LiYF4:Yb/LiYF4 and LiYF4:Yb,Er/LiYF4 core/shell nanocrystals with luminescence decay times similar to YLF laser crystals and the upconversion quantum yield of the Yb,Er doped nanocrystals

We developed a procedure to prepare luminescent LiYF4:Yb/LiYF4 and LiYF4:Yb,Er/LiYF4 core/shell nanocrystals with a size of approximately 40 nm revealing luminescence decay times of the dopant ions that approach those of high-quality laser crystals of LiYF4:Yb (Yb:YLF) and LiYF4:Yb,Er (Yb,Er:YLF) with identical doping concentrations. As the luminescence decay times of Yb3+ and Er3+ are known to be very sensitive to the presence of quenchers, the long decay times of the core/shell nanocrystals indicate a very low number of defects in the core particles and at the core/shell interfaces. This improvement in the performance was achieved by introducing two important modifications in the commonly used oleic acid based synthesis. First, the shell was prepared via a newly developed method characterized by a very low nucleation rate for particles of pure LiYF4 shell material. Second, anhydrous acetates were used as precursors and additional drying steps were applied to reduce the incorporation of OH− in the crystal lattice, known to quench the emission of Yb3+ ions. Excitation power density (P)-dependent absolute measurements of the upconversion luminescence quantum yield (ΦUC) of LiYF4:Yb,Er/LiYF4 core/shell particles reveal a maximum value of 1.25% at P of 180 Wcm−2. Although lower than the values reported for NaYF4:18%Yb,2%Er core/shell nanocrystals with comparable sizes, these ΦUC values are the highest reported so far for LiYF4:18%Yb,2%Er/LiYF4 nanocrystals without additional dopants. Further improvements may nevertheless be possible by optimizing the dopant concentrations in the LiYF4 nanocrystals.

In addition, the shell has to be sufficiently thick to reduce surface quenching of the luminescence as much as possible. Shell forming reactions with very low nucleation rates are therefore advantageous, since they reduce the loss of shell material due to the nucleation of new particles of pure shell material. Furthermore, not only the core-shell interface but also the core particle itself should contain only a negligible number of defects and impurity centers leading to luminescence quenching. Similar to laser crystals, the incorporation of OH − into the fluoride host lattice must therefore be avoided during the synthesis.
The aim of this work was to develop a procedure for the preparation of LiYF4:Yb/LiYF4 core/shell particles with optical properties close to those of LiYF4:Yb bulk crystals. This included measures to reduce the incorporation of OH − into the crystal lattice of LiYF4 as previously realized for NaYF4:Yb,Er/NaYF4 core/shell nanocrystals with high upconversion luminescence quantum yield [90]. Most challenging, however, was the search for a new shell forming reaction with the above mentioned properties, as the method used to form the shell of NaYF4:Yb,Er/NaYF4 core/shell nanocrystals cannot be applied to LiYF4 nanoparticles.
After comparing the optical properties of the resulting core/shell particles with those of a high-quality laser crystal of LiYF4:Yb, we prepared Yb,Er-doped particles and studied the upconversion quantum yield of LiYF4:Yb,Er/LiYF4 core/shell particles.

Chemicals and materials
Anhydrous rare earth acetates were prepared as described previously [90]. Purified oleic acid was purchased from Fisher Scientific, 1-octadecene (technical grade, 90%) from Alfa Aesar, ammonium fluoride and lithium acetate from Sigma Aldrich. All materials were used as received.

