A. Morphology and structural characterization of GdOF: Yb³⁺, Er³⁺ hollow nanospheres

Figure 1(a) shows SEM images of the Gd(OH)₃: Yb³⁺, Er³⁺ nanospheres. The as- prepared Gd(OH)₃: Yb³⁺, Er³⁺ powder consists of uniform nanospheres of about 100-nm diameter. Figure 1(b) shows the TEM image of the above-mentioned Gd(OH)₃: Yb³⁺, Er³⁺ nanospheres. It is worth noting that as-synthesized Gd(OH)₃: Yb³⁺, Er³⁺ nanospheres are solid. The nanospheres become hollow after they are stored in the water for one month [as shown in Fig. 1(c)] by an Ostwald ripening process.Reference Dong, Ferri, Adalseinsson, Schonhoff, Sukhorukov and Mohwald⁵²

FIG. 1.

(a) SEM images of Gd(OH)₃: Yb³⁺, Er³⁺ samples, (b) TEM images of Gd(OH)₃: Yb³⁺, Er³⁺ samples, (c) TEM images of hollow Gd(OH)₃: Yb³⁺, Er³⁺ samples.

Figure 2(a) shows the SEM image of the GdOF: Yb³⁺, Er³⁺ nanosphere precursor. It is noted that uniform, monodisperse GdOF: Yb³⁺, Er³⁺ hollow nanospheres inherit the spherical morphology and good size dispersion from the Gd(OH)₃: Yb³⁺, Er³⁺. The morphology, uniform shape, and hollow structure of nanospheres were not adversely affected by the calcination process [Fig. 2(b)]. The TEM micrograph [Fig. 2(c)] shows strong contrast between the dark edges and the light center confirms that all the nanospheres are hollow. HRTEM image [Fig. 2(d)] of nanocrystalline particles shows lattice planes continuously extending in each nanocrystalline area that shows good nanocrystallinity of the particle. The distance of the lattice planes was calculated to be 0.316 nm using the HRTEM, which is consistent with the (012) planes of rhombohedral phase GdOF: Yb³⁺, Er³⁺.

FIG. 2.

(a) SEM images of GdOF: Yb³⁺, Er³⁺ samples, (b) SEM images of GdOF: Yb³⁺, Er³⁺ samples after calcination at 400 °C, (c) TEM images of GdOF: Yb³⁺, Er³⁺ samples after calcination at 400 °C, (d) HRTEM images of GdOF: Yb³⁺, Er³⁺ samples after calcination at 400 °C.

Figure 3 shows the typical x-ray diffraction patterns of the GdOF: Yb³⁺, Er³⁺ nanospherical precursor and the crystalline GdOF: Yb³⁺, Er³⁺ product obtained by annealing at 400 °C for 4 h. No obvious diffraction peaks appear in the GdOF: Yb³⁺, Er³⁺ nanospherical precursor, which indicates that as-synthesized GdOF: Yb³⁺, Er³⁺ hollow nanosphere precursor is amorphous before calcination. After calcining at 400 °C for 4 h, all of the diffraction peaks are indexed to the rhombohedral phase of GdOF [Joint Committee on Powder Diffraction Standards (JCPDS) No. 50–0569]. Furthermore, no other impurity peaks could be detected, indicating that the Yb³⁺ and Er³⁺ ions have been thoroughly doped into the GdOF host lattice. The peak positions of the GdOF: Yb³⁺, Er³⁺ nanospheres do not shift because these doping ions have similar ionic radii as Gd³⁺ (94 pm). Moreover, no additional peaks of other phases have been found, indicating the formation of pure rhombohedral GdOF: Yb³⁺, Er³⁺ phase.

FIG. 3.

The x-ray powder diffraction patterns of GdOF: Yb³⁺, Er³⁺ samples: (a) GdOF: Yb³⁺, Er³⁺ precursors, (b) GdOF: Yb³⁺, Er³⁺ after calcination at 400 °C.

B. Formation mechanism of GdOF hollow nanospheres

The growth process of the GdOF: Yb³⁺, Er³⁺ hollow nanospheres from Gd(OH)₃: Yb³⁺, Er³⁺ was investigated in detail at different reaction temperatures and times. Figure 4 is a series of TEM images of the samples heated at different reaction temperatures for 4 h. No hollow product was obtained at 120 °C [Fig. 4(a)]. The particles start to become hollow at center and the shell becomes thinner as temperatures rises to 180 °C [Figs. 4(b)–4(d)]. Most of the hollow spheres break down when the reaction temperature reaches 200 °C (as seen in Fig. S1). From the above systematic TEM analysis, it can be understood that the formation of GdOF: Yb³⁺, Er³⁺ hollow nanospheres occurs by the Kirkendall mechanism because of competitive atomic diffusions between two phases upon the hydrothermal treatment as related before.Reference Chiang and Chiang⁶⁶

FIG. 4.

