease nonradiative relaxation and in turn to yield relatively strong UC emission, they have served as very efficient UC matrixes. For instance, NaREF₄ (RE = Y, Gd and Lu) is one kind of the efficient NIR-vis UC host materials.^23–27 Therefore, Bi³⁺-based fluorides can combine the advantages of both fluorides and Bi³⁺ ions, they are suggested to be novel UC materials with efficient luminescence and have been attracted more and more attention.^28–31 As one member of Bi³⁺-based fluorides, NaBiF₄ belongs to the hexagonal system with the space group P3̄ and the unit cell parameters: a = 6.144 Å, c = 3.721 Å (ref. 32) and has the same structure as NaREF₄,^1–3 thereby, the addition of Er³⁺ and Yb³⁺ can not generate the crystal lattice distortion and exhibits the capacity of heavily rare-earth ion doping. The Yb³⁺/RE³⁺ (RE = Er, Tm, and Ho) doped hexagonal NaBiF₄ nanocrystals has demonstrated to be one potential excellent UC material.³¹

In past several decades, rare-earth doped NaREF₄ materials with high dispersability were synthesized in organic solvents by hydro(solvo)thermal methods.^{1–11,33–35} In contrast with the above methods, the most advantage of low-temperature molten-salt method is an economical mass, convenient, effective, facile and environmental friendly approach. And it is very easy to obtain the materials with clean surface, chemical purification, and few residual impurities below the melting point of the as-grown crystals. Ammonium nitrate (NH₄NO₃) exhibits some inherent advantages, including abundance in nature, low melting point, and easy deliquescence in water/alcohol causing to easily separate from the flux and the target compound. Thus, this synthetic route has been has been adopted to fabricate hexagonal NaYF₄:Er³⁺/Yb³⁺,³⁶ tetragonal LiYF₄:Er³⁺/Yb³⁺,³⁷ cubic BaGdF₅:Ce³⁺/Er³⁺/Yb³⁺ (ref. 38) and monazite LaPO₄:Eu³⁺.³⁹ Recently, the Yb³⁺/RE³⁺ (RE = Er, Tm, and Ho) doped hexagonal NaBiF₄ nanocrystals have been fabricated in ethylene glycol at room temperature.³¹ Unfortunately, the degree of the crystalline and their UC emissions need to be further improved. Therefore, here, hexagonal NaBiF₄:Er³⁺/Yb³⁺ (h-NaBiF₄:Er³⁺/Yb³⁺) micro-/nanocrystals will be prepared by the low-temperature molten-salt synthesis in NH₄NO₃ flux. In this synthetic route, the phase structure, shape and size of the samples could be highly determined by the intrinsic structure of target compound and the growth surroundings (such as the reaction temperature, the reaction time, and the usage amounts of raw materials and flux). To obtain high-quality NaBiF₄:Er³⁺/Yb³⁺ micro-/nano-crystals, it is very necessary to fastidiously control these parameters. Herein, we will mainly concentrate upon the influence of the usage amount of reaction materials NaF and flux NH₄NO₃, the reaction temperature and the reaction time on the chemical composition, the morphology and size of the resultant products. On this basis, the corresponding UC properties will be investigated and discussed in detail. Finally, the UC emission dynamics of h-NaBiF₄:Er³⁺/Yb³⁺ will be studied by the pumping power, and temperature dependences of UC emissions and fluorescence decays.

2. Experimental

2.1 Material preparation

NaBiF₄:Er³⁺/Yb³⁺ samples were synthesized by the low temperature molten-salt method in NH₄NO₃ flux. Reagent-grade NaF, NH₄NO₃, Bi(NO₃)₃·5H₂O, Yb(NO₃)₃·5H₂O and Er(NO₃)₃·5H₂O powders were used to synthesize hexagonal Er³⁺/Yb³⁺-codoped NaBiF₄ crystals. All the reagent-grade powders and NH₄NO₃ with different usage amounts were mixed well. The mixtures were put into 15 ml capacity crucibles. After the lids were tightly closed, the crucibles were placed in an oven. The sealed tank was heated to different reaction temperature (160, 175, 210 and 250 °C), and held for different hours (0.5, 1, 2, 4, 8, 12 and 24 h) in an oven, and then cooled to room temperature naturally. The detailed experimental parameters including the reaction temperature, reaction time, the doping concentration of Er³⁺ and Yb³⁺, the NaF usage amount and the usage amount of NH₄NO₃ flux were listed in Table 1. After being washed several times with deionized water and ethanol, the precipitates were dried at 70 °C for 12 hours in vacuum.

Table 1 The detailed experimental parameters and crystal phase of the as-obtained samples

Sample	NaF/Ln molar ratio	NH4NO3/Ln molar ratio	Reaction temperature (°C)	Reaction time (h)	Crystal phase
S1	3	5	250	24	c-BiF3 + t-BiOF
S2	4	5	250	24
S3	5	5	250	24	h-NaBiF4
S4	6	5	250	24	h-NaBiF4
S5	8	5	250	24	h-NaBiF4
S6	5	8	250	24	h-NaBiF4
S7	5	10	250	24	h-NaBiF4
S9	6	8	250	24	h-NaBiF4
S10	6	10	250	24	h-NaBiF4
S11	8	8	250	24	h-NaBiF4
S12	8	10	250	24	h-NaBiF4
S13	5	5	210	24	h-NaBiF4
S14	5	5	180	24	h-NaBiF4 + c-BiF3 + t-BiOF
S15	5	5	160	24	m-NH4BiF4 + c-BiF3 + t-BiOF
S16	3	5	250	12	h-NaBiF4
S17	3	5	250	8	h-NaBiF4
S18	3	5	250	4	h-NaBiF4
S19	3	5	250	2	h-NaBiF4
S20	3	5	250	1	h-NaBiF4 t-BiOF
S21	3	5	250	0.5	h-NaBiF4 t-BiOF

---

2.2 Material characterization

The samples were examined by X-ray diffraction (XRD), transmission electron microscopy (TEM), energy-dispersive X-ray spectroscopy (EDS) and photoluminescence (PL). XRD analyses were carried out on a Bruker D8-Advance diffractometer with graphite-monochromatized Cu Kα radiation (40 kV/60 mA, graphite monochromator, λ = 0.1541 nm). The size, morphology and chemical compositions of products were determined by a transmission electron microscopy (TEM, JEOL2010) operating at 200 kV and a JSM 6700F scanning electron microscope (SEM) equipped with the energy dispersive X-ray spectrum (EDS). Structural information of the nanocrystals was measured by a high-resolution transmission electron microscopy (HRTEM).

