1 Introduction

Lanthanide-doped upconversion nanoparticles (UCNPs) have attracted considerable attentions as a result of their outstanding luminescent performance compared with traditional photoluminescent (PL) nanomaterials such as florescent organic dyes and quantum dots (QDs) [1,2,3,4,5,6,7]. Owning to their fine photostability, low photobleaching, long lifetimes, minimal phototoxicity and deep penetration depth [8,9,10], UCNPs are promising in three-dimensional (3D) flat panel displays [11], white light emitting diodes [12, 13], solar cells [14], and especially in biolabels and bioimaging [15,16,17,18]. However, UCNPs remain suffering from weak luminescent intensity, low efficiency, high power excitation and a confined excitation wavelength during UC process [19,20,21,22,23]. During last decades, a lot of efforts have been made to enhance UC intensity, such as broadband sensitization [24, 25], surface passivation [26,27,28,29], energy transfer modulation [30, 31], surface plasmon coupling [32]. In addition, exploring new host matrix is also an efficient method, because the local crystal field of host matrix and the interaction between host and doping ions have a strong impact on the UC process. As widely confirmed, host lattice determines the distance, spatial position, coordination numbers and the type of surrounding anions of doping ions [33, 34].

As a recently prevalent topic, Sc-based nanomaterials have become hotspots. Sc belongs to the rare earth family. However, it has no 4f electrons and behaves more like transition metals, which endows Sc-based nanomaterials unique properties, such as different UC luminescence behaviors from those of Y/Ln-based nanomaterials and remarkable negative thermal expansion constant [35,36,37,38,39]. Currently, alkali (Na, K, Rb, Cs) scandium fluorides with various morphologies and size including M2NaScF6 (M = K, Rb, Cs), NaScF4, Na3ScF6 and KSc2F7 have been synthesized and investigated [40,41,42,43,44]. However, alkaline earth ions (Ca2+, Sr2+ and Ba2+) owning closed ionic radii with lanthanide ions were almost ignored. Recently, Yang’s group have successfully synthesized SrSc2F7 and Ba3Sc2F12 crystals with various morphologies by hydrothermal process and investigated the fluorescence of the new host, which indicates that SrSc2F7 and Ba3Sc2F12 are excellent host material for UC luminescence [45, 46]. However, the research on synthesis and UC luminescence properties of nanoscaled alkaline rare earth fluorides with uniform morphology, and good dispersibility is still absent to the best of our knowledge.

Herein, we report the precise synthesis of Ba3Sc2F12 nanocrystals by thermolysis method. The synthesis conditions, including temperature, reaction time and ratio of reactants, are optimized to control the crystal growth process and to get high-quality products. Then, we carefully explore the UC luminescence properties of Yb/Er and Yb/Ho doped Ba3Sc2F12 nanocrystals and the corresponding energy transfer mechanisms. In order to obtain the strongest UC luminescent intensity, we modify the doping concentration of sensitizer ions Yb3+ from 10 mol% to 98 mol% and activator ions Er3+ or Ho3+ from 1 mol% to 8 mol%. We also notice that with Yb3+ doping concentration increasing, pure tetragonal phase Ba3Sc2F12 can gradually transform to hexagonal phase Ba4Yb3F17 nanocrystals.

2 Experimental

2.1 Materials

Sc2O3 (99.99%), Yb2O3 (99.99%), Er2O3 (99.99%), Ho2O3 (99.99%) and Ba(CH3COO)2 (99.99%) were purchased from HWRK Chem Co. Ltd. (Beijing China). CF3COOH (99%) was obtained from Lingfeng Chemical Reagent Co. Ltd. (Shanghai China). Aleic acid (OA, 90%) and 1-octadecene (ODE, 90%) were purchased from Sigma-Aldrich. All the materials were used directly without further purification.

2.2 Characterization

Powder X-ray diffraction (XRD) data were recorded via a Rigaku D/max 2550 with Cu Kα radiation (λ = 0.15418 nm). The morphology and size of nanocrystals were analyzed by the Hitachi 7700 transmission electron microscope (TEM) with an acceleration voltage of 100 kV. High-resolution TEM (HRTEM) images were obtained through Tecnai G2 F20 field emission transmission electron microscope (FESEM) equipped with an energy-dispersed X-ray system (EDX) operating at 200 kV. Inductively coupled optical emission spectrometry (ICP-OES) analysis was performed on a PE Optima 5300DV spectrometer. The UC luminescence spectra were collected using an Edinburgh F920 spectrometer with an external continues wave diode 980 nm laser at room temperature. All luminescence studies were carried out under identical conditions.