Synthesis of LiYF4
:Yb(18%) and LiYF4:Yb(18%), Er(2%) core particles 25 mL of oleic acid (HOA) and 25 mL 1-octadecene (ODE) were combined in a 250 mL three-neck round bottom flask and connected to a reflux condenser, heating mantle and temperature controller. The stirred solvent was degassed at 100 °C under vacuum (< 0.1 mbar, Schlenk-line) until the evolution of gas (water and air) came to an end. After cooling to room temperature using a water bath and switching to nitrogen flow, 0.6599 g (10 mmol) of anhydrous lithium acetate and 5 mmol of anhydrous rare earth acetates were added: 1.0642 g (4 mmol) of yttrium acetate, 0.3152 g (0.9) mmol of ytterbium acetate and 0.0344 g (0.1) mmol of erbium acetate in the case of LiYF4:Yb(18%),Er(2%) particles or 1.0908 g (4.1 mmol) of yttrium acetate and 0.3152 g (0.9 mmol) of ytterbium acetate in the case of LiYF4:Yb(18%) particles. The stirred mixture was again heated to 100 °C under vacuum and left at this temperature until the evolution of gas (acetic acid) ceased. The apparatus was refilled with nitrogen and the clear solution heated to 300 °C. Heating was stopped after 10 min and the mixture allowed to cool to 100 °C. To remove the acetic acid released during the heating step at 300 °C, vacuum was again applied and the solution degassed for one hour at 100 °C.
Subsequently, the apparatus was switched to nitrogen atmosphere and 0.9259 g (25 mmol) of dry NH4F was added at 100 °C. To remove air, the vessel was subjected three times to a short vacuum (3-5 s) and refilled with nitrogen. The mixture was then heated to 300 °C and further stirred at this temperature and in nitrogen atmosphere for two hours. After cooling to room temperature, the cloudy, slightly yellowish solution was centrifuged. The yellow supernatant was decanted from the precipitate and controlled for dissolved particles by adding the equivalent amount of ethanol. In general, no precipitation was observed unless the heating time at 300 °C was reduced from 2 h to 30 min. In this special case, the solid was separated by centrifugation for further analysis and either dried or re-dispersed in hexane. The solid content of the reaction mixture obtained by decantation of the supernatant was combined with 10 mL of hexane and vigorously shaken to extract the core particles adhering to the LiF solid formed as by-product. After centrifugation, the core particles were precipitated by adding 10 mL of ethanol to the supernatant and separated by centrifugation. Subsequently, the particles were washed by redispersing the precipitate in hexane again, adding 10 mL of ethanol, and collecting the product by centrifugation. Colloidal solutions of the core particles were prepared by immediately re-dispersing some of this product in hexane. The remaining product was dried, yielding a matte white powder. Thermogravimetric analysis indicated a solid, non-volatile content of the dry powder of about 85 wt.%-90 wt.%, the remaining 10 wt.%-15 wt.% is attributed to the organic ligand shell of the particles.

Synthesis of core/shell particles with LiYF4 shell
For the synthesis of core/shell particles, the core particles and the precursors for the LiYF4 shell were combined in a molar ratio of 1:7. The shell is therefore expected to increase the volume of the core particles by a factor of 8 and their diameter by a factor of 2. 50 mL of HOA and 50 mL of ODE were combined in a 250 ml three-neck round bottom flask and connected to a reflux condenser, heating mantle and temperature controller. The stirred solvent was degassed at 100 °C under vacuum (< 0.1 mbar) at a Schlenk-line until the evolution of gas (water and air) stopped. After cooling to room temperature and switching to nitrogen flow, 4.6557 g (17.5 mmol) of anhydrous yttrium acetate was added. The stirred mixture was again heated to 100 °C under vacuum and left at this temperature until no more acetic acid was released. The apparatus was refilled with nitrogen and the clear solution heated to 300 °C. Heating was stopped after 10 min and the mixture was allowed to cool to 100 °C. To remove the acetic acid released during the heating step at 300 °C, vacuum was again applied and the solution was degassed at 100 °C until the evolution of gas (acetic acid) was finished.
In the meantime, 16.7 mL of HOA and 16.7 mL of ODE were combined in a second set-up consisting of a 250 mL three-neck round bottom flask, reflux condenser, heating mantle, and temperature controller. The stirred solvent was degassed at 100 °C under vacuum (< 0.1 mbar) at a Schlenk-line to remove water and oxygen. After cooling to room temperature and switching to nitrogen flow, 2.5 mmol of LiYF4:Yb,(Er) core nanoparticles were added, taking into account the mass of the organic content of the particles. The stirred mixture was again heated to 100 °C under vacuum and left at this temperature until no gas evolved from the solution.
The apparatus was refilled with nitrogen and the solution heated to 300 °C. Heating was stopped after reaching 300 °C and the mixture was allowed to cool to 100 °C. Vacuum was applied again and the solution degassed at 100 °C until no further gas evolved.
After switching to nitrogen atmosphere, 2.7237 g (105 mmol) LiF and the 100 °C hot solution of yttrium oleate were added to the solution of the LiYF4:Yb,(Er) core particles at 100 °C. The solution was again degassed at 100 °C. After bubbling had stopped (max. 1 h), the apparatus was refilled with nitrogen and the solution heated to 300 °C. Heating was stopped after 6 h at 300 °C and the mixture was allowed to cool to room temperature overnight. Subsequently, the cloudy suspension was mixed with 300 mL HOA and 300 mL ODE to optimize the extraction of the particles and enhance their colloidal stability. The stirred suspension was degassed at 100 °C under vacuum (< 0.1 mbar) at a Schlenk-line until the evolution of gas (water and air) came to an end. The apparatus was refilled with nitrogen and the suspension heated to 300 °C. Heating was stopped after 10 min and the mixture allowed to cool to room temperature.
The cloudy suspension was centrifuged and treated as given above for the core particles, except that larger amounts of hexane and ethanol (40 mL each instead of 10 mL) were used for purification. In contrast to the synthesis of the core particles, most of the particles precipitated from the supernatant of the reaction mixture after the addition of ethanol. These particles were purified twice by precipitation with ethanol and by dissolving the precipitate in hexane.
For comparison, this core/shell synthesis procedure was also performed in the absence of core particles. In this case, the solution of the core particles was replaced by 33 mL of a deaerated 1:1 mixture of HOA and ODE. In contrast to the core/shell synthesis, the LiYF4 product could not be extracted from the LiF by-product with hexane, because of the very large particle size. For the same reason, the precipitate obtained by centrifugation of the reaction mixture was directly investigated without heating the precipitate in 300 mL HOA and 300 mL ODE.