Downconversion (DC) luminescence spectra were performed using an Edinburgh Instruments FLS920 fluorescence spectrometer equipped with both continuous (450 W) and pulsed xenon lamps. UC emission spectra were carried out upon 976 nm excitation with a pump power of ∼200 mW (power density ∼ 20 W cm⁻²), provided by a mode-locked picosecond Ti:sapphire laser (700–1000 nm, pulse width ≤ 1.5 ps, Tsunami, Spectra-Physics). The measurements were repeated in an integrating sphere in order to reduce the effect of the scattering on the UC properties. Power-dependence of UC emissions of Er³⁺ was studied upon excitation by a power-controllable 976 nm diode laser (DPL-II, Module HTL98M10) with two emission monochromators to record the emission spectra in the wavelength range of 200–850 nm (with a Hamamatsu R928 photomultiplier tube). Its maximum power output is 5 W. To eliminate the noise from tunable Ti:sapphire laser, 980 nm LD, and both continuous (450 W) and pulsed xenon lamps, different filters were kept in front of the detector. The emission monochromator's slits were set as small as possible to maximize the instrumental resolution. The best wavelength resolution was 0.05 nm. The line intensities and positions of the measured spectra were calibrated according to the FLS920 correction curve and standard mercury lamp. All the spectra were measured at room temperature and under identical conditions for each serial test. Luminescence decays were recorded using a customized UV to mid-infrared steady-state and phosphorescence lifetime spectrometer (FSP920-C, Edinburgh) equipped with a digital oscilloscope (TDS3052B, Tektronix) and a tunable mid-band OPO pulse laser as the excitation source (410–2400 nm, 10 Hz, pulse width ≤ 5 ns, Vibrant 355II, OPOTEK). For the low-temperature measurements, the samples were mounted on a closed-cycle cryostat (10–350 K, DE202, Advanced Research Systems).

3. Results and discussion

3.1 Phase structure

When the molar ratio of NaF to Ln(NO₃)₃ (hereinafter to be referred as NaF/Ln) was 3, the resultant product consists of cubic BiF₃ (c-BiF₃) (JCPDS no. 74-0144) and tetragonal BiOF (t-BiOF) (JCPDS no. 86-1478) (Fig. 1a) and the XRD peaks due to hexagonal NaBiF₄ (h-NaBiF₄) (JCPDS no. 041-0796) were not observed. When the NaF/Ln molar ratio increased to 4, h-NaBiF₄ phase appeared and played a predominant role (Fig. 1a), and yet a small quantity of c-BiF₃ could be still detected. As the NaF/Ln molar ratio increased to 5, the diffraction peaks due to c-BiF₃ disappeared and those due to t-BiOF continued to weaken. The further increasing NaF/Ln molar ratio to 8, the diffraction peaks due to the t-BiOF vanished, all the diffraction peaks were in good agreement with those of pure h-NaBiF₄ and no trace of characteristic peaks for other impurity phases was observed, implying that the samples obtained have the same crystallographic structure as pure h-NaBiF₄ crystal. Thus, the h-NaBiF₄ can be synthesized at NaF/Ln molar ratio above 5, and the NH₄NO₃ molten solution system can be developed to prepare the h-NaBiF₄ materials.

Fig. 1 The XRD patterns of the samples prepared under different conditions. (a) Different molar ratios of NaF/Ln; (b) different temperatures; (c) different times; (d) doped with 2 mol% Er³⁺ and different Yb³⁺ concentrations; (e) doped with 3 mol% Er³⁺ and different Yb³⁺ concentrations; (f) doped with 20 mol% Yb³⁺ and different Er³⁺ concentrations.

From Fig. 1b, it can be seen that the variance of the reaction temperature also highly affects the phase structure of the products. For the sample obtained at 210 and 250 °C for 12 hours, only h-NaBiF₄ (JCPDS no. 041-0796) was obtained. When the reaction temperature decreased to 175 °C, the t-BiOF and monoclinic NH₄BiF₄ (m-NH₄BiF₄) (JCPDS no. 041-0796) were formed. With the reaction temperature further decreasing to 160 °C, the m-NH₄BiF₄ became more and more and played a predominant role while the amount of h-NaBiF₄ significantly decreased. As is known, the thermodynamic growth regime is highly dependent upon a sufficient supply of thermal energy, and the crystal with the most stable structure preferential grows.⁴⁰ Since h-NaBiF₄ is thermodynamically more stable than m-NH₄BiF₄ and t-BiOF, m-NH₄BiF₄ and t-BiOF are susceptible to transform to h-NaBiF₄. The energy barrier hinders the formation of the h-NaBiF₄, a sufficient supply of thermal energy must be required to surmount this energy barrier to adjust the environment of Bi³⁺ (Er³⁺ and Yb³⁺) and Na⁺ occupation sites.⁴¹ Therefore in this thermal system m-NH₄BiF₄ and t-BiOF can be formed at low temperature, relatively high reaction temperature is more favorable for this transformation (m-NH₄BiF₄ → NaBiF₄ and t-BiOF → h-NaBiF₄) in the molten salt process and the pure h-NaBiF₄ crystals can be synthesized at the reaction temperature of 210 °C and above.

To study the impact of the reaction time on the phase structure and the growth mechanism, the XRD patterns of the products prepared for different reaction times of 0.5, 1, 2, 4, 8, 12 and 24 h, as shown in Fig. 1c, indicates that the crystal phase of the samples is strongly dependent on the reaction time. When the reaction time was 0.5 h, a great amount of h-NaBiF₄ and a small amount of t-BiOF were also observed and m-NH₄BiF₄ (JCPDS no. 041-0796) was hardly detected. The t-BiOF became less and less with the elongation of the reaction time from 0.5 hour to 4 hours and disappeared as the reaction time extended to 8 and 12 hours. It is demonstrated that the longer reaction time is also beneficial to transform from t-BiOF to h-NaBiF₄ and the pure h-NaBiF₄ crystals can be also obtained when the reaction time exceeds 8 h. Moreover, in comparison to (101) and (201) XRD peaks owing to he hexagonal NaBiF₄, the (110) and (110) XRD peaks becomes stronger with the extension of the reaction time, implying that the [0001] direction is the preferred direction of crystal growth.

XRD patterns of the products with different concentrations of Er³⁺ or Yb³⁺ for 24 hours are shown in Fig. 1d, e and f with h-NaBiF₄ standard card (no. 41-0796). All the XRD peaks are in good agreement with those of the h-NaBiF₄, showing that all the products have the same crystal structure as pure h-NaBiF₄. Because of the replacement of Bi³⁺ with smaller Er³⁺ and Yb³⁺ ions in the host lattice, the diffraction peaks shift slightly to higher angle side as a result of the shrinkage of the unit-cell volume.

