Uniform, luminescent Eu:LuF3 nanoparticles

1 Uniform, luminescent Eu:LuF3 nanoparticles. Ana I. Becerro,* Daniel González-Mancebo, Manuel Ocaña Instituto de Ciencia de Materiales de Sevilla (CSIC-University of Seville). c/ Américo Vespucio, 49. 41092 Seville (Spain). * Corresponding author. E-mail address: [email protected]. Tel no: +34 95448945. Fax no: 954460165 2 Abstract A simple procedure for the synthesis orthorhombic, uniform, LuF3 particles with two different morphologies (rhombus- and cocoon-like) and nanometer and sub-micrometer size, respectively, is reported. The method consists in the aging, at 120ºC for 2 h, a solution containing [BMIM]BF4 ionic liquid (0.5 mL) and lutetium acetate (in the case of the rhombi) or lutetium nitrate (in the case of the cocoons) (0.02 M) in ethylene glycol (total volume 10 mL). This synthesis method was also adequate for the synthesis of Eu3+-doped LuF3 particles of both morphologies, whose luminescence properties were investigated in detail. The experimental observations reported herein suggest that these materials are suitable phosphors for optoelectronic as well as in-vitro biotechonological applications. Keywords: LuF3, nanoparticles, ionic liquid, luminescence, concentration quenching. 3 Introduction Lanthanide-doped rare earth (RE) fluoride matrices such as REF3 and NaREF3 have become a research focus in the optical materials field owing to their unique applications in optical telecommunication, lasers, and biochemical probes (Downing et al. 1996; Yi et al. 2004; Wang et al. 2006). In comparison with oxygen-based systems, fluorides possess very low vibrational energies and therefore the quenching of the excited states of the lanthanide ions is minimal (Diamente et al. 2007). In addition, RE fluorides exhibit high thermal and chemical stability, which make them ideal for luminescent purposes. It is highly desirable, for many of the mentioned applications, to have uniform nano- or submicronparticles with controlled size and shape, as well as good dispersion (Yi et al. 2004; Wang et al. 2006; Sun et al. 2007). A number of highly monodisperse, nanometer and submicrometer size YF3, LaY3 and GdF3 particles have been reported in the literature with very promising results for their use as both upconversion and down conversion phosphors (Mai et al. 2006; Wang et al. 2007; Rao et al. 2014; Carniato el at. 2013). However, fewer reports can been found on the synthesis of uniform Lu-based fluoride hosts in spite of their excellent luminescence and even better laser performance than their Y-based counterparts ( Maunier et al. 2002; Sudesh and Asai, 2003; Liu et al. 2011; Li et al. 2011). In particular, most papers published about the synthesis of LuF3 report either heterogeneous, polydisperse submicroparticles (Xiao et al. 2007; Li et al. 2008; Lin et al. 2013) or aggregated nanoparticles (Qiu et al. 2014). Only Li et al. (2008) reported a synthesis method adequate to obtain uniform LuF3 particles of micrometer-sized. In all these cases, the synthesis methods used involve either relatively high temperatures (≥ 180 ºC) or/and the use of toxic surfactants (DDBAC) or complexing agents (EDTA). Herein, we report a really simple (it does not require the use of toxic additives) low temperature (120ºC) route for the synthesis of uniform LuF3 particles, through controllable release of F- anions from an ionic liquid, with two different morphologies (rhombus- and cocoon-like shape) and narrow size distribution. We have investigated the influence of different Lu and F precursors, their concentrations, solvents, aging temperature and heating method on the morphology of the synthesized particles. The same method used to obtain uniform LuF3 particles has also been successfully applied 4 to obtain Eu3+-doped LuF3 particles of both morphologies, whose luminescence properties have been investigated in detail. Experimental section 1. Synthesis of samples The procedure for the synthesis of the particles was as follows. Rhombus-like LuF3 particles were obtained using lutetium acetate hydrate (0.02 M Lu(CH3COO)3 · xH2O, Sigma Aldrich, 99.9%) as the lutetium source, while lutetium nitrate hydrate (0.02 M Lu(NO3)3 · xH2O, Sigma Aldrich, 99.99%) was necessary for the synthesis of cocoon LuF3 particles. In both cases, the lutetium salt