Particle characterization
Transmission electron microscopy (TEM) and powder X-ray diffraction (XRD) were used to determine the mean size, the size distribution [91], as well as the crystal phase and the phase purity of all particles [92,93], respectively.

Single-crystal of LiYF4:Yb(20%)
A Czochralski-grown laser crystal of high-quality LiYF4:Yb(20%) was purchased from Optogama UAB, Lithuania. The diskshaped crystal had a thickness of 2 mm and a diameter of 8 mm.

Absolute measurement of upconversion quantum yields at different excitation power densities
The quantum yields of the upconversion emission ( UC) of the nanoparticle powders were determined absolutely at different excitation power densities (P) with a calibrated customdesigned integrating sphere setup previously reported [94]. The setup included a very stable 8 W 976 nm laser diode as excitation light source and two filter wheels with neutral density filters of known transmittance, both in the excitation channel for controlled attenuation of the excitation power density (P) in small steps. Another filter wheel with edge, band pass, and neutral density filters is placed in the detection channel to avoid detector saturation. Detection of the transmitted and emitted photons was done with a silicon CCD. The design and calibration of the setup, the beam profile characterization, and the measurement procedure were described in detail in a recent publication [94]. P-dependent  UC were obtained from P-dependent upconversion luminescence (UCL) measurements by integration over all Er 3+ emission bands between 370 and 890 nm. For the calculation of  UC, see Eq. (1), the number of emitted photons (Nem) was obtained from blank and spectrally corrected UCL spectra. The number of absorbed photons (Nabs) was derived from transmission measurements of the sample and a non-emissive blank, as previously described [94]  UC (P) = Nem/Nabs, for em < abs (1)