3.2 Morphology

Fig. 2–5 illustrate some representative SEM images of h-NaBiF₄:Er³⁺/Yb³⁺ prepared under different reaction conditions, as listed in Table 1. When the molar ratio of NaF/Ln was 3.0, the corresponding sample is composed of a large number of tiny nanoparticles attached to the surface of large polyhedral particles with size of ca. 0.6 μm. The product prepared at the NaF/Ln molar ratio of 4 mainly consists of many prismatic microrods with 4–5 μm in length and ca. 0.8 μm in diameter. Their lateral surfaces (Fig. 2a and 3a) are covered with a layer of minute irregular nanoparticles while their end planes are coarse, concave and irregular. When the molar ratio of NaF/Ln was increased to 5.0, the product maintains rod-like morphology and becomes even uniform. Their sizes and end planes keep nearly unchanged, whereas their lateral surfaces become relatively smooth and have some fine slit along the rods. With the increasing NaF/Ln molar ratio to 6.0, the morphology of the product still exhibits prismatic microrods, whenas their length and diameter slightly decrease to 3–4 μm and ca. 0.8 μm, respectively. Some small holes appear on their surfaces. Further increasing NaF/Ln molar ratio to 8.0 deduces the prisms incomplete and their length and diameter are 3–5 μm and 0.6 μm, respectively. There are a great number of relatively large holes on their side surfaces of these prisms.

Fig. 2 Representative SEM images of the as-obtained products at 250 °C for 24 hours with the NH₄NO₃/Ln molar ratio of 5, and different NaF/Ln molar ratios of 3 (a), 4 (b), 5 (c), 6 (d) and 8 (e).

Fig. 3 Representative SEM images of the as-obtained products at 210 °C for 24 hours with the NaF/NH₄NO₃/Ln molar ratio of 5/8/1 (a), 5/10/1 (b), 6/8/1 (c), 6/10/1 (d), 8/8/1 (e) and 8/10/1 (f).

Fig. 4 Representative SEM images of the as-obtained products under for 24 hours with the NH₄F/Ln molar ratio of 5 and NH₄NO₃/Ln molar ratios of 8 at reaction temperature of 210 °C (a), 180 °C (b) and 160 °C (c).

Fig. 5 Representative SEM images of the as-obtained products under at 250 °C for 24 h with the NH₄F/Ln molar ratio of 5 and NH₄NO₃/Ln molar ratios of 8 for the reaction time of 0.5 h (a), 1 h (b), 2 h (c), 4 h (d), 8 h (e) and 12 h (f).

To study the impact of the variation of the amount of flux NH₄NO₃ upon the morphology and size of the products, the SEM images of the samples with different NH₄NO₃/Ln molar ratios were shown in Fig. 3. The NH₄NO₃ usage amount produces a great influence upon their sizes. When the NaF/Ln molar ratio was fixed to be 5 and 6, the molar ratio of NH₄NO₃/Ln increased to 8 and 10, the resultant products still maintain prism-like morphology. Meanwhile, their diameters also nearly stay at the same level (ca. 0.8 μm), and the mean lengths decrease to ca. 3.0 μm. For the NaF/Ln molar ratio of 8, the NH₄NO₃/Ln molar ratio enhances from 5 through 8 to 10, the diameters of these rods also keep almost invariable (ca. 0.5 μm), the lengths experience a significant downward from 3.0 μm through 2.0 μm to 1.0 μm. That is to say, the larger NH₄NO₃/Ln molar ratio is, the smaller the aspect ratio length/diameter becomes.

To elucidate the effect of the reaction temperature on the morphology the detailed temperature-dependent experiments were carried out under the similar reaction conditions. When the reaction temperature decreases from 250 °C to 210 °C, the corresponding sample takes on prismatic shape with the length of ca. 2.0–3.0 μm, the diameter of ca. 0.3–0.5 μm and irregular end planes (Fig. 4a). In comparison with S3, the length and diameter simultaneously have a downward trend. Further decreasing reaction temperature to 180 °C deduces the prismatic rods disappear, and some tiny nanoparticles aggregate together into the prismatic embryonic form (Fig. 4b). When the reaction temperature is 160 °C, the prismatic embryonic form vanish, a great number of tiny nanoparticles have come into being (Fig. 4c).

To shed light on the morphology evolution of h-NaBiF₄ crystal, taking S3 as an example, the morphologies of the intermediates at different growth stages are shown in Fig. 5 and 6. The intermediates exhibit different morphologies at different reaction times. The reaction treatment for 0.5 hour leads to the formation of small hexagonal prisms with average sizes of 10–50 nm in length and 20–50 nm in diameter (Fig. 5a and 6a). The high-magnification TEM image indicates that the fringe distance (d-spacing) is 0.311 nm (Fig. 6b), corresponding to the distance between neighboring (110) planes of h-NaBiF₄ (JCPDS no. 041-0796). The selected area electron diffraction (SAED) pattern, as shown in Fig. 6c, shows the diffraction rings corresponding to (110), (111), (210), (102) and (310) planes of h-NaBiF₄. The compositional analysis by EDS pattern spectrum (as shown in Fig. 6d) reveals that the sample is composed of Er, Yb, Bi, Na and F elements. When the reaction time was increased to 1.0 h, some short and imperfect hexagonal prismatic particles are formed (Fig. 5b). TEM image reveals that these prismatic products are actually polycrystalline, constructed by small particles (Fig. 6e). The holding time of 2.0 h reduced the these short and imperfect hexagonal prisms to evolve into the regular ones with the length of 200–500 nm and the diameter of 150–200 nm and convex end planes (Fig. 5c). When the reaction time was extended to 4 and 8, the hexagonal prisms continue to grow large while the small particles become less and less (Fig. 5d and e). As the reaction time elongated to 12 h, most of small nanoparticles disappeared and transformed to the hexagonal prismatic rods or plates (Fig. 5f). When the reaction time further extended to 24 h, these irregular nanoparticles depleted and converted to the hexagonal prismatic rods or plates (Fig. 2c). The above results reveal that the rod-like particles undergo dissolution–recrystallization transformation in order to minimize the surface energy of the system, and the h-NaBiF₄ prisms grow along c-axis direction at the consumption of smaller nanoparticles by Ostwald ripening process.⁴²

Fig. 6 High-magnification TEM image (a), high-resolution TEM image (b), SAED pattern (c); and EDS spectra (d) of as-prepared NaBiF₄:3 mol% Er³⁺/20 mol% Yb³⁺ sample S21 for 1.0 h, and high-magnification TEM image (e) of as-prepared NaBiF₄:3 mol% Er³⁺/20 mol% Yb³⁺ sample S20 for 1.0 h.

