Effect of Defect States on the Upconversion Emission Properties in KLu 2 F 7 Nanocrystalline

The inﬂuence of the introduction of Li + ions on the upconversion (UC) properties of KLu 2 F 7 :Yb 3 + , Er 3 + nanocrystalline has been investigated in detail. It is found that the UC emission intensity of 10 mol% Li + ions doped KLu 2 F 7 :Yb 3 + , Er 3 + nanocrystalline is enhanced about 13 times in comparison with that of Li + -free sample. Under the excitation of 980 nm laser diode (LD), a signiﬁcant improvement of the green to red emission ratio (GRR) in the 10 mol% Li + -doped sample is observed by increasing pump power density. Li + ions could occupy the cationic sites or the interstitial sites in the KLu 2 F 7 host matrix, which probably eliminate the defect states via the charge compensation and tailor of the crystal ﬁeld, leading to the promotion of the UC emission efﬁciency. Thus, the introduction of Li + ions provides new opportunities for enhancing the scope of applications of Yb 3 + , Er 3 + doped KLu 2 F 7 nanocrystalline ranging from infrared solar cells to volumetric multiplexed bio-probe. The preparation of UC nanoparticles (NPs) that exhibits anti- Stockes emission is important for applications in ﬁelds as diverse as solar cells, photovoltaics, biological imaging, and photocatalysis.

The preparation of UC nanoparticles (NPs) that exhibits anti-Stockes emission is important for applications in fields as diverse as solar cells, photovoltaics, biological imaging, and photocatalysis. [1][2][3][4] In particular, lathanide-doped fluorides have some distinct advantages superior to oxide-based UC materials due to their low phonon energy, high signal-to-noise ratio and excellent chemical stability. However, the applications of lanthanide-doped fluorides are still constrained because of the low UC efficiency and the restricted tunability of the luminescence color output. Various attempts have been devoted to improving these aspects, including both internal adjustments and external approaches, such as varying the crystal phase or morphology of the NPs, adjusting the concentration of the doped rare earth, introducing a co-dopant sensitizer, surface coating or adopting of a core-shell structure. [5][6][7][8][9][10][11] The excitation and relaxation dynamics of energy levels involved in UC process can be manipulated to vary the relative luminescence intensity in different UC bands of a lanthanide ion or to realize a combination of UC emission bands from other lanthanide ions. 12,13 To improve the UC emission efficiency and adjust the relatively color, a core-shell structure or a combination of activators is selected, which are aimed at reducing the surface defects. 14,15 However, the synthesis of NPs featuring controllable color with high chromatic purity remains a formidable challenge, as lanthanide ions generally have more than one metastable excited state and the defects are unavoidable in NPs. Nowadays, a new class of KLn 2 F 7 NPs with orthorhombic crystallographic structure has drawn a widespread interest, in which the lanthanide ions are distributed in arrays of tetrad clusters. This unique arrangement enables the preservation of excitation energy within the sub-lattice domain and effectively minimizing the migration of excitation energy to defects. Based on this, the adjustment of UC emission color and the unusual four-photon violet UC process from Er 3+ doped KYb 2 F 7 are observed. 16 Nevertheless, the influence of defects on the photoluminescence properties of KLn 2 F 7 NPs has not been recognized. With the small cationic radius, Li + ions can be doped easily into the host lattice substitutionally or interstitially, which will tailor the crystal field around the Ln 3+ ions and have a non-negligible influence on defect states, resulting in the change of the UC properties. 17,18 Although the increase UC emission intensity has been obtained via Li + ions doped in NaYF 4 :Yb 3+ /Tm 3+ , NaGdF 4 :Yb 3+ /Er 3+ and NaLuF 4 :Yb 3+ , Tm 3+ /Ho 3+ NPs, [19][20][21] the investigation on Li + doped KLn 2 F 7 has not been carried out.
In this study, we report a facile and mild hydrothermal process of the synthesis of water-phased KLu 2 F 7 NPs. The mechanism of the z E-mail: yuyu6593@126.com enhanced UC emission intensity of the as-synthesized KLu 2 F 7 :Yb 3+ , Er 3+ NPs with Li + ions doped under the excitation of 980 nm LD is investigated. Additionally, the explanation for the power density dependence of the UC color is given. With Li + ions doped, the elimination of the traps in the NPs as well as the tailor of the crystal field around Er 3+ ions play vital roles in the UC properties, which will be a meaningful research direction.