2.3 Synthesis of RE(CF3COO)3 precursors (RE = Sc, Yb, Er, Ho)

In a typical synthesis, RE2O3 (2 mmol) was dissolved in 2 ml deionized water under vigorous magnetic stirring. CF3COOH was added to above mixture at room temperature. Then, the solution was heated to 100 °C and kept for 12 h until the solution became transparent. The final products were dried in a vacuum oven at 60 °C for 24 h.

2.4 Synthesis of Ba3Sc2F12:Yb/Er nanocrystals

In a typical synthesis, 0.200 mmol Ba(CH3COO)2, 0.312 mmol Sc(CF3COO)3, 0.080 mmol Yb(CF3COO)3 and 0.008 mmol Er(CF3COO)3 were added into a 100 ml three-necked round-bottom flask containing 17 ml oleic acid and 3 ml octadecene. The mixture was heated to 120 °C and kept for 30 min with magnetic stirring under an argon atmosphere. Subsequently, the mixture was heated to 310 °C and maintained for 50 min before cooling down to room temperature. Finally, the as-prepared samples were precipitated by addition of ethanol, collected by centrifugation at 9000 r·min−1 for 5 min, washed with water and ethanol several times, and finally redispersed in cyclohexane for further characterization.

3 Results and discussion

Ba3Sc2F12 and lanthanide ions doped Ba3Sc2F12 nanocrystals were prepared via the modified thermolysis method [47]. The crystal growth process is clearly presented in Fig. 1, where aleic acid (OA) was used as surfactant and noncoordinating 1-octadecene (ODE) as reaction solvent for its high boiling point. The carboxyl group of OA can coordinate the rare earth elements, while the long hydrocarbon chain acts as a passivating ligand which can prevent the final nanoparticles from agglomeration. Then, by carefully tailoring experimental conditions, high-quality Ba3Sc2F12 nanocrystals with great dispersibility, narrow size distribution and good crystallinity could be readily acquired.

Fig. 1
figure 1

Scheme of growth process of Ba3Sc2F12 and Ba3Sc2F12:Ln (Ln = Yb, Er, Ho) nanocrystals via thermolysis method

Full size image

Figure 2a1–e1 shows the effect of Ba2+ content on the morphology and crystal structure of Ba3Sc2F12:Yb/Er nanocrystals. At low Ba2+ content, most of the products were irregular nanoparticles and only few tetragonal nanosheets were observed (Fig. 2a1). XRD data indicate that the products are mixture of cubic phase ScF3 and tetragonal phase Ba3Sc2F12 (Fig. 2e1). With the increase in Ba2+ content, pure tetragonal phase Ba3Sc2F12 could be acquired. When the molar ratio of Ba2+ to Sc3+ was 0.50, uniform and well dispersed Ba3Sc2F12 nanosheets were synthesized and the particle size was about 35 nm × 35 nm. However, at high Ba2+ content (Ba:Sc = 2.00), majority of products were tetragonal nanosheets with the size of 5 mm × 5 nm. These results indicate that Ba2+ content has considerable influence on the shape and phase evolution of Ba3Sc2F12 nanocrystals and that appropriate Ba2+ content is crucial to the synthesis of uniform and pure Ba3Sc2F12 nanocrystals.

Fig. 2
figure 2

Ba3Sc2F12 nanocrystals synthesized under different reaction conditions: a1–d1 TEM images and e1 XRD patterns of Ba3Sc2F12 nanocrystals synthesized under different ratios of Ba to Sc (0.25, 0.50, 1.00, 2.00); a2–d2 TEM images and e2 XRD patterns of Ba3Sc2F12 nanocrystals synthesized at 280, 290, 300 and 310 °C, respectively; a3–h3 TEM images and i3 XRD patterns of Ba3Sc2F12 nanocrystals synthesized at different reaction time (0, 10, 20, 30, 40, 50, 60, 90 min)