A series of TEM images of the samples obtained at different reaction temperatures: (a) 120 °C, (b) 130 °C, (c) 150 °C, and (d) 180 °C for 4 h.

Figure 5 shows a series of TEM images of the samples obtained from different reaction time when heated at 150 °C. After 1.0–4.0 h, the products become hollow. However, when the reaction time exceeds 16 h by heated at 150 °C, most of hollow spheres break down (seen in Fig. S2).

FIG. 5.

A series of TEM images of the samples investigated with different reaction times at 150 °C: (a) 0 h, (b) 1 h, and (c) 4 h.

In general, the starting materials have a great effect on the morphology of the final products. Therefore, a series of experiments with different fluorine sources such as ammonium fluoride (NH₄F), and hydrofluoric acid (HF) as starting materials were performed while keeping other parameters constant. The results show that the different fluorine sources determine the morphology and structure of the final products. As shown in Fig. S3(a), when the reaction was performed in the presence of NH₄F, no hollow spherical particles were obtained. The XRD pattern of the crystalline sample after calcination shows that the powder is the cubic phase of gadolinium (III) oxide (Gd₂O₃) (JCPDS No. 86-2477). It is obvious that Kirkendall process does not take place under this condition [Fig. S3(b)]. When HF acid of different amounts such as 20, 40, 60, and 100 μL was used as fluorine source, ytterbium ion and erbium ion codoped gadolinium fluoride (GdF₃: Yb³⁺, Er³⁺) powders with different morphology were obtained as seen in Fig. S4. It indicates that the uniform and monodisperse Gd(OH)₃: Yb³⁺, Er³⁺ solid nanospheres reaction to form GdF₃: Yb³⁺, Er³⁺ when a trace amount of HF acid (20 μL) was introduced into the reaction system. When HF was increased to 100 μL, the hollow microplate product with diameters in the range of 0.5–1.0 μm was observed. Figure S4(e) displays a series of XRD patterns of the as-synthesized samples with 20, 40, 60, and 100 μL HF. It indicates that the uniform and monodisperse Gd(OH)₃: Yb³⁺, Er³⁺ phase gradually disappears, and GdF_3: Yb³⁺, Er³⁺ phase gradually appears when HF acid was introduced into the reaction system.

At this point, an interesting question arises, namely, how these different fluorine sources impact the morphology and structure of final products? It can be speculated that the appropriate amount of H⁺ can play an important role in the reaction process. There was too low H⁺ (pH = 4.83) when the NH₄F was used as fluorine source. So, the Gd(OH)₃: Yb³⁺, Er³⁺ would not dissolve and no GdOF: Yb³⁺, Er³⁺ was formed. However, the dissolution of Gd(OH)₃: Yb³⁺, Er³⁺ had become dominant compared with the growth of GdOF: Yb³⁺, Er³⁺ when the HF acid was used as fluorine source, so that the uniform and monodisperse spherical morphology of Gd(OH)₃: Yb³⁺, Er³⁺ would not be preserved.Reference Zhang, Shi, Sun, Zhao and Stucky⁴⁵ The experiments prove that the uniform and monodisperse hollow GdOF: Yb³⁺, Er³⁺ nanospheres can be obtained only when the TiF₄ was the fluorine source, as described below.

In our reaction system, F⁻ anions can be hydrolyzed to form a small concentration of HF in the hydrothermal environment and Ti⁴⁺ ions can react with the OH⁻ to form TiO₂ and H⁺. Once the hydrogen ions (H⁺) were formed, they would react with F⁻ anions once again to give rise to more HF in this process. Subsequently, the formed appropriate amount of HF would react with the Gd(OH)₃: Yb³⁺, Er³⁺ to produce the GdOF: Yb³⁺, Er³⁺ and hold the former spherical morphology. Previous studies have demonstrated that the Kirkendall processReference Fujihara, Koji and Kimura^42–Reference Antic-Fidancev, Holsa, Krupa and Lastusarri⁴⁴ will take place during this stage because the equilibrium will be set up between the dissolution of Gd(OH)₃: Yb³⁺, Er³⁺ and growth of GdOF: Yb³⁺, Er³⁺. These appropriate amounts of GdOF: Yb³⁺, Er³⁺ will form hollow sphere driven by the minimization of the total energy of the system. The titanium element is present as TiO₂ clusters with positive static potential. The whole reaction can be described in accordance with following process:

(1)$${\rm{F}}^{\rm{ - }} \,{\rm{ + }}\,{\rm{H}}_{\rm{2}} {\rm{O}}\, \to \,{\rm{HF}}\,{\rm{ + }}\,{\rm{OH}}^{\rm{ - }} {\rm{\quad,}}$$

(2)$${\rm{Ti}}^{{\rm{4 + }}} \,{\rm{ + }}\,{\rm{OH}}^{\rm{ - }} \, \to \,{\rm{TiO}}_{\rm{2}} \,{\rm{ + }}\,{\rm{H}}^{\rm{ + }} {\rm{\quad,}}$$

(3)$${\rm{F}}^{\rm{ - }} \,{\rm{ + }}\,{\rm{H}}^{\rm{ + }} \, \to \,{\rm{HF\quad,}}$$

(4)$${\rm{Gd(OH)}}_{\rm{3}} \,{\rm{ + }}\,{\rm{HF}}\, \to \,{\rm{GdOF}}\,{\rm{ + }}\,{\rm{2H}}_{\rm{2}} {\rm{O}}{\rm{\quad.}}$$

On the other hand, it is well known that the dissolution rate of Gd(OH)₃: Yb³⁺, Er³⁺ and growth rate of GdOF: Yb³⁺, Er³⁺ will be enhanced with increasing reaction temperature. Therefore, the void is enlarged by increasing the reaction temperature, which is in good agreement with the TEM results. The hollow process preferentially takes place in the center of the nanospheres as reported in previous reference literature.Reference Pang, Wang and Zeng⁶⁸ For validating our conjecture, we measured the zeta potentials of the Gd(OH)₃: Yb³⁺, Er³⁺ precursor, residual solution, GdOF: Yb³⁺, Er³⁺ hollow nanospheres, and the ICP analysis of the residual solution. The results show that the residual solution contains about 0.038M Ti, which is in agreement with the molar concentration of TiF₄ (0.04 M). It indicates that the titanium element remains in the residual solution. The measured zeta potentials are +10 mv for Gd(OH)₃: Yb³⁺, Er³⁺ and +26.5 mv (130 °C), +28.0 mv (150 °C), and +34.2 mv (180 °C) for GdOF: Yb³⁺, Er³⁺, respectively. The results show that the Gd(OH)₃: Yb³⁺, Er³⁺ solid nanospheres and GdOF: Yb³⁺, Er³⁺ hollow nanospheres have a positive static potential. It qualitatively suggests that the TiO₂ cluster with positive static potential will be excluded from surface of GdOF: Yb³⁺, Er³⁺ nanospheres because of the static repulsion and remain in the residual solution. All above-mentioned analyses clearly corroborate that the Kirkendall process is the causal factor in the present system. A schematic illustration for the effect of different fluorine sources for the final product and the formation procession of the GdOF: Yb³⁺, Er³⁺ uniform, monodisperse hollow nanosphere at different reaction temperature is presented in Fig. 8.

High-resolution x-ray photoelectron spectra (XPS) were measured to characterize the surface electronic structure and chemical element valence by referencing the C1s line to 284.5 eV, as shown in Fig. 6. The XPS spectra reveal that all the products contain Gd, Yb, and Er as seen in Fig. 6(A).Reference Petoral, Soderlind, Klasson, Suska, Fortin, Abrikossova, Selegard, Kall, Engstrom and Uvdal^69, Reference Milanov, Toader, Parala, Barreca, Gasparotto, Bock, Becker, Ngwashi, Cross, Paul, Kunze, Fishcher and Devi⁷⁰ The bonding energy (BE) at 685.5 eV measured in the samples of GdOF: Yb³⁺, Er³⁺ and GdF₃: Yb³⁺, Er³⁺ can be attributed to the F. However, there is no F in the Gd₂O₃: Yb³⁺, Er³⁺ sample. Figure 6(B) displays the corresponding O1s spectra of Gd₂O₃: Yb³⁺, Er³⁺, GdOF: Yb³⁺, Er³⁺, and GdF₃: Yb³⁺, Er³⁺, respectively. The high BE component at 528.6 and 529.4 eV is due to the presence of lattice O^2- on the surface of the samples of Gd₂O₃: Yb³⁺, Er³⁺ and GdOF: Yb³⁺, Er³⁺, respectively.Reference Pang, Wang and Zeng⁶⁸ The second peak at 531 eV can be attributed to the presence of a hydroxyl group and/or carbonate group for the samples of Gd₂O₃: Yb³⁺, Er³⁺ and GdOF: Yb³⁺, Er³⁺, respectively.Reference Du, Zhang, Yan, Sun and Yan⁴¹ The final peaks at 532.9, 532.4, and 532.6 eV can be assigned to molecular water adsorbed on the surface of Gd₂O_3: Yb³⁺, Er³⁺, GdOF: Yb³⁺, Er³⁺, and GdF_3: Yb³⁺, Er³⁺ products, respectively. These results once again show that different fluorine sources have a great effect on the morphology, component and structure of the final products and match with the above-mentioned SEM and TEM results very well.