3.3 Formation mechanism

Herein we have experimentally and independently study the impacts of (1) the reaction time, (2) the reaction temperature, (3) the usage amount of NaF (4) the usage amount of NH₄NO₃ flux upon the resultant morphology and chemical composition. Based on the above results, a possible phase and morphology evolution mechanism is described as follows.

At the beginning, Ln(NO₃)₃·5H₂O (Ln = Bi, Yb and Er) and NaF are mixed with the desired amount of NH₄NO₃ salts to form a precursor, which then is reacted above the melting point to provide a fluid phase. Since the melting point of NH₄NO₃, Bi(NO₃)₃·5H₂O, Yb(NO₃)₃·5H₂O and Er(NO₃)₃·5H₂O is less than 100 °C, they quickly melt, dissociate and form Bi³⁺, Yb³⁺, Er³⁺, NH₄⁺ and NO₃⁻ ions during the heat-up process, whereas too high melting point of NaF (ca. 993 °C) cannot melt in such low experimental temperature, howbeit NaF can be partly dissolved to form Na⁺ and F⁻ ions due to the existence of crystalliferous water. In addition, some Bi³⁺ ions tend to experience a hydrolysis process and generate BiO⁺ ions during warming-up process.

When the usage amount of F⁻ is beyond the stoichiometric ratio of h-NaBiF₄, Na⁺, Ln³⁺ (Ln = Bi, Yb, Er) and NH₄⁺ ions in the molten salt interact with F⁻ to generate m-NH₄BiF₄ and h-NaBiF₄ small nuclei. If the number of F⁻ ions can not exceed this ratio, the m-NH₄BiF₄ and c-BiF₃ take precedence to produce in this molten-salt system. Additionally, those BiO⁺ ions react with F⁻ ions and produce t-BiOF small nuclei. In thermodynamics, m-NH₄BiF₄ and t-BiOF are lower than h-NaBiF₄ and c-BiF₃, and susceptible to evolve inevitably to h-NaBiF₄ and c-BiF₃ seed through dissolution–renucleation process. These nuclei quickly aggregate together and grown into t-BiOF and h-NaBiF₄ crystals. As the reaction temperature enhances and reaction time extends, the rate of transformation from m-NH₄BiF₄ and t-BiOF to h-NaBiF₄ becomes larger, more m-NH₄BiF₄ and t-BiOF convert to h-NaBiF₄ by dissolution–recrystallization process.

On the basis of the general theory of crystal growth, the growth crystal is relevant to the relative growth rate of different crystal faces. The different crystal rates give rise to diverse appearance of the crystalline. Since h-NaBiF₄ micro-/nano-crystals have a hexagonal shape, their surfaces are typical (001) top/bottom planes and six prismatic side planes of the energetically equivalent (100) family. In this molten salt system, the growth rate on the c-axis in the crystal growth process is quicker than that on the directions perpendicular to c-axis, thereby the sample takes on the hexagonal prism. At the same time, the selective adhesion of some simple ions such as NH₄⁺ ions and Na⁺ ions on some specific crystalline planes can change the epitaxial growth rate of different crystallographic direction.³⁸ In our experimental surrounding, NH₄⁺ or Na⁺ ions preferentially adsorb the (001) crystal plane and thereby remarkably prohibit the c-axis (i.e. [001]) directed growth, there is a tendency to shorten the length of prism. The observation of the concave structure at top/bottom facets suggests that the growth rate of the prismatic side facets is little faster than that of the top/bottom facets. These prism-like crystals exhibit the presence of some defects (such as holes and linear slits) on their side surfaces of these prisms, which may be reflected on the preferential 2D-nucleation at the border of the growing face. This behavior can be resulted form the morphological instability generated by the competition between the super-saturation (destabilizing factor) and the surface tension (stabilizing factor).^43,44

3.4 Upconversion emission

To study the effects of Er³⁺ and Yb³⁺ concentrations on the UC properties of h-NaBiF₄:Er³⁺/Yb³⁺ products, Fig. 7 depicts the room temperature UC emission spectra of h-NaBiF₄:Er³⁺/Yb³⁺ samples with different doping concentrations upon 980 nm excitation. The strong green UC emission band is observed in the region of 512–573 nm, attributed to the ²H_11/2/⁴S_3/2 → ⁴I_15/2 transitions of Er³⁺. The red UC emission between 630–700 nm belongs to ⁴F_9/2 → ⁴I_15/2 transition of Er³⁺. It is striking that the predominant green emission is much stronger than the red emission. Thus, it can emit intense green light with high purity. Besides green and red UC emission, other weak UC emissions at 370–390, 400–428, and 445–460 nm, stem from the ⁴G_11/2, ²H_9/2, ⁴F_3/2 and ⁴F_5/2 states to the ⁴I_15/2 ground state of Er³⁺,^9–12 respectively. The UC emission in the range between 485 and 505 nm is attributed to the overlap of the cooperative UC of Yb³⁺,^{45–47 4}F₇₂–⁴I_15/2 transition of Er³⁺ (ref. 9–12) and the ³P₁–¹S₀ transition of Bi³⁺ ion^14,15 (see Fig. 7a). It is clearly seen from Fig. 7a and b that the UC emission is strongly dependent upon the concentration of Er³⁺ and Yb³⁺ ions. In the samples with the constant Er³⁺ concentration (2 mol% and 3 mol%) the UC emission intensities except ⁴F_7/2–⁴I_15/2 UC emission increase as the Yb³⁺ concentration increases from 10 mol% to 20 mol%, and the maximum UC emission intensity appears at 20 mol% Yb³⁺ doping concentration. Then further increasing Yb³⁺ doping concentration (such as 30 mol%, 40 mol%) deduces the UC emission to fall (Fig. 7a and b). In similar fashion in the samples with the constant Yb³⁺ doping (20 mol%), the UC emission starts an upward trend as the Er³⁺ doping concentration increases from 0.5 mol% to 2 mol%, arrives at a peak at 3.0 mol% Er³⁺ doping concentration, then further increasing Er³⁺ concentration deduces to fall (see Fig. 7c). Thus, NaBiF₄:3 mol% Er³⁺/20 mol% Yb³⁺ sample exhibits the strongest UC emission, and possesses stronger green UC emission about 4 times than hexagonal NaYF₄:2 mol% Er³⁺/20 mol% Yb³⁺ nanocrystals under experimental condition (Fig. 7d). The size and morphology of the latter exhibit small conical hexagonal prisms with ca. 50 nm in diameter and 80–100 nm in length (see Fig. 7b and e in ref. 36).