Full size image

The morphology and phase development caused by Ba2+ content can be explained by the crystallization speed-controlled model [48,49,50]. The nanocrystal growth was initiated with the nucleation process of Ba2+, Sc3+ and F−. In addition, the change of monomer concentration in the solution is mainly caused by the nucleation process, while the following growth process has relatively slight impact. The precursors in our experiments were Ba(CH3COO)2 and Sc(CF3COO)3, while the ratio of F− to Sc3+ was set as 3, so the ratios of F− to Ba2+ and Ba2+ to Sc3+ changed depending on the amount of Ba2+ added. Low concentration of Ba2+ shall decelerate the particle crystallization rate and produce relatively small number of nuclei, leading to a high monomer concentration, so highly uniform tetragonal Ba3Sc2F12 nanosheets were obtained under Ba2+:Sc3+  = 0.50. However, according to the mechanism of chemical reaction equilibrium, Ba2+ content was too low to form Ba3Sc2F12 when the ratio of Ba2+ to Sc3+ was 0.25, so the cubic phase ScF3 was obtained. On the contrary, high Ba2+ content can enhance the crystallization speed which results in rapid depletion of RE(CF3COO)3 in the solution. When the monomer concentration was reduced to a critical level rapidly, Ba3Sc2F12 nanocrystals with smaller size and worse crystallinity appeared.

Then, the effect of the reaction temperature on the morphology and crystal structure of Ba3Sc2F12 nanocrystals were explored. Corresponding results are listed in Fig. 2a2–e2. At 280 and 290 °C, most of the products were of irregular morphology and only little tetragonal Ba3Sc2F12 nanosheets were obtained. XRD results in Fig. 2e2 show that the products were not pure tetragonal phase. As the reaction temperature increased to 300 °C, pure tetragonal phase was acquired. However, the morphology was not uniform and many small round Ba3Sc2F12 nanosheets were still obtained. XRD results show that the products were only pure tetragonal phase. Further increase the temperature to 310 °C, uniform tetragonal nanosheets with an average size of 35 nm × 35 nm were acquired. Herein, 310 °C was suitable temperature to synthesize uniform Ba3Sc2F12 nanocrystals.

In order to track the morphology and phase evolution of Ba3Sc2F12 nanocrystals, we monitored the nanocrystal growth process. As is shown in Fig. 2a3–h3, when the reaction time was set as 0, 10 or 20 min, the products were composed of irregular small nanoplates and little small tetragonal nanosheets. Corresponding XRD results indicate the existence of impurities (Fig. 2i3). When the reaction was prolonged to 30 min, pure tetragonal phase Ba3Sc2F12 nanocrystals could be obtained, which can be seen from the XRD patterns. TEM images show fewer small irregular and more uniform tetragonal nanosheets as the time increased. However, when the reaction time was prolonged to 90 min, the corners and tips of tetragonal nanosheets were smoothed and TEM images show irregular nanoparticles. TEM images of Ba3Sc2F12 nanocrystals at different reaction time in Fig. 2a3–h3 obviously demonstrate a typical Ostwald ripening process [51, 52]. In addition, 50 min was enough for Ba3Sc2F12 nanocrystals ripening process.

Based on the above results and discussion, we knew that the suitable ratio of Ba2+ to Sc3+ was 0.5, the best reaction temperature was 310 °C and 50 min was enough for the synthesis of Ba3Sc2F12 nanocrystals in current system.

To explore the UC luminescence property and to get the strongest UC luminescence intensity of Ba3Sc2F12 nanocrystals, we synthesized a series of Ba3Sc2F12:Yb/Er (x/2 mol%) nanocrystals under above optimized conditions of which the ratio of Ba2+ to Sc3+ is 0.50, reaction temperature is 310 °C and reaction time is 50 min. The doping content of sensitizer ion Yb3+ was tuned from 10 mol% to 98 mol% firstly. Corresponding TEM images and XRD patterns are listed in Fig. 3. When the concentrations of Yb3+ were 10 mol%–30 mol%, the products obtained were pure tetragonal phase Ba3Sc2F12 nanocrystals with uniform size. With the concentration of Yb3+ further increasing to 40 mol%–50 mol%, the products became big tetragonal nanosheets mixed with irregular small nanosheets, which were tetragonal phase Ba3Sc2F12 and tetragonal phase Ba4Yb3F17, respectively. When the concentration of Yb3+ was more than 60 mol%, the products were pure tetragonal phase Ba4Yb3F17. Inductively coupled plasma mass spectroscopy (ICP-MS) analysis was carried out to characterize the actual concentration of Yb3+ doped into the as-synthesized BamSc1-xFn:Yb/Er (x/2 mol%) nanocrystals, which are listed in Table S1. Considering the same charge and adjacent ionic radii of Sc3+ (0.0745 nm) and Yb3+ (0.0868 nm), we believe that Yb3+ has substituted Sc3+. Figure S1 shows the evolution of (310) diffraction patterns of Ba3Sc2F12:Yb/Er (x/2 mol%). As Yb3+ substitutes Sc3+, the unit cell of Ba3Sc2F12 would swell, leading to the shift of XRD peak to small angles according to the Bragg’s law, which is consistent with XRD results.