FIG. 6.

The XPS spectrum of as-obtained samples: (A) the survey XPS spectra of F 1s, O 1s, C 1s, Gd 4d, Yb 4d, and Er 4d for different of samples: (a) Gd₂O₃: Yb³⁺, Er³⁺, (b) GdOF: Yb³⁺, Er³⁺, and (c) GdF₃: Yb³⁺, Er³⁺, (B) detailed XPS spectrum of O 1s for the Gd₂O₃: Yb³⁺, Er³⁺ nanosphere, the GdOF: Yb³⁺, Er³⁺ hollow nanosphere, the GdF₃: Yb³⁺, Er³⁺ microplate, respectively.

C. Luminescence properties

It is well known that GdOF is a good host matrix for UC luminescenceReference Du, Zhang, Yan, Sun and Yan⁴¹ due to its chemical durability and lower phonon energy. As reported in reference literature, the rare earth activator ion can be effectively doped into the lanthanide oxides, lanthanide fluoroxides, and lanthanide fluorides host matrices. In this contribution, a stoichiometric amount of rare earth ions (Yb³⁺ and Er³⁺) were doped into the Gd₂O₃, GdOF, and GdF₃ host lattices to investigate the UC luminescence properties, respectively. We chose the GdOF:18% Yb³⁺/2% Er³⁺ hollow nanospheres that were heated at 150 °C as a typical sample because the normalized UC emission spectra of GdOF:18% Yb³⁺/2% Er³⁺ with different shell thicknesses are similar under 980 nm light excitation. Figure 7 shows the normalized UC emission spectra of Gd₂O₃:18% Yb³⁺/2% Er³⁺, GdOF:18% Yb³⁺/2% Er³⁺, and GdF₃:18% Yb³⁺/2% Er³⁺ under 980 nm light excitation. It is well known that codoping not only increases the intensity of luminescence but also induces UC luminescence between the donor and acceptor ions.Reference Milanov, Toader, Parala, Barreca, Gasparotto, Bock, Becker, Ngwashi, Cross, Paul, Kunze, Fishcher and Devi⁷⁰ In our work, Yb³⁺ was chosen as the codopant with Er³⁺ because it has a large absorption cross section at 980 nm and energy transfer (ET) occurs as a result of the large spectral overlap between the Yb³⁺ emission ²F_5/2-²F_7/2 and the Er³⁺ absorption ⁴I_11/2-⁴I_15/2 bands. However, for the samples Gd₂O₃:18% Yb³⁺/2% Er³⁺, GdOF:18% Yb³⁺/2% Er³⁺, and GdF₃:18% Yb³⁺/2% Er³⁺, the emission spectra are quite different. The UC emission color can be adjusted from red to yellow to green when the host matrix was varied from Gd₂O₃: 18% Yb³⁺/2% Er³⁺ to GdOF: 18% Yb³⁺/2% Er³⁺ to GdF₃: 18% Yb³⁺/2% Er³⁺ with all other parameters remaining constant. We can see that the obvious green emissions from 500 to 600 nm regions, which correspond with the ²H_11/2-⁴I_15/2 and ⁴S_3/2-⁴I_15/2 transitions of Er³⁺, and the red lines from 600 to 700 nm can be attributed to the transition of ⁴F_9/2-⁴I_15/2. The enhanced relative intensity of green emissions to red emissions can be attributed to the attenuation of the nonradiative transition rate when the host lattices change from Gd₂O₃ to GdOF to GdF₃.Reference Vetrone, Boyer, Capobianco, Speghini and Bettinelli^71–Reference Bai, Song, Pan, Lei, Wang, Ren, Lu, Dong, Dai and Fan⁷³

FIG. 7.

Normalized emission spectra (λ_ex = 980 nm) in different samples: (a) Gd₂O_3: Yb³⁺, Er³⁺, (b) GdOF: Yb³⁺, Er³⁺, and (c) GdF_3: Yb³⁺, Er³⁺.

FIG. 8.

Illustration of the effect of different fluorine sources for the final product and the formation procession of the GdOF: Yb³⁺, Er³⁺ uniform, monodisperse hollow nanosphere at different reaction temperatures.