Time-resolved luminescence measurements
Luminescence decay kinetics were measured with an Edinburgh Instruments spectrofluorometer FLS-980 equipped with an electrically modulated 8 W 978 nm laser diode (40 μs long square pulse), a 100 mW 375 nm LED (5 μs long square pulses), and a red extended PMT (Hamamatsu R2658P). The decay curves were analyzed as previously reported [94].
If the reaction mixture was heated for a sufficiently long time, no particles could be precipitated from the supernatant by the addition of ethanol. Instead, the LiYF4 particles were extracted from the LiF solid by washing with hexane.
We minimized the incorporation of UCL-quenching OH − in the crystal lattice of the LiYF4 particles by reducing the water content of the reaction mixture. Similar to our previous synthesis procedure for NaYF4:Yb,Er/NaYF4 core/shell upconversion particles with high  UC, we used anhydrous rare earth acetates RE(OAc)3 as precursors. In addition, high temperatures of up to 300 °C were applied for drying the solvent and releasing the acetic acid HOAc (see Experimental) [90]. Moreover, we avoided the use of LiOH or polar solvents like methanol and employed anhydrous lithium acetate LiOAc as lithium source. The solution of oleates required for the synthesis of the nanocrystals was thus formed by the reaction RE(OAc)3 + 2LiOAc + 5HOA → RE(OA)3 + 2LiOA + 5HOAc↑ The above mentioned protocol was chosen because it yields undoped and doped LiYF4 particles with narrow size distribution and a mean size of approximately 20 nm. Core particles in this size range were found to yield NaYF4:Yb,Er/NaYF4 core/shell particles with the highest upconversion quantum yield [90]. The particles have the well-known square bipyramidal shape, as shown in Fig. 1(a) for Yb 3+ doped LiYF4 nanocrystals.
The small particle size indicates that a large number of LiYF4 seeds are formed during the reaction. The reaction described above is therefore well suited to prepare a large number of LiYF4 core particles with small size but not well suited to grow a LiYF4 shell.
The latter requires a reaction where only a small number of crystal seeds nucleate, because otherwise the undesired formation of new particles of pure shell material competes with the formation of a shell on the core particles. Seeds of pure shell material are a particular problem when thick shells are required for complete passivation of the core, as, in this case, a large amount of shell precursor material has to be combined with only a small number of core particles. Despite these disadvantages, the above mentioned reactions are frequently used to synthesize both, the LiYF4 particles as well as the LiYF4 shell. To nevertheless minimize the nucleation of new LiYF4 particles, the solution of the shell precursor is often added very slowly or in small portions to the solution of the core particles [78,81,88].
This time-consuming and complex procedure can be avoided by forming the shell via a reaction with a very low nucleation rate. We recently used this strategy by employing small -NaxYF3+x particles with low x as shell precursor in the synthesis of β-NaYF4:Yb,Er/NaYF4 core/shell particles [90,95]. Unfortunately, the same reaction cannot be used to form the shell of LiYF4:Yb/LiYF4 core/shell particles, since an analogue metastable phase does not exist in the LiYF4 system.

The new shell forming reaction
We found, however, that a suitable shell-forming reaction is the reaction of solid lithium fluoride with a solution of yttrium oleate in oleic acid/octadecene Y(OA)3 + 4LiF → LiYF4 + 3LiOA Similar to the reaction described above, the solution of yttrium oleate was prepared by dissolving anhydrous yttrium acetate in oleic acid/octadecene at high temperatures. In order to increase the reaction rate, we used Y(OA)3 and LiF in a molar ratio of 1 to 6 instead of the stoichiometric ratio of 1 to 4.
When this reaction takes place in the presence of LiYF4:Yb core particles, core/shell particles with narrow size distribution are obtained, as shown in Fig. 1(b) for a molar ratio of core and shell material of 1 to 7. The size histograms in Fig. 1(c) in fact show an increase in size from 22 nm of the core particles in Fig. 1(a) to 43 nm of the core/shell particles in Fig 1(b). This increase of the long axis of the square bipyramids corresponds to a shell thickness of about 5 nm, because the shell thickness d is connected to the apex angle  of the bipyramids of about 60° by d = h•sin (/2) where h equals half the increase of the long axis (see Fig. S1 in the Electronic Supplementary Material (ESM) for details).
In the absence of core particles, however, the same reaction yields very large LiYF4 particles with sizes exceeding 300 nm ( Fig. 1(d)). The diffractograms in Fig. 2 confirm complete conversion to LiYF4 after a reaction time of 6 h. The large particle size confirms that the shell reaction itself forms only a small number of crystal seeds.
The reaction is therefore not suitable for the synthesis of LiYF4 nanoparticles with a small size, but very well suited to grow a shell on core particles which act as crystal seeds in the solution. The XRD data confirm that the core and the core/shell particles, as well as the large particles in Fig. 1(d), crystallize in the tetragonal LiYF4 phase (Fig. 2 and Fig. S2 in the ESM). A closer inspection of the materials formed at early stages of the synthesis reveals, however, that LiYF4 is not the first product formed in either of the two procedures. Figure 2 displays the XRD data of the solids formed in the shell forming reaction in the absence of core particles. When this reaction is stopped already after 2 h at 300 °C, the solid in the reaction mixture consists almost entirely of LiF and only a very small amount of LiYF4 particles (Fig. 2 (b)). A second solid can be precipitated in small amount from the supernatant of the reaction mixture by adding ethanol (Fig. 2 (a)). The XRD data show that this solid consists mainly of YF3 and some LiYF4 indicating that the following reaction took place (neglecting excess LiF) Y(OA)3 + (3+x)LiF → (1−x)YF3-NPs + xLiYF4 + 3LiOA Figure 2 shows that after two hours the value of x is very small, but with increasing reaction time the amount of YF3 in the supernatant decreases while the amount of LiYF4 increases in both, the supernatant and in the solid product produced. These results indicate that the YF3 particles formed at the beginning of the reaction completely dissolve at later stages of the synthesis (x→1) and yield LiYF4 particles that partly adhere to the excess of solid LiF. It is not yet clear whether nucleation of the LiYF4 seeds takes place heterogeneously on the surface of the LiF solid or homogeneously in solution, followed by adhesion of the seeds to the LiF solid.
The first product formed in the synthesis of the doped LiYF4 core particles seems to be also lanthanide trifluoride. When the core particle synthesis is stopped after 30 min rather than after 2 h, a white powder precipitates when ethanol is added to the supernatant of the reaction mixture. The XRD data indicate that the solid consists of very small oleate capped doped YF3 particles (Fig. S3 in the ESM). After 2 h of reaction at 300 °C, no solid can be precipitated any more from the supernatant, and extraction of the LiF solid yields nothing but the doped LiYF4 particles shown in Fig. 1 (see also the XRD data in Fig. S2 in the ESM). This indicates that small doped YF3 particles are formed intermediately in the synthesis of the core particles Y(OA)3 + 2LiOA + 5NH4F → (1−x)YF3-NPs + xLiYF4-NPs + (2−x)LiF↓+ 5NH3↑+ 5HOA Similar to the shell forming reaction, the small particles dissolve at later stages of the synthesis (x → 1) leading to doped LiYF4 particles that adhere to the LiF by-product.