Fig. 7 The UC emission spectra of the h-NaBiF₄:Er³⁺/Yb³⁺ samples with different concentrations of Er³⁺ and Yb³⁺: (a) 2 mol% Er³⁺/xYb³⁺ (x = 10, 20, 30 and 40 mol%); (b) 3 mol% Er³⁺/xYb³⁺ (x = 10, 20 and 30 mol%); (c) xEr³⁺/20 mol% Yb³⁺ (x = 0.5, 1, 2, 3 and 4 mol%).

Fig. 8 shows the UC emission spectra of NaBiF₄:3 mol% Er³⁺/20 mol% Yb³⁺ prepared under different conditions (including the reaction time, the reaction temperature, NaF/Ln and NH₄NO₃/Ln molar ratio). It can be seen that the UC emissions are similar in the line, and the only difference between these bands are their intensities. From the figure plugged in Fig. 8a, the as-obtained sample at the NaF/Ln molar ratio of 5 has the highest UC emission intensity, and the strong green emission can be observed under several mW 980 nm excitation. The samples prepared at the NaF/Ln molar ratio of 3 including c-BiF₃ and t-BiOF possess the lowest UC intensity, whereas the NaF/Ln molar ratio of 4 leads to the formation of h-NaBiF₄, thus substantially improving the intensity of UC emission. The resulting samples at the NaF/Ln molar ratio of 5, 6, and 8 exhibit prism with the almost large diameter (ca. 0.8 μm), their mean lengths become short. Similar to the hexagonal NaYF₄:2 mol% Er³⁺/20 mol% Yb³⁺, the green UC emission gradually decreased as the NaF/Ln molar ratio increased from 5 through 6 to 8, that is to say, the larger the length of the prism is, the stronger the UC emission is (see Fig. 3 and Table 1).

Fig. 8 The UC emission spectra of the h-NaBiF₄:Er³⁺/Yb³⁺ samples with different conditions. (a) Different Na/Ln molar ratios; (b) different NH₄NO₃/Ln molar ratios; (c) different reaction temperatures; (d) different reaction times.

In addition, the change of the NH₄NO₃ usage amount can alter UC emission intensities of the samples (Fig. 8b). For the fixed NaF/Ln molar ratio (5, 6 and 8, respectively), as NH₄NO₃/Ln molar ratio gradually increases, the UC emission intensity of the corresponding sample also gradually reduces (plugged in Fig. 8b). This can be also related to the morphology and size of the sample. The increase of NH₄NO₃/Ln molar ratio deduces the length of the rod to shorten, thereby the UC emission of the corresponding sample also decreases.

With the enhancing reaction temperature and elongating reaction time, the UC emission became stronger and stronger (Fig. 8c and d). As mentioned above, the higher reaction temperature and longer reaction time are favorable for the transition from m-NH₄BiF₄, t-BiOF to h-NaBiF₄ in the molten salt procedure and the proportion of h-NaBiF₄ increasing. Thus, the UC emission intensity of the samples is dramatically improved.

To determine the nature of the UC emission mechanisms responsible, we study the evolution of the intensities as a function of the pumping power for NaBiF₄:3 mol% Er³⁺/20 mol% Yb³⁺ sample. The UC emission intensity I should scale as Pⁿ at low power limit (far from the saturation), where n stands for the number of pumping phonons absorbed per UC phonon emitted and P represents the pump power.^48–51 A plot of ln I vs. ln P yields a straight line with the slope n. Fig. 9 shows the plots for ⁴G_11/2, ²H_9/2, ⁴F_9/2 ²H_11/2, ⁴S_3/2, ⁴F_7/2 and emission. The slopes for ²H_11/2 and ⁴S_3/2 and are 1.70 and 1.61, respectively, which are close to 2, indicating that two-phonon process deal with the UC mechanism responsible for populating these emitting levels, whereas the slope for ⁴F_9/2, ⁴G_11/2 and ²H_9/2 is 2.12, 2.4 and 2.18, respectively, implying that it is necessary to feed ⁴G_11/2 and ²H_9/2 level via three-phonon process. In addition, the slope for ²H_11/2 and ⁴G_11/2 is relatively larger than that for ⁴S_3/2 and ²H_9/2, indicating that a faster growth of the former than the latter. This can be resulted from the thermal effect. As the pump power increases to high enough, the strong absorption brings forth the enhancement of the sample temperature at the irradiated region, generating the thermal effect and the fluorescence quenching. The small energy gap of ²H_11/2/⁴S_3/2 and ⁴G_11/2/²H_9/2 pairs deduces more Er³⁺ ions to feed the ²H_11/2 and ⁴G_11/2 upper states via the thermal distribution as the pump power rise gradually. Therefore, the ²H_11/2 and ⁴G_11/2 state has a large slope n. In these samples, the predominant green emission in UC emission fluorescence indicates a larger population of the ²H_11/2/⁴S_3/2 level and a smaller population of the ⁴F_9/2 level. In this case, the ²H_11/2(Er³⁺1) + ⁴I_15/2(Er³⁺2) → ⁴I_9/2(Er³⁺1) + ⁴I_13/2(Er³⁺2) cross relaxation between Er³⁺ and the ²F_7/2(Yb³⁺) + S_3/2(Er³⁺) → ²F_5/2(Yb³⁺) + ⁴I_13/2(Er³⁺)^{9–12,48–51} energy back transfer becomes important. As a result, two NIR excitation photons are used to feed the ⁴I_13/2 state of Er³⁺. Since ⁴F_9/2 level is populated from ⁴I_13/2 through energy transfer between Er³⁺ and Yb³⁺. Consequently, the ⁴F_9/2 population, and thus red emission becomes a three-photons process.

Fig. 9 Logarithmic plot of the integrated UC emission intensity of h-NaBiF₄:3 mol% Er³⁺/20 mol% Yb³⁺ sample at 295 K after 980 nm excitation.