Fig. 3
figure 3

a–l TEM images and m, n XRD patterns of BamSc1-xYbxFn: Er (2 mol%) with different amounts of Yb3+ doped (x = 10 mol%–98 mol%)

Full size image

To prove the transformation process from pure tetragonal phase Ba3Sc2F12 to hexagonal Ba4Yb3F17, TEM and high-resolution TEM (HRTEM) images of 20 mol%, 40 mol% and 80 mol% Yb3+ doped samples are presented in Fig. S2. The HRTEM image of 20 mol% doped sample reveals a highly crystalline nature and the interplanar distance between the adjacent lattice fringes is 0.299 nm, which matches well with (310) lattice planes of tetragonal phase Ba3Sc2F12. The HRTEM image of 80 mol% Yb3+ doped product shows that the interplanar distance between the adjacent lattice fringes is 0.294 nm, which agrees well with (214) lattice planes of hexagonal phase Ba4Yb3F17. While 40 mol% Yb3+ doped products are the mixture of tetragonal phase Ba3Sc2F12 and hexagonal Ba4Yb3F17 nanocrystals. The two crystal structures grow independently and exist as a physical mixture rather than, for example, as core–shell particles [39].

Similar results were also observed when 2 mol% Ho3+ ions were used as activator. Corresponding TEM images and XRD patterns are listed in Fig. S3, which indicate that BamSc1−xFn:Yb/Ho (x/2 mol%) nanocrystals transform from pure tetragonal Ba3Sc2F12:Yb/Ho to Ba4Yb3F17:Sc/Ho with Yb3+ concentration varying from 0 mol% to 98 mol%.

Figure 4a displays UCPL emission spectra of BamSc1−xFn:Yb/Er (x/2 mol%) nanocrystals with variable Yb3+ concentration from 0 mol% to 98 mol%. All samples exhibit strong red UC luminescence, which is contrary to NaYF4:Yb/Er (20/2 mol%) nanocrystals that generate strong green UC emission under 980-nm laser excitation. The UC emission intensity of the products also changed with the variation of Yb3+ concentration. When Yb3+ concentration varied from 0 mol% to 25 mol%, the UC luminescence intensity of Ba3Sc2F12:Yb/Er nanocrystals increased gradually. As is known to us all, Yb3+ has a sufficient absorption cross section matching well with common 980-nm laser excitation source. Herein, we reason that introduction of an elevated amount of Yb3+ dopants in the Ba3Sc2F12 host lattice decreases Yb3+–Er3+ interatomic distance and thus facilitates energy transfer from Yb3+ to Er3+. In addition, high Yb3+ contents help sustain multi-step excitation without depletion of the intermediate, which will benefit UC luminescence process. However, as the Yb3+ ions content further increased from 25 mol% to 98 mol%, the luminescent intensity fell down quickly as a result of concentration quenching phenomenon, which means that a high Yb3+ content leads to an increased possibility of long-distance energy transfer that takes excitation energy to lattice or surface defects instead of luminescence [53, 54]. So, for Ba3Sc2F12:Yb/Er system, the optimum Yb3+ concentration for efficient UC under 980-nm excitation was 25 mol%.

Fig. 4
figure 4

a UC emission spectra of BamSc1−xYbxFn:Er (2 mol%) (x = 0, 10, 15, 20, 25, 30, 40, 50, 60, 70, 80, 90 and 98, mol%); b UC emission spectra of Ba3Sc2F12:Yb/Er (25 mol%/x mol%) (x = 1, 2, 4, 6 and 8, mol%) under the excitation of 980-nm NIR diode laser at room temperature

Full size image

Then, we optimized Er3+ concentration. ICP-MS analysis of the actual concentration of Er3+ doped into the as-synthesized Ba3Sc2F12:Yb/Er (25/x mol%) nanocrystals is given in Table S2. Figure 4b shows the UC luminescence spectra of Ba3Sc2F12:Yb/Er (25/x mol%) nanocrystals with variable Er3+ contents ranging from 1 mol% to 8 mol%. As an elevated amount of Er3+ dopants will increase the amount of UC emission centers, the UC intensity becomes stronger and stronger as the Er3+ contents increase from 1 mol% to 4 mol%. However, more than 4 mol% Er3+ doping would lead to the decrease in intensity because high doping level could inevitably lead to localized concentration quenching of activator emissions caused by enhanced cross relaxation between the lanthanide ions. Herein, the optimal Er3+ content is 4 mol%.