was added to ethylene glycol (EG, C2H6O2, Fluka, 99.5%), keeping magnetic stirring and heating at 70 ºC to favor dissolution. The solutions were then cooled down to room temperature and admixed with 0.5 cm3 of 1-butyl-2-methylimidazolium tetrafluoroborate ([BMIM]BF4, C8H15BF4N2, Aldrich ≥ 97 %). After homogenization, the final solutions (total volume = 10 cm3) were aged for 2 h in tightly closed test tubes using an oven preheated at 120 ºC. The resulting dispersions were cooled down to room temperature, centrifuged to remove the supernatants and washed, twice with ethanol and once with double distilled water. For some analyses, the powders were dried at room temperature. The Eu3+-doped particles were synthesized with the same procedure using Eu3+ acetate hydrate (Eu(CH3COO)3 · xH2O, Sigma Aldrich, 99.9%) and Eu3+ nitrate hydrate (Eu(NO3)3 ·xH2O, Sigma Aldrich, 99.9%) in stoichiometric amounts. The lanthanide ions (Lu + Eu) concentration was kept constant (0.015 mol·dm−3) in all experiments, whereas the Eu/(Lu + Eu) molar ratio was varied from 0.5% to 40% in order to investigate the effect of this parameter on the morphological and luminescent properties of the precipitated particles. 2. Characterization techniques The shape and size of the particles was examined by both transmission electron microscopy (TEM, Philips 200CM) and scanning electron microscopy (SEM-FEG Hitachi S4800). Particle size distributions were obtained from the micrographs by counting several hundreds of particles. Additional information on the size and colloidal stability of the particles in aqueous suspension (0.5 mg·mL-1 of solid) was obtained 5 from Dynamic Light Scattering (DLS) measurements. The experiments were carried out using a Malvern Zetasizer Nano-ZS90 equipment, which was used as well to measure the zeta potential of the suspensions. The crystalline structure of the prepared particles was assessed by X-ray diffraction (XRD) using a Panalytical, X´ Pert Pro diffractometer (CuKα) with an X-Celerator detector over an angular range of 10º < 2θ < 70º, 2θ step width of 0.05º, and 10 s counting time. The crystallite size was calculated using the Scherrer formula from the full width at half maximum of several single reflections. Unit cell parameters of the undoped and the Eu3+-doped particles were calculated from the XRD patterns using the Rietveld method with the TOPAS software. Starting crystallographic parameters were taken from those reported for orthorhombic LuF3 in ICDD PDF 00-032-0612. Nominal Eu3+ contents were added to the structure. Refined parameters were: scale factor, zero error, background coefficients, unit cell parameters and atomic coordinates and atomic displacement factors of the lanthanide atoms. The infrared spectra of the powders, diluted in KBr pellets, were recorded in a Jasco FT-IR-6200 Fourier transform spectrometer. Finally, the excitation and emission spectra of the Eu3+-doped LuF3 particles, dispersed in water (2.5 mg·mL-1), were measured in a Horiba Jobin Yvon spectrofluorimeter (Fluorolog3). The emission spectra were transformed to the CIE color coordinates system using a 2º observer. Results and discussion It is well known that uniform particles can be obtained by precipitation through a slow and controlled release of the precipitating anions or cations in the reaction medium (Matijevic, 1993). We have used, in the present study, a strategy based on the controlled release of fluoride anions using an ionic liquid ([BMIM]BF4) as fluoride precursor and two different lutetium salts (lutetium acetate hydrate and lutetium nitrate hydrate), which provide different particle morphology. The required fluoride anions are liberated by the ionic liquid as a consequence of its hydrolysis with the hydration water molecules of the lutetium precursor (Jacob et al. 2006). Ionic liquids are non-volatile, non-flammable and thermally stable organic salts with low melting point, which show a 6 superior capability for the solvation and stabilization of metal ions, which gives to them the possibility of acting as capping agents or surfactants. Ethylene glycol (EG) was selected as solvent because polyols have been amply shown to be suitable solvents for the synthesis of nanomaterials, as they may act not only as solvent but also as capping agent, thus limiting particle growth (Feldmann, 2003). 