To investigate UC emission mechanism and study the temperature dependence UC characteristics, the UC emission spectra of NaBiF₄:3 mol% Er³⁺/20 mol% Yb³⁺ at different temperatures under the excitation wavelength of 980 nm were depicted in Fig. 11a and b. The intensity of the UC emission band attributed to the ⁴S_3/2 → ⁴I_15/2 and ²H_9/2 → ⁴I_15/2 transition of Er³⁺ exhibits the same temperature dependent behaviors, which may be resulted from that the population on the ²H_9/2 state of Er³⁺ can be excited from S_3/2/²H_11/2 state via the cross relaxation process of ⁴S_3/2(Er³⁺1) + ⁴I_13/2(Er³⁺2) → ²H_9/2(Er³⁺1) + ⁴I_15/2(Er³⁺2).^9–12 They stay at the same level with increasing temperature from 10 to 150 K, then experience a significant downward trend. The other UC emissions exhibit another similar temperature dependent behavior. With decreasing temperature the emission intensities increase initially, reach a maximum at 150 K, and then decrease (Fig. 11c). The reduction in the emission can be resulted from the decrease in resonant transfer because of the thermal depopulation of the higher-energy ⁴I_11/2 Er³⁺ levels, which has a resonant match with the ⁴F_7/2 level. The decrease of emission intensity can be originated from the temperature dependence of the multi-phonon de-excitation of the ⁴S_3/2(²H_11/2) and ⁴F_9/2 level. Since it is most likely for a relatively tightly coupled pair to transfer to a single ion is and such pairs have the probability to relax back to ⁴I_J levels again, the UC fluorescence quenching takes place. The fluorescence quenching becomes more effective than there occurs for the excitation into ⁴F_7/2 of Er³⁺ aggravating the fluorescence quenching with increasing temperature. For ⁴F_7/2, the emissions originating mainly from relatively isolated ions cause to reduce rate relaxation rate of ion-pair and to decrease the UC emission.^49–54

Fig. 10 The excitation spectra (a) and emission spectra (b) of h-NaBiF₄:3 mol% Er³⁺/20 mol% Yb³⁺ and pure h-NaBiF₄; (c) the cooperative UC spectra of h-NaBiF₄:20 mol% Yb³⁺.

Fig. 11 (a and b) The UC emission spectra of the NaBiF₄:3 mol% Er³⁺/20 mol% Yb³⁺ sample at different temperatures (the excitation wavelength is 980 nm). (c and d) The UC emission intensity of NaBiF₄:3 mol% Er³⁺/20 mol% Yb³⁺ sample as a function of the temperature; (e and f) the R_HS and R_GS of NaBiF₄:3 mol% Er³⁺/20 mol% Yb³⁺ samples a function of the temperature.

In order to better understand the temperature-dependent behavior, R_HS (the ratio of ²H_11/2 → ⁴I_15/2 to ⁴S_3/2 → ⁴I_15/2) and R_GH (the ratio of ⁴G_11/2 → ⁴I_15/2 to ²H_9/2 → ⁴I_15/2) as a function of temperature are shown as Fig. 11d, respectively. R_HS and R_GH possess the same temperature dependence characteristics. R_HS and R_GH firstly increase as the temperature enhances from 10 K to 200 K, then have a downward trend and hit rock bottom when the temperature increases to 250 K, finally recover to have an upward trend. Such phenomenon that the decreasing trend at 200–250 K indicates that there must be some additional UC channel energy transfers are involved with Yb³⁺ cooperative UC followed by the energy transfer to Bi³⁺, and the nonradiative energy transfer from the excited levels of Er³⁺ to the ³P₁ manifold of Bi³⁺, followed by energy back energy to another manifolds of Er³⁺.

To authenticate the existence of energy transfer process between Bi³⁺ and Er³⁺, the excitation and emission spectra of NaBiF₄:3 mol% Er³⁺/20 mol% Yb³⁺ and NaBiF₄ have been measured and are depicted in Fig. 10. The excitation and emission wavelengths are 245 and 545 nm, respectively. In the excitation spectra (Fig. 10a), there exists broad absorption bands owing to Bi³⁺ ions,^15–19 while some emission bands due to the transitions of Er³⁺ ions are also observed in the emission spectra (Fig. 10b),^1–12 implying that the interaction between Bi³⁺ and Er³⁺ takes place. In addition, the anti-Stokes UC fluorescence spectra of NaBiF₄:20 mol% Yb³⁺ samples under 980 nm excitation at room temperature was also shown in Fig. 10c. The observed anti-Stokes in the range of 470–510 nm is believed to be cooperative UC causing by the simultaneous radiative relaxation of the excited Yb³⁺–Yb³⁺ pairs accompanied by the emission of a visible photon in the expression ²F_5/2(Yb³⁺)–²F_5/2(Yb³⁺) → ²F_7/2(Yb³⁺) + ²F_7/2(Yb³⁺) + hν^45–47 and the emission from the ³P₁ excited state to the ¹S₀ ground state of Bi³⁺.^14,15,22

To further validate the rationality of the energy transfer from Bi³⁺ to Er³⁺, the UC emission spectra of NaYF₄:2 mol% Er³⁺/20 mol% Yb³⁺ at different temperatures and the emission decay curves of ⁴S_3/2 → ⁴I_15/2 transition in h-NaBiF₄:3 mol% Er³⁺/20 mol% Yb³⁺ (the former) and h-NaYF₄:2 mol% Er³⁺/20 mol% Yb³⁺ (the latter) were measured under the excitation wavelength of 980 nm, as shown in Fig. 12. Both h-NaBiF₄ and h-NaYF₄ belong to the hexagonal system with the same space group P3̄, and the crystal cell parameter of h-NaBiF₄ (a = 6.144 Å, c = 3.721 Å)³² is slightly larger than that of h-NaYF₄ (a = 5.96 Å, c = 3.53 Å).⁵⁵ In the same doping concentration, the mean distance between Yb³⁺ ions in the former is larger than that in the latter. The cooperative UC intensity depends upon the shortest distances between Yb³⁺ ions, and the smaller the distance between Yb³⁺ ions, the stronger the cooperative UC emission,^45–47 thereby, the cooperative UC intensity of the former is weaker than that of the latter. As a matter of fact, the UC emission intensity of the former is stronger than that of the latter (see figure). It is suggested that there exists the energy transfer between Er³⁺ ion or Yb³⁺ ion and Bi³⁺ ion in h-NaBiF₄:3 mol% Er³⁺/20 mol% Yb³⁺ sample. In addition, the decay curves of both the samples could be well fitted into a single exponential function (Fig. 12a). It is obvious that the lifetime of the ⁴S_3/2 state in the former (267 μs) is longer than that in the latter (147 μs). Moreover, the former exhibits longer rise and delay times than the latter. This further affirms the existence of energy transfer between Er³⁺ ion and Bi³⁺ ions in h-NaBiF₄:3 mol% Er³⁺/20 mol% Yb³⁺.