Experimental
The Li + doped KLu 2 F 7 :Yb 3+ , Er 3+ NPs were prepared by a hydrothermal method. In a typical experiment, 0.77 mmol Lu 2 O 3 (99.99%, Aladdin, China), 0.16 mmol Yb 2 O 3 (99.99%, Aladdin, China) and 0.02 mmol Er 2 O 3 (99.99%, Aladdin, China) powder were dissolved in dilute nitrate solution and the residual nitrate was removed by heating and evaporation, resulting in the formation of clear solution of Ln(NO 3 ) 3 (Ln = Lu, Yb, Er). Then, 0.05 mmol LiNO 3 (99.99%, Aladdin, China) water solution was added into the above nitrate solution under stirring. 7 mmol KF (99.0%, Aladdin, China) was dissolved in 10 mL deionized water and dropped into the nitrate solution. The result solution was transferred into a 100 mL Teflonlined autoclave and kept at 180 • C for 4 h. The NPs were separated via centrifugation and washed twice with water and once with absolute ethyl alcohol. The products (KLu 2 F 7 :16%Yb 3+ , 2%Er 3+ , 5%Li + ) were obtained as a kind of white powder after drying at 80 • C in a baking oven. The other concentration (0%, 10%, 15% and 20%) Li +doped KLu 2 F 7 :16%Yb 3+ , 2%Er 3+ samples were synthesized by a similar procedure except that different moles of LiNO 3 involved in the reaction.
X-Ray powder diffraction (XRD) was performed using a D8 Focus diffractometer (Bruker) with Cu-Kα radiation (λ = 0.15405 nm) in the 2θ range from 10 • to 80 • . The particle morphology and size were studied by the Scanning Electronic Microscopy (SEM) with FEI-Quanta 600. The UC photoluminescence spectra of the samples under a 980 nm infrared laser excitation were recorded by HITACHIU-F-7000 spectrophotometer at room temperature. The thermoluminescence (TL) curves were measured with a FJ-427 A TL meter (Beijing Nuclear Instrument Factory). Weight of the measured samples was constant (0.002 g). Prior to the TL measure, the samples were first exposed to the radiation from ultraviolet (UV) light (365 and 254 nm) for about 20 min, then heated from room temperature to 550 K with a rate of 1 K/s. The photoluminescence decay curves were measured by FS980 fluorescence spectrophotometer.

Results and Discussion
Fig. 1a shows the XRD patterns of 0, 5, 10, 15 and 20 mol% Li + ions doped KLu 2 F 7 :Yb 3+ /Er 3+ NPs. All samples are orthorhombic structure according to the standard card of KYb 2 F 7 JCPDF 27-0459. No impurity peaks are observed, indicating that these KLu 2 F 7 :Yb 3+ /Er 3+ samples synthesized with different amounts of Li + ions doped are of high purity. The positions of the diffraction peaks shift slightly when Li + ions are introduced. Fig. 1b shows the main diffraction peak shift toward larger angle gradually from 27.86 • to 28.14 • as the concentration of Li + ions increases in the range of 0-15 mol%. It indicates that the lattice shrinks when Li + ions are introduced. The peaks shift to smaller angles when Li + ion concentration further increase to 20 mol%, implying that the lattice expands. The crystal cell volume is calculated by the cell parameters in a unit cell with crystal structure of KLu 2 F 7 . The cell volume of the samples decreases linearly when the concentration of Li + ions changes from 0 to 10 mol% as shown in Table I and the insert of Fig. 1b, which is consistent with Vagard's law. [23][24][25] Li + ions are expected to replace K + and Lu 3+ ions, which leads to lattice shrinking. When the doped concentration increases, the confusion trend of the cell volume could suggest the occupying of the interstitial sites of host matrix. Based on the Shannon theory, the effect ionic radius of K + is 1.37 Å in four-coordination, Lu 3+ is 0.861 and 0.977 Å in six-and eightcoordination and Li + is 0.590, 0.760 and 0.920 Å in four-, six-and eight-coordination, respectively. 22 Therefore, Li + ions are speculated to substitute the cationic crystal sites or enter into the interstitial sites of host matrix. It should be noted that both the substitution of cationic sites and the occupation of interstitial sites would tailor the local crystal field around Er 3+ ions in the host lattice, and then influence their anti-Stokes luminescence.