For the case of activator Ho3+, we also optimized the doping content of Yb3+. The UC luminescence spectra of Ba3Sc2F12:Yb/Ho (x/2 mol%) nanocrystals are exhibited in Fig. S4a, which shows that the optimal Yb3+ content was 30 mol%. Then, we changed the concentration of Ho3+ from 1 mol% to 8 mol%, finding that the optimal Ho3+ content for Ba3Sc2F12:Yb/Ho (30/x mol%) was 4 mol%, of which the results are shown in Fig. S4b. The actual concentrations of Yb3+ and Ho3+ doped into the as-synthesized Ba3Sc2F12:Yb/Ho samples are listed in Tables S3, S4, respectively.

To further understand the underlying UC mechanism of Ba3Sc2F12:Yb/Er (25/4 mol%), we explored the number of photons involved in the UC luminescence process (Fig. 5a) and UC lifetime (Fig. 5b). According the equation I∞Pn, where I represents the UC luminescence intensity, P is the pumping power, and n is the number of required pumping photons; the emission intensity (I) is proportional to Pn. The logarithmic curves between the UC emission intensity and pumping power are shown in Fig. 5a, where the slope of each curve denotes the number of photons required. Based on the value of n presented in Fig. 5a, the UC emissions of 545 and 654 nm all follow a two-photon process. The decay time of energy transition from 4S3/2 and 4F9/2 to 4I15/2 for Er3+ is 0.54 and 0.41 ms, respectively. Figure 5c shows the proposed mechanism of UC emission spectra of Ba3Sc2F12:Yb/Er (25/4 mol%) nanocrystals. Under 980-nm laser excitation, Yb3+can be excited to 4F5/2 level, from where it transfers its energy to the nearby Er3+ 4I11/2 state and returns to its ground state. Populating on the metastable level 4I11/2, Er3+ can be either excited to its higher state 4F7/2 via absorbing another photon or relax to its lower state 4I13/2 via non-radiative transition from the level 4I13/2, Er3+can also be excited to its higher state 4F9/2 by absorbing a 980-nm photon. The transition of Er3+ at 4F9/2 state to the ground state gives rise to a 654-nm photon, while the Er3+ at 4F7/2 state will relax to 2H11/2 and 4S3/2, from where they emit 525, 545 nm photons and transit back to the ground state 4I15/2. For Ho3+, the red (654 nm) and green (545 nm) emission originates from the characteristic radiative transition from 5F5 and 5F4 of Ho3+ to its 5I8 level and their decay time are 0.35 and 0.25 ms, respectively, as depicted in Figure S5.

Fig. 5
figure 5

a lg–lg plots of UC emission intensity in 648 nm (red) and 548 nm (green) intensity for Ba3Sc2F12:Yb/Er (25 mol%/4 mol%) nanocrystals; b luminescent decay curves of Er3+ in 648 nm (red) and 545 nm (green); c proposed UC mechanism under excitation of 980-nm NIR diode laser at room temperature (E: energy)

Full size image

4 Conclusion

In summary, high-quality tetragonal phase Ba3Sc2F12 nanosheets with uniform morphology and good dispersibility have been obtained via the thermal decomposition method. Then, we adjusted the doping contents of sensitizer Yb3+ and activator ions Er3+ or Ho3+ to get the strongest UC luminescent intensity, demonstrating that the optimum doping concentrations for Yb3+/Er3+ and Yb3+/Ho3+ are 25 mol%/4 mol% and 30 mol%/4 mol%, respectively. Interestingly, when the Yb3+ doping concentration is more than 60 mol%, final products could transform from pure tetragonal phase Ba3Sc2F12 to hexagonal Ba4Yb3F17 nanocrystals. This study would provide a new kind of UC nanomaterials with strong red emission, and further complements the investigation of Sc-based nanomaterials, which may have promising applications that the Ln/Y-based materials don't possess.