Fig. 12 (a) The UC emission spectra of the NaYF₄:2 mol% Er³⁺/20 mol% Yb³⁺ sample at different temperatures (the excitation wavelength is 980 nm). (b) The decay curves UC emission intensity of 532 nm of NaBiF₄:3 mol% Er³⁺/20 mol% Yb³⁺ and NaYF₄:2 mol% Er³⁺/20 mol% Yb³⁺ samples.

Based on the above analysis, the energy diagram showing the proposed UC emission mechanism and energy transfer process in NaBiF₄:Er³⁺/Yb³⁺ materials was depicted in Fig. 13. The initially increasing Yb³⁺ doping rapidly promotes the rate of energy transfer from Yb³⁺ to Er³⁺, enhancing the population on the green-emitting participating ⁴F_7/2 and red-emitting participating ⁴I_11/2 manifolds, which results in the improvement of the UC emission. Too high Yb³⁺ doping concentration induces an efficient energy back-transfer process ⁴F_7/2(Er³⁺) + ²F_5/2(Yb³⁺) → ⁴I_11/2(Er³⁺) + ²F_5/2(Yb³⁺)^9–12 and ⁴F_9/2(Er³⁺) + ²F_5/2(Yb³⁺) → ⁴I_13/2(Er³⁺) + ²F_5/2(Yb³⁺),^9–12 causing the decrease of the population on the ⁴F_7/2, ⁴I_11/2, ²H_11/2/⁴S_3/2 and ⁴F_9/2 manifold, accordingly the green UC emission and red one both experiences a significant downward trend. At low Er³⁺ doping concentration, there are not enough Er³⁺ ions to be populated or excited. The increase of Er³⁺ concentration makes the average distance between Er³⁺ ions shrink and the interaction between ions intensifies. This induces the population on the ⁴F_7/2, ⁴I_11/2, ²H_11/2, ⁴S_3/2 and ⁴F_9/2 states increase, and gives rise to the increase of the population on the green and red-emitting manifolds, thus, the red and green emission experiences an upward trend. When the Er³⁺ doping concentration achieves a certain crisis, the average interaction distance between Er³⁺ and Yb³⁺ is beyond the critical distance of the energy back transfer ⁴I_11/2(Er³⁺) + ²F_7/2(Yb³⁺) → ⁴I_15/2(Er³⁺) + ²F_5/2(Yb³⁺) (EBT),^9–12 which leads to a considerate drop the population on the red-emitting ⁴F_9/2 manifold and the green-emitting ²H_11/2/⁴S_3/2 and violet emitting ⁴G_11/2 manifold. As a consequence, the intensity of the visible emission abruptly falls off, while the green one quickly rises. Obviously, 4 mol% exceeds the critical concentration, whereas 3 mol% is below this threshold.

Fig. 13 Schematic energy level diagrams of Bi³⁺, Er³⁺ and Yb³⁺ ions as well as the proposed energy transfer of NaBiF₄:Er³⁺/Yb³⁺ under excitation of 980 nm.

4. Conclusions

The hexagonal NaBiYF₄:Er³⁺/Yb³⁺ micro-/nanocrystals were synthesized via low-temperature molten-salt method in NH₄NO₃ flux. The influences of the reaction time, the reaction temperature, the content of NaF and NH₄NO₃ on the crystal phase of the resulting products were studied at length. This research also investigates and discusses in detail the effects of shapes, sizes, doping concentration of Er³⁺ and Yb³⁺, the pumping power and the temperature on the UC emission of NaBiYF₄:Er³⁺/Yb³⁺ micro-/nanocrystals. The results show that the NaBiF₄:3 mol% Er³⁺/20 mol% Yb³⁺ sample exhibits the strongest UC emission and can be comparable to the most efficient UC hexagonal NaYF₄:Er³⁺/Yb³⁺ micro-/nanocrystals. Two-phonon process takes responsibility to populate ⁴S_3/2/²H_11/2 of Er³⁺, whist the ⁴F_9/2 is involved in a three-photon process. The energy transfer between the excited levels of Er³⁺ and the ³P₁ manifold of Bi³⁺ and cooperative UC of Yb³⁺ are favorable to improve upconversion emission of Er³⁺. This work is significant not only to provide one novel synthesis method of Bi-based fluoride UC materials but also for the better understanding of their essential UC emission mechanisms (particularly the energy transfer from Bi³⁺ to Er³⁺).

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (61204003), Train Object Program of Jiangxi Province Young Scientists, Natural Science Foundation of Jiangxi Province (No. 20142BAB206011 and 2011BAB206029) and Luodi program of Jiangxi Province (No. KJLD13030), respectively.