The images in Fig. 2 represent the regular hexagonal prism of the NPs with Li + ions doped. The introduction of 5 and 10 mol% Li + ions leads to a larger grain size compared to Li + free sample, which probably has influence on the UC property. Further increasing of Li + concentration to 20 mol%, the particles with a much larger size hexagon sheet are formed due to interstitially crystal sites occupied by the Li + ions, and the crystal size reaches to a micrometer scale. It is evident from SEM images that Li + ions incorporated into the KLu 2 F 7 lattice serve as a controller for the morphology. UC photoluminescence properties of a series of KLu 2 F 7 :16%Yb 3+ , 2%Er 3+ , x%Li + (x = 0, 5, 10, 15 and 20) excited by 980 nm LD with the power of 50 W/cm 2 at room temperature are depicted in Fig. 3. The position of the emission peaks of Er 3+ is not affected by introducing Li + ions. The dominant green emissions of 525 and 543 nm are assigned to the transitions of 2 H 11/2 → 4 I 15/2 and 4 S 3/2 → 4 I 15/2 of Er 3+ ions, respectively, and the red emission of 668 nm is attributed to the 4 F 9/2 → 4 I 15/2 transition. With an increasing concentration of Li + ions, the integrated emission intensities increase firstly and then decrease. The UC luminescence intensity reaches its maximum in 10 mol% Li + ions doped sample, which is 13 times than that of Li + -free sample. The UC intensity decreases when the concentration of Li + ions reaches to 20 mol%. These phenomena may derive from the contributions that the introduction of Li + ions could tailor Er 3+ ions' local environment, decrease the defect states. A small fraction of Li + ions substituted in the lattice could induce the asymmetric environment around Er 3+ ions, which promotes the fast energy transfer from Yb 3+ to Er 3+ ions. Furthermore, defects are unavoidably formed during the synthesis process, indicating that the introduction of Li + ions is expected to influence the UC properties by changing defect states as well.
Thermoluminescence (TL) spectra are presented to reveal the change of defect states caused by Li + ions. As shown in Fig. 4, two TL peaks located around 325 and 387 K can be identified in Li +free sample. However, the TL peaks located around 325 and 387 K are insignificant when Li + ions are doped, indicating the traps are eliminated. During the synthesis process under high temperature and pressure, the substitution of K + ions with Yb 3+ and Er 3+ ions is unavoidable, though the replacement of Lu 3+ ions is dominant. Therefore, the introduction of the vacancy defects (V K ) cannot be avoided because of the charge balance principle. Moreover, the surface defects are inevitable in nanoparticles. Both of these defects contribute to the observation of the traps. 26 It could be expected that the introduction of Li + ions eliminates the traps efficiently, since the substituting of Lu 3+ ions with Li + ions producing the vacancy of F − (Li Lu +2V • F ) accomplishes the charge compensation. Thus, the introduction of Li + ions could improve the UC emission efficiency significantly.
The photoluminescence decay curves of 4 S 3/2 → 4 I 15/2 (543 nm) and 4 F 9/2 → 4 I 15/2 (668 nm) of KLu 2 F 7 :16%Yb 3+ , 2%Er 3+ , x%Li + (x = 0, 5, 10, 15 and 20) NPs are recorded as shown in Fig. 5. All the decay curves of the samples could be well fitted to a single exponential function:  Where I is the photoluminescence intensity of Er 3+ ions, A 1 is constant; t is time; and τ 1 is the decay time for the exponential component. As shown in Fig. 5, the introduction of Li + ions prolongs the decay time of 4 S 3/2 and 4 F 9/2 levels in Er 3+ ions indicating that the local environment around Er 3+ ions are tailored with the Li + ions doped. With an increasing concentration of Li + ions, the lifetimes of 4 S 3/2 and 4 F 9/2 states increase firstly and then decrease, and such trend is similar to the UC emission intensity change. The decay time of 10 mol% Li + -doped sample is longer than other samples, indicating that the UC emission efficiency would be improved with proper concentration of Li + ions. The large increase in decay time contributes to a significant increase of UC photoluminescence intensity, which has been studied and reported. 