Yb:YLF laser crystal
To confirm the passivation of the LiYF4:Yb core particles by the inert LiYF4 shell, we determined the luminescence decay kinetics of the Yb 3+ dopant known to sensitively respond to environmental effects and the presence of quenchers [96,97]. Since our particles contain a high concentration of Yb 3+ , rapid energy migration occurs via adjacent Yb 3+ ions. In the absence of a LiYF4 shell, this leads to efficient luminescence quenching at the particle surface. Even in the presence of a LiYF4 shell, however, quenching can occur at the core/shell interface or at defects in the crystal lattice. The lifetime of the excited Yb 3+ state is therefore a good indicator of the quality of the core/shell particle. Figure 3 displays the Yb 3+ decay curves of our core and core/shell particles containing 18% Yb 3+ in the core. For comparison, also the Yb 3+ decay curve of a commercial high-quality LiYF4:Yb(20%) laser crystal was measured which Figure 3 Luminescence decay of the Yb 3+ emission of powders of monodoped LiYF4:Yb and co-doped LiYF4:Yb,Er core and core/shell particles and of a LiYF4:Yb laser crystal. The fastest decay is observed for LiYF4:Yb and LiYF4:Yb,Er core particles due to severe surface quenching. The Yb 3+ luminescence of the corresponding core/shell particles decays significantly slower than the emission of the core particles. Due to energy transfer from Yb 3+ to Er 3+ , the emission of the Yb,Er-doped core and core/shell particles decays faster than the emission of the corresponding Yb-doped particles lacking the activator Er 3+ . The decay of the LiYF4:Yb core/shell particles is already very similar to the slow decay of the laser crystal. had a size of several millimeters and was grown by the Czochralski process.
The decay times given in the figure are intensity-weighted average lifetimes derived from multiexponential fits of the luminescence decay profiles. Remarkably, the decay time of the LiYF4:Yb(18%)/LiYF4 core/shell particles of 2.1 ms is almost identical to the decay time of the laser crystal: The laser crystal shows a lifetime of the Yb 3+ emission of approximately 2.5 ms, in agreement with the specification given by the supplier of the crystal (2.1 ms).
In the absence of surface passivation, a much faster decay time of Yb 3+ of 230 μs is observed showing that the LiYF4 shell increases the Yb 3+ luminescence lifetime by a factor of 9.3. These results indicate that our improved synthesis method not only leads to a very efficient passivation by the shell but also to a low defect concentration in the particle cores and at the core/shell interface. This leads to core/shell particles with optical properties very similar to those of the bulk material.