References

1. B. Zhou, B. Y. Shi and X. G. Liu, Nat. Nanotechnol., 2015, 10, 924.
2. M. Haase and H. Schäfer, Angew. Chem., Int. Ed., 2011, 50, 5808.
3. S. Y. Han, R. R. Deng, X. J. Xie and X. G. Liu, Angew. Chem., Int. Ed., 2014, 53, 2.
4. G. Y. Chen, J. W. Seo, C. H. Yang and P. N. Prasad, Chem. Soc. Rev., 2013, 242, 8304.
5. W. F. yang, X. Y. Li, D. Z. Chi, H. J. Zhang and X. G. Liu, J. Nanotechnol., 2015, 25, 482001.
6. X. M. Li, F. Zhang and D. Y. Zhao, Chem. Soc. Rev., 2015, 44, 1346.
7. X. D. Wang, O. S. Wolfbeis and R. J. Meier, Chem. Soc. Rev., 2013, 42, 7834.
8. G. Y. Chen, H. L. Qiu, P. N. Parasad and X. Y. Chen, Chem. Rev., 2014, 110, 5161.
9. W. Zhang, P. Huang, E. Ma, H. M. Zhu and X. Y. Chen, Chem. Soc. Rev., 2015, 44, 1379.
10. H. Dong, L. D. Sun and C. H. Yan, Nanoscale, 2013, 5, 5657.
11. L. P. Tu, X. M. Liu, F. Wu and H. Zhang, Chem. Soc. Rev., 2015, 44, 1331.
12. H. Dong, L. D. Sun and C. H. Yan, Chem. Soc. Rev., 2015, 44, 1608.
13. J. F. Suyver, J. Grimm, M. K. van Veen, D. Biner, K. W. Krämer and H. U. Güdel, J. Lumin., 2006, 117, 1.
14. P. Boutinaud, Inorg. Chem., 2013, 52, 6028.
15. R. H. P. Awater and P. Dorenbos, J. Lumin., 2017, 184, 221.
16. G. J. Gao, M. Y. Peng and L. Wondraczek, J. Mater. Chem. C, 2014, 2, 8085.
17. R. V. Yadav, R. S. Yadav, A. Bahadur and A. K. Singh, Inorg. Chem., 2016, 55, 10928.
18. A. Urbina-Frías, T. López-Luke, H. Desirena, P. Salas, A. Terres-Casatro and E. De la Rosa, Opt. Mater., 2015, 48, 92.
19. M. Z. Yang, Y. Sui, S. C. Lü, M. J. Wang, X. J. Wang, M. H. Wu, Y. Wang, Y. T. Q. Lü and W. F. Liu, J. Alloys Compd., 2011, 509, 8590.
20. J. P. Fu, R. Pang, Y. L. Jia, W. Z. Sun, L. H. Jiang, S. Zhang and C. Y. Li, J. Lumin., 2017, 181, 240.
21. P. P. Lei, P. Zhang, S. Yao, S. Y. Song, L. L. Dong, X. Liu, X. L. Liu, K. M. Du, J. Feng and H. J. Zhang, ACS Appl. Mater. Interfaces, 2016, 8, 27490.
22. N. Niu, F. He, S. L. Gai, C. X. Li, H. Huang and P. P. Yang, J. Mater. Chem., 2012, 22, 21613.
23. K. J. Chong, T. Hirai, S. Hashimoto and N. Ohno, J. Lumin., 2017, 122–123, 149.
24. A. Escudero, E. Moretti and M. Ocaña, CrystEngComm, 2014, 16, 3274.
25. J. M. Zhao, H. L. Pan, X. He, Y. S. Wang, L. Gu, Y. S. Hu, L. Q. Chen, H. Z. Liu and S. Dai, Nanoscale, 2013, 5, 518.
26. W. B. Niu, S. L. Wu, S. F. Zhang, J. Li and L. Li, Dalton Trans., 2011, 40, 3305.
27. H. X. Mai, Y. W. Zhang, L. D. Sun and C. H. Yan, J. Phys. Chem. C, 2007, 111, 13721.
28. W. J. Yao, Q. Y. Tian, Z. H. Wu, S. Y. Cui, Z. G. Dai and W. Wu, J. Mater. Chem. C, 2016, 4, 6327.
29. H. L. in, D. K. Xu, A. M. Li, D. D. Teng, S. H. Yang and Y. L. Zhang, J. Mater. Chem. C, 2015, 3, 11754.
30. D. T. Klier and M. U. Kumke, J. Mater. Chem. C, 2015, 3, 11228.
31. P. P. Lei, R. An, S. Yao, Q. S. Wang, L. L. Dong, K. M. Du, J. Feng and H. J. Zhang, Adv. Mater., 2017, 22, 00505.
32. P. Fedorov, et al., Russ. J. Inorg. Chem., 1979, 24, 157.
33. B. Shao, Z. W. Yang, Y. D. Wang, J. Li, J. Z. Yang, J. B. Qiu and Z. G. Song, ACS Appl. Mater. Interfaces, 2015, 7, 25211.
34. J. Y. Liao, Z. W. Yang, H. J. Wu, D. Yan, J. B. Qiu, Z. G. Qiu, Y. Yang, D. C. Zhou and Z. Y. Yin, J. Mater. Chem. C, 2013, 1, 6541.
35. S. L. Gai, C. X. Li, P. P. Yang and J. Lin, Chem. Rev., 2014, 114, 2343.
36. X. Y. Huang, G. H. Hu, Q. J. Xu, X. X. Li and Q. M. Yu, J. Alloys Compd., 2014, 616, 652.
37. X. Y. Huang, Opt. Mater. Express, 2014, 4, 2381.
38. X. Y. Huang, L. Jiang, X. X. Li and A. Q. He, J. Alloys Compd., 2017, 6721, 374.
39. X. Y. Huang, Opt. Mater., 2015, 50, 81.
40. D. M. Mathews, R. B. Ambekar, K. A. Tyagi and J. Köhler, J. Alloys Compd., 2004, 377, 162–166.
41. F. Zhang, Y. Wan, T. Yu, F. Q. Zhang, Y. F. Shi, S. H. Xie, Y. G. Li, L. Xu, B. Tu and D. Y. Zhao, Angew. Chem., Int. Ed., 2007, 46, 7976.
42. A. Naduviledathu Raj, T. Rinkel and M. Hasse, Chem. Mater., 2014, 26, 5689.
43. M. Y. Ding, C. H. Lu, L. H. Cao, Y. R. Ni and Z. Z. Xu, CrystEngComm, 2014, 4, 29165.
44. X. Qu, H. Song, X. Bai, G. Pan, B. Dong, H. Zhao, F. Wang and R. Qin, Inorg. Chem., 2008, 47, 9654.
45. E. Nakazawa and S. Shionoya, Phys. Rev. Lett., 1970, 25, 1710.
46. D. G. Deng, S. Q. Xu, R. Q. Bao, S. L. Zhao, B. L. Wang, H. P. Wang and H. D. Ju, J. Phys. D: Appl. Phys., 2009, 42, 105111.
47. X. Y. Huang, Chin. Opt. Lett., 2010, 8, 780.
48. M. Pollnau, D. R. Gamelin, S. R. Lüthi, H. U. Güdel and M. P. Hehlen, Phys. Rev. B: Condens. Matter Mater. Phys., 2000, 61, 3337.
49. F. Vetrone, J. C. Boyer, J. A. Capobianco, A. Speghini and M. Bettinll, J. Appl. Phys., 2004, 96, 661.
50. J. F. Suyver, A. Aebisher, S. García-Revilla, P. Gerner and H. U. Güdel, Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 71, 125123.
51. J. P. van der Ziel, F. W. Ostermayer Jr and L. G. Van Uitert, Phys. Rev. B: Condens. Matter Mater. Phys., 1970, 2, 4432.
52. K. Y. Wu, J. B. Cui, X. Kong and Y. J. Wang, J. Appl. Phys., 2011, 110, 053510.
53. A. M. Pires and O. A. Serra, J. Appl. Phys., 2005, 98, 063529.
54. D. D. Li, Q. Y. Shao, Y. Yan and J. J. Jiang, J. Phys. Chem. C, 2014, 118, 22807.
55. Y. S. Liu, H. M. Zhu, R. F. Li, L. Q. Liu and X. Y. Chen, Angew. Chem., Int. Ed., 2013, 52, 1128.

---