27 Fig. 6a demonstrates the UC photoluminescence spectra of KLu 2 F 7 :16%Yb 3+ , 2%Er 3+ , 10%Li + with great enlargement of GRR under different excitation power density by normalizing the red emission. The GRR can be tuned by varying the excitation power density. The emission ratios are calculated by the integrated emission intensities of 525 and 543 nm to 668 nm. The change of GRR can be attributed to a fast and efficient energy transfer process due to the participation of more intensive photons in UC emission process. Since the electrons could be excited frequently before they decay under higher power density of 980 nm LD excitation, UC emission favors the transition from higher energy levels, which would promote the electrons stimulated from 4 S 3/2 level. Thus, the emission intensity of the 525 nm peak increases drastically in comparison with 543 nm and the energy transition 4 G 11/2 → 4 I 11/2 (380 nm) and 2 H 9/2 → 4 I 11/2 (408 nm) occur as shown in Fig. 6a1. Therefore, an increasing excitation power density of 980 nm LD would not only adjust the UC emission color efficiently but promote the UC photoluminescence intensity in short wavelength. In addition, with the introduction of Li + ions, the GRR is enlarged more greatly as the increasing excitation power density compared with that of Li + -free sample (Fig. 6a2) with the assistance of the elimination of the traps and the tailor of the crystal field around of Er 3+ ions. To visualize the improvement and the tunable color of the UC emission, the photographs of the UC luminescence of KLu 2 F 7 : 16%Yb 3+ , 2%Er 3+ , 10%Li + NPs dispersed in the absolute ethyl alcohol (2 wt%) excited by 980 nm LD are shown in Fig. 6b. With an increasing power density, a tunable color from yellow to green is clearly observed with the naked-eye. Fig. 6c shows the photograph model that the cuvette containing 2%wt sample is keeping away from convex lens gradually, which indicates a decrease of excitation power density. Evidently, as further apart from the focus, the GRR decreases, verifying the dependence of GRR on the power density, which are corresponding to the results of Fig. 6b. Furthermore, based on the spectral data of the UC photoluminescence, variation of color points of the 10 mol% Li + ions doped KLu 2 F 7 : Yb 3+ , Er 3+ NPs as a function of power density are obtained, which is illustrated in the Commission International de I'Eclairage France (CIE) 1931 chromaticity diagram in Fig. 6d. Obviously, the color coordinates of UC emission falls in yellow regions under the excitation of 980 nm LD with a low power density. It ranges from yellow to green as the pomp power density increase gradually. Fig. 7 is the schematic partial energy level diagram and energy transfer processes involved in the UC process. The energy transfer from Yb 3+ ions to Er 3+ ions can promote the electrons from the ground state to the 4 F 7/2 level of Er 3+ ions under the excitation of Figure 6. Pump power density dependence of UC photoluminescence spectra (a), the photographs (b), photograph model (c) and color coordinates (d) of KLu 2 F 7 :16%Yb 3+ , 2%Er 3+ , 10%Li + excited by 980 nm LD. The inset of (a) shows the amplification of UC spectra in short wavelength (a1) and the power density dependence of GRR of x = 0 and 10 samples (a2).  980 nm LD. With the assistance of non-radiative relaxation process, the electrons can then relax non-radiatively to the 2 H 11/2 and 4 S 3/2 levels, which contributes to the green emitting as the electrons finally relax to the 4 I 15/2 level. It can be seen that the red emission derives from relaxation of electrons from 4 F 9/2 to 4 I 15/2 level. The excited electrons relax to 4 I 13/2 from 4 I 11/2 level non-radiatively, and then populate the red emitting 4 F 9/2 state. A high power density excitation of 980 nm LD would be beneficial to excite the electrons to a higher energy level of Er 3+ ions. It provides much more opportunities for electrons to be excited at 4 G 11/2 level, which would contribute to the observation of the near-ultraviolet region (380 nm) or depopulate to purple emitting at 2 H 9/2 (408 nm) levels. The elimination of the traps with the introduction of Li + ions and the tailor of the crystal field around Er 3+ ions would improve the UC efficiency significantly.

Conclusions
In summary, KLu 2 F 7 : Yb 3+ , Er 3+ NPs doping with different concentration of Li + ions are synthesized via the hydrothermal process. Due to the introduction of Li + ions, the enhancement of UC emission intensity and the decrease of TL intensity are observed. The optimized concentration of Li + ions for the maximum UC emission intensity in KLu 2 F 7 :16%Yb 3+ , 2%Er 3+ NPs is 10 mol%. Furthermore, the dependence of the adjustment UC color on the excitation power density is investigated, and the UC photoluminescence color is tuned from yellow to green with the increasing power density. The introduction of Li + ions could tailor the crystal field around Er 3+ ions and eliminate the traps, which would contribute to the dramatical enhancement of UC emission intensity and a significant enlargement of the GRR as the increase of the excitation power density. Thus, the introduction of Li + ions will greatly enhance the scope of the applications of KLu 2 F 7 : Yb 3+ , Er 3+ NPs.