Yb,Er doped LiYF4 core and core/shell nanocrystals
Next, we used our newly developed method to prepare upconverting nanocrystals with core particles doped with Yb 3+ and Er 3+ . The TEM images in the upper part of Fig. 4 indicate that co-doping with 2% Er has only a weak influence on the reactions, as the particle size of the core and the core/shell particles are similar to those shown in Fig. 1. This is further confirmed by the particle size histograms and the XRD data displayed in Figs. S4 and S2 in the ESM, respectively. Figure 3 also includes the decay kinetics of the Yb 3+ emission of our co-doped systems, the LiYF4:Yb,Er core and core/shell nanoparticles. The figure shows that co-doping of the particle cores with 2% of Er 3+ decreases the lifetime of the excited Yb 3+ ions in both, the core as well as the core/shell particles. This was to be expected, since energy transfer from Yb 3+ to Er 3+ ions is known to occur in LiYF4:Yb,Er crystals [59,60]. This energy transfer results in upconversion emission of the co-doped particles, as shown in Fig. 4. Compared to the upconversion emission spectrum of the microcrystalline β-NaYF4:Yb,Er  [98,99]. The arrows indicate direct excitation of Er 3+ at 375 nm and UC excitation at 976 nm and the main Er 3+ emission bands in the green, red, and NIR spectral region ( 850 nm). (c) Normalized luminescence spectra (lower panel) of the upconversion emission of (1) 18 nm LiYF4:Yb(18%),Er(2%) core particles, (2) the corresponding LiYF4:Yb(18%),Er(2%)/LiYF4 core/shell particles, and, for comparison, (3) a microcrystalline NaYF4:Yb(18%),Er(2%) upconversion phosphor. The latter displays a different crystal field splitting of the emission lines due to the different point symmetry of the rare-earth lattice sites compared to LiYF4:Yb,Er. All spectra were recorded with the same spectral resolution. upconversion phosphor (Fig. 4), the crystal field splitting of the Er 3+ emission lines in the luminescence spectrum of LiYF4:Yb,Er is different, since the point symmetry of the rare earth sites in the inverse scheelite (CaWO4) structure of LiYF4 is S4 rather than C3h [1,2,100,101].
The decay kinetics of the green and red Er 3+ emission of the co-doped core/shell particles are displayed in Fig. 5. The figure clearly shows that direct excitation of the Er 3+ ions at 375 nm results in shorter decay times than excitation at 978 nm. This was to be expected, as excitation at 978 nm mainly excites the Yb 3+ ions which show a comparatively long lifetime of about 0.86 ms even in the presence of Er 3+ (Fig. 3). Since energy transfer from Yb 3+ to Er 3+ persists as long as the number of excited Yb 3+ ions is sufficiently high, excitation of Er 3+ via the ETU mechanism results in a longer decay time of the red and green Er 3+ emission than short-pulsed direct excitation at 375 nm. The decay times given in Fig. 5 are again intensity-weighted average lifetimes derived from multiexponential fits of the luminescence decay profiles. This method yields very low residuals in all cases. However, to be able to compare the luminescence kinetics of our core/shell particles with literature data reported for a single crystal of LiYF4:Yb,Er, we fitted our decay profiles also with a double exponential tail fit. Both methods yield fits of comparable quality.
The results of the double exponential fits are summarized in Table 1, together with the data reported for the single crystal. When the Er 3+ ions are excited directly, the decay times of the red and green Er 3+ emission of the core/shell particles are even slightly longer than those of the single crystal. This result, together with the close match between the Yb 3+ lifetimes of the core/shell particles and the Yb 3+ laser crystal (see Fig. 3), shows again that our LiYF4 shell protects the dopant ions in the core particles very well. When the materials are excited in the NIR, involving light absorption by Yb 3+ and subsequent energy transfer to the emissive Er 3+ ions, the red emission bands of the core/shell particles and the single crystal show very similar decay times, whereas the green emission of the nanocrystals decays slightly faster. The latter indicates that the energy transfer processes are not identical in both materials. The reason could be a different distribution of the Er 3+ ions in the core/shell particles and the single crystal, which could possibly favor cross-relaxation in the former known to decrease the lifetime of the green emitting 2 S3/2 state of Er 3+ in LiYF4 [102]. A detailed analysis of the distribution of dopant ions, however, is currently beyond reach. Figure 6 displays absolute measurements of the P-dependent upconversion quantum yields ( UC) of the Yb,Er doped samples. As highlighted in Fig. 6, the inert LiYF4 shell significantly increases  UC, the enhancement factor depending on P. The largest enhancement of more than a factor of 100 is observed at low P and decreases at higher P. A maximum  UC value of 1.25% is determined at 180 W·cm −2 for the core/shell particles with the 5 nm LiYF4 shell. This is almost one order of magnitude lower than the maximum  UC of the microcrystalline -NaYF4:Yb,Er upconversion phosphor. A lower quantum yield is, however, to be expected, as bulk -NaYF4:Yb,Er is reported to be the most efficient upconversion material known today.