1. Morphology, colloidal stability and crystal structure of the precipitated LuF3 particles. Figure 1 (top) shows the TEM micrographs of the particles obtained after aging, at 120 ºC in a conventional furnace for 2 hours, a solution containing lutetium acetate (0.02 M) and [BMIM]BF4 in EG. Using the same experimental conditions but changing the lutetium source to lutetium nitrate (0.02 M) led to the particles shown in the TEM micrographs of Figure 1 (bottom), which only needed 30 minutes aging to reach their final morphology. These experimental conditions allowed the synthesis, in both cases, of particles, homogeneous in both shape and size, and with a high degree of dispersion. The morphology of the particles is shown in detail in the insets of each of the TEM micrographs. As observed, using lutetium acetate as the lutetium source led to the synthesis of rhombus-like particles with two rounded and two sharp vertices. Their long axis length was ~180-200 nm (standard deviation = 30) while the short axis length was ~100 nm (standard deviation = 20). However, when lutetium nitrate was used as the lutetium precursor, cocoon-like particles (300 (30) nm x 150 (20) nm) were obtained. Although the TEM micrographs shown in Figure 1 suggested that the particles were uniform and non-aggregated, this aspect was further confirmed by dynamic light scattering (DLS) measurements conducted on an aqueous suspension of both types of particles (pH 6.8). The average hydrodynamic diameter obtained by DLS was 190 nm for the rhombus-like particles and 220 nm for the cocoons (Figure 2). These data confirm the absence of particles aggregation and reveal that the particles are stable in an aqueous medium at close to neutral pH. The zeta potential values calculated on the same suspensions were 46.9 mV and 46.3 mV for the rhombus- and cocoon-like particles, respectively, which suggest that the colloidal stability of the suspensions was mainly due to the electrostatic repulsion among positively charged particles. The crystalline structure of both types of the precipitated particles was assessed from their X-ray diffraction patterns (Figure 3, top). Both patterns match the standard PDF 7 ICDD 00-32-0612 corresponding to LuF3, indicating that in both cases the LuF3 particles exhibit an orthorhombic symmetry with space group Pnma. The width of the reflections is clearly different from one pattern to the other, the cocoon particles showing the wider reflections. As a consequence, the crystalline domain size in the cocoon particles (calculated using the Scherrer formula on single reflections) was of 10 nm while that of the rhombus-like particles was of 15-20 nm. If one compares these values with the particles dimensions in any of both cases, it can be easily concluded that the particles were not single crystals but they were formed by the aggregation of small subunits. This fact will be analyzed in more detail later in this paper. FTIR spectroscopy was used to gain additional information on the samples composition. The spectra of both types of LuF3 particles (Figure 3, bottom) are very similar to each other and exhibit the absorptions associated with fluoride ions (< 1000 cm-1), along with broad bands at about 3400 cm-1 and 1640 cm-1 due to absorbed water. The weak features in the 1500-1300 cm-1 region are due to carbonate anions coming from the adsorption of atmospheric CO2. Since no additional impurities could be observed in the FTIR spectra of the samples, the presence of EG species on the particles surface, previously observed for other Ln-doped particles prepared in EG (Singh et al. 2012), can be discarded. 2. Influence of the synthesis conditions on the morphology of the precipitated particles. The set of experimental conditions mentioned above is essential to obtain uniform LuF3 particles. The conditions for the synthesis of the LuF3 rhombus- and cocoon-like particles were so restrictive that keeping the lutetium source constant (as lutetium acetate or lutetium nitrate) and changing, one by one, each of the other experimental parameters (fluoride source, lutetium concentration, ionic liquid concentration, type of solvent, aging temperature and heating source), led to ill-defined, heterogeneous, aggregated or bigger particles, depending on the