Introduction

Considerable interests in rare-earth (RE) doped UC phosphors have burgeoned since the development of synthetic methods for creating highly processable UC materials with applications in biolabeling1, 2, sub-band gap energy harvesting in photovoltaics3, 4, and security printing5. Due to its potential applications in biological fluorescence and volumetric displays, much effort has been devoted to the investigation on UC in RE3+ doped materials over the past years. Recently, a few basic works have been reported in order to explain the origin of the relatively high UC efficiencies6, the ability to adjust the color purity through manipulating dopant concentrations7, and the special phenomenon that result from surface effects in nanoscale UC phosphors8, 9.

Among all the rare-earth ions, Er3+ ion is one of the best candidates as an activator in up-conversion luminescence on account for its longer lifetimes of metastable energy levels and more homogeneous energy level array. By employing Yb3+ ions as sensitizer into Er3+ doped materials, the UC luminescence performance can be improved because of a larger absorption cross section at NIR of Yb3+ and the high-efficiency energy transfer from Yb3+ to Er3+ ions10,11,12.

A number of researches on Er3+/Yb3+ co-doped systems have been focused on exploring the UC mechanisms and developing novel hosts in recent years13. In order to reduce the multi-phonon nonradiative relaxation and realize the high efficient UC luminescence, low maximum phonon energy hosts are required and necessary. Compared with the popular oxide materials, it is well-known that fluorides are more efficient hosts for RE3+ ions due to their low energy phonons to produce strong UC fluorescence. Up till now, fluorides like Na5Lu9F32 have been investigated widely14,15,16. However, the studied materials of fluoride Na5Lu9F32 mainly concentrate on nano-crystals and powders17, and the Er3+/Yb3+ co-doped UC materials have never been in-depth studied in Na5Lu9F32 single crystals. Compared with nano-crystal, single crystal provides higher transmission for lights and thermal stability as well as good chemical durability. There are scarce reports on Na5Lu9F32 single crystals because of the difficulty in the crystal growth. Thus, Na5Lu9F32 in form of single crystal is an excellent candidate