Absolute measurement of the upconversion quantum yields
The reason for the higher quantum yield of -NaYF4:Yb,Er compared to LiYF4:Yb,Er could be the multisite formation (disordered structure) known to occur in the -NaYF4 host lattice, resulting in ideal resonance conditions for the Yb-to-Er energy transfer in this material [ [72].  UC measurements of relatively large diamondshaped LiYF4:20%Yb,2%Er/LiYF4 core/shell nanocrystals (length about 200 nm, width ≥ 70 nm) in cyclohexane revealed a value of 1.05% [78]. Significantly higher  UC values of 2.28% and 5.7% (P = 6.2 W·cm −2 ) have been reported, however, for Cd 2+ co-doped LiYF4:20%Yb,2%Er,10%Cd/LiYF4 core/shell nanocrystals and for single crystals of LiYF4 doped with only 2% Figure 6 Upconversion quantum yield of powders of LiYF4:Yb,Er core and core/shell particles as function of excitation power density P. The quantum yield of the 18 nm LiYF4:Yb(18%),Er(2%) core particles is one to two orders of magnitude lower than the quantum yield of the corresponding LiYF4:Yb(18%),Er(2%)/LiYF4 core/shell particles. Yb 3+ and 1% Er 3+ , respectively [78,103]. The latter can indicate a different optimum doping concentration for LiYF4 crystals compared to ß-NaYF4, which was also observed for SrF2 single crystals [104]. Methods for increasing multisite formation in LiYF4 nanocrystals and a systematic search for optimal dopant concentrations may therefore be useful strategies to further increase  UC of LiYF4:Yb,Er nanocrystals.

Conclusion and outlook
In conclusion, we prepared LiYF4:Yb/LiYF4 core/shell nanocrystals displaying a decay time of the Yb 3+ emission very similar to Czochralski-grown laser crystals of LiYF4:Yb. Core/ shell particles co-doped with Er 3+ show similar decay kinetics of the red and green emission band as single crystals of LiYF4 containing the same dopant concentrations. These results confirm the high quality of our core/shell synthesis procedure. The long lifetimes of the excited dopant ions indicate a very low defect density in the core/shell particles and an efficient shielding of the approximately 20 nm large particle cores by the inert shell of approximately 5 nm thickness. This was achieved by introducing two important modifications into the previously used synthetic route: We developed a new method to prepare the LiYF4 shell, which is based on a reaction with a very low tendency to form separate new particles of pure shell material. This not only ensures that the vast majority of the shell precursor reacts with the core particles, but also that the shell can be prepared via a simple heating-up method without the necessity of slowly adding a precursor solution. By using anhydrous rare earth acetates and anhydrous lithium acetate for the synthesis and drying the oleic acid/octadecene solvent at higher temperatures, we could reduce the probability of incorporating OH − into the fluoride lattice as previously reported for the synthesis of highly luminescent NaYF4:Yb,Er/ NaYF4 core/shell nanocrystals [90]. Co-doping the core particles with Er 3+ leads to energy transfer from Yb 3+ to Er 3+ resulting in upconversion emission. Absolute measurements of the excitation power density (P)-dependent upconversion quantum yield ( UC) show the highest values reported so far for LiYF4:18%Yb,2%Er/LiYF4 core/shell nanocrystals of small size and without additional dopants. However, despite the efficient surface shielding confirmed by time-resolved luminescence measurements, the values are approximately one order of magnitude lower compared to the quantum yield of microcrystalline NaYF4:Yb,Er upconversion phosphor powder.
We assume that LiYF4 lacks the disordered structure of -NaYF4, which is believed to be the main reason for the high upconversion efficiency of -NaYF4. Furthermore, the optimum dopant concentrations could be different for the two hosts.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.