parameter changed, as described in detail in Table 1 and observed in Figure 1SI and 2SI. 3. Formation mechanism of the LuF3 particles To elucidate the mechanism for the formation of the LuF3 particles, time-dependent experiments were carried out in the synthesis conditions of the rhombus-like and the cocoon particles. TEM images of the corresponding intermediates are shown in Figures 4a and 4b, respectively. Precipitation of three types of particles, consisting of ribbons, 8 100 nm elongated particles and tiny ill-defined particles (< 10 nm), was observed in both cases in the first stages of the reaction (10 min). Increasing reaction time (15 min) led to the increase in the size of the elongated particles as well as to the decrease in the number of ribbons and tiny particles, the latter being practically missing in the case of the cocoon series. These facts suggest that the elongated particles were formed by the dissolution of the ribbons and tiny particles followed by re-precipitation. However, it cannot be discarded that the tiny particles served as primary nanobuilding units to reassemble and rearrange in a three-dimensional array to form the elongated particles. In fact, the XRD data described above, point to the latter mechanism. The elongated particles already resembled, at this stage of the reaction (15 min), the final rhombus and cocoon shapes. Finally, while the cocoon particles were fully formed after 30 minutes reaction, the rhombus-like particles needed a much longer reaction time (120 min) to complete formation. This observation suggest that the lutetium acetate salt used for the synthesis of the LuF3 rhombi acted as capping agent, thus delaying the release of the Lu3+ ions and, therefore, the formation of the particles. Such delay was not possible in the case of the Lu(NO3)3 used for the synthesis of the cocoons, which, consequently, formed faster. Finally, the XRD patterns registered at each reaction stage (Figure 4c and 4d) indicated that no symmetry change took place with reaction time. The patterns just showed the crystallization of orthorhombic LuF3, which occurred after 10 minutes of reaction in both cases. The proposed reaction mechanism can be used to understand the change in the morphological characteristics of the LuF3 particles precipitated under different experimental conditions, as described in the previous section (Table 1). The ionic strength (given by the reactant concentrations) and the dielectric constant of the solvent (42.5, 37 28.8 for glycerol, ethylene glycol, and butylene glycol respectively, at 25 ºC, respectively) determine the equilibrium between attractive and repulsive forces in a colloidal system, thus controlling the aggregation process (Privman et al. 1999). 4. Synthesis, structural characterization and luminescent properties of Eu3+-doped LuF3 rhombus and cocoon particles. Eu3+ doping was used, as a proof on concept, to analyze the luminescent properties of the resultant LuF3 rhombus- and cocoon-like phosphors. The synthesis method used to 9 obtain the undoped particles was also successfully applied to obtain LuF3 particles of both morphologies doped with Eu3+. Different doping levels were used to evaluate the effect of the Eu3+ content on the luminescent properties of the phosphors. Nominal concentrations were 2.5, 5.0, 10, 20, 30 and 40 mol % Eu3+ in LuF3. It was found that the same shape and crystal symmetry as those of the undoped materials were reproduced for the doped samples for all doping levels in both, the rhombus and cocoon particles. However, a significant decrease in particle size (Table 2) was observed when the doping level increased from 20% to 30% in the case of the rhombus-like particles and from 30% to 40% in the case of the cocoon-like particles (Figure 5). Changes in particle size were observed as well by Li et al. (2008) in their study of lanthanide fluorides LnF3 (Ln=La to Lu) crystals. To confirm the presence of Eu3+ in the LuF3 crystal structure, that is, replacing Lu3+ in their crystallographic sites, the unit cell volumes corresponding to the Eu-doped LuF3 rhombus- and cocoon-like particles were calculated by the Rietveld analysis of their XRD patterns (Figure 3SI). The values obtained have been plotted in Figure 6 for both types of particles. The cell volume increased with increasing Eu3+ content, in agreement with the bigger ionic radius of Eu3+ compared with that of Lu3+ (Shannon, 1976). In addition, the increase in unit cell volume was linear, satisfying the Vegard´s law (Vegard, 1921) and indicating, therefore, the formation of a solid solution of Eu3+ in LuF3 for both types of morphologies. To evaluate the luminescent properties of the Eu3+-doped LuF3 particles, their excitation spectra were recorded by monitoring the emission of the characteristic 5D0-7F1 transition of Eu3+ at 593 nm. Figure 7a shows the excitation spectrum corresponding to the 10% Eu3+-doped LuF3 rhombus-like particles. The rest of compositions as well as the Eudoped LuF3 cocoon-like particles showed very similar excitation spectra to this one. The spectrum showed a set of sharp and well defined excitation bands, the most intense at 393 nm, which correspond to the f-f electronic transitions characteristic of the Eu3+ ions (Yan and Li, 2005) and that have been indexed in the figure. The emission spectra of the Eu3+-doped LuF3 rhombus particles (molar % Eu3+ from 0.5% to 30%) recorded after excitation at 393 nm, are shown in Figure 7b. All of them showed the emission lines corresponding to the well-known 5D0-7FJ (J= 0, 1, 2, 3, 4) transitions of Eu3+. The 5D0-7F1 transition is due to the magnetic dipole transition of 16 Table 3: Asymmetry ratio (R) of the 5D0-7F2 transition to the 5D0-7F1 transition as a function of Eu3+ content. Molar % Eu3+ R (Rhombi) R (Cocoons) 10 0.22 0.26 20 0.26 0.28 30 0.31 0.28 40 0.33 0.36 17 FIGURE CAPTIONS Figure 1: TEM micrographs of the LuF3 particles synthesized after aging at 120 ºC a solution containing lutetium acetate (a) and lutetium nitrate (b) and [BMIM]BF4 in ethylene glycol. Aging time: 2h and 30 min in a and b, respectively. The insets show a magnified image particle of each particle type. Figure 2: Size distribution of the two types of particles shown in Figure 1 calculated by Dynamic Light Scattering. Figure 3: Top: XRD patterns of the two types of particles shown in Figure 1. The PDF file shown at the bottom corresponds to orthorhombic LuF3. Bottom: FTIR spectra of both types of particles. Figure 4: a and b) TEM micrographs of three different intermediates in the formation of the rhombus- and cocoon-like LuF3 particles. c and d) XRD patterns corresponding to the intermediates shown in a and b, respectively. Figure 5: TEM micrographs of the Eu3+-doped LuF3 rhombus- (top) and cocoon- (bottom) like particles with different Eu contents. Figure 6: Unit cell volume of the Eu:LuF3 rhombus- and cocoon-like particles as a function of Eu3+ concentration calculated by the Rietveld analysis of their corresponding XRD patterns (the latter shown in Figure 3SI). The error bars are approximately the size of the symbols. Figure 7: a) Excitation spectrum of the 5%Eu-doped LuF3 rhombus-like particles recorded at the emission wavelength of 588 nm. b) Emission spectra of different Eucontaining LuF3 rhombus like particles recorded at an excitation wavelength of 394 nm. c) CIE diagram showing the chromatic coordinates of the rhombus- (square) and cocoon- (circle) like Eu:LuF3 particles. d) Integrated area of the emission spectra of the rhombus- and cocoon-like Eu:LuF3 particles as a function of Eu3+ concentration. 18 Figure 1 19 Figure 2 10 100 1000 cocoons rhombi Intensity (a.u.) Hydrodynamic diameter (nm) 20 Figure 3 20 30 40 50 60 70 Intensity (a.u.) 2 Cocoons Rhombus PDF 00-032-0612 4000 3500 3000 2500 2000 1500 1000 500 Wavenumber (cm-1) Intensity (a.u.) Cocoons Rhombi 21 Figure 4 10 min 15 min 120 min 5 0 0 n m 15 min 30 min 10 min a ) b ) 20 30 40 50 60 70 Intensity (a.u.) 2 15 min 30 min PDF ICCD: 032-0612 5 min 10 min 120 min 20 30 40 50 60 70 Intensity (a.u.) 2 15 min 30 min PDF I 00-032-0612 5 min 10 min c d 22 Figure 5 20% 30% 30% 40% 23 Figure 6 010 20 30 40 188 189 190 191 192 193 194 195 Cocoons Rhombi Unit Cell Volume Å3 % Eu3+ in LuF3 24 Figure 7 575 600 625 650 675 700 725 30% Eu 20% Eu 10% Eu 5% Eu 2.5% Eu 0.5% Eu Wavelength (nm) Intensity (a.u.) 5D0 7F1 5D0 7F2 5D0 7F4 % Eu3+ in LuF3 5D0 7F3 250 300 350 400 450 7F0 5H8 Intensity (a.u.) Wavelength (nm) 7F0 5L6 7F0 5L7 7F0 5D4 7F0 5G2 7F0 5F4 7F0 5I6 0 5 10 15 20 25 30 35 40 Rhombus-like particles Cocoons Integrated area (540-750 nm) % Eu in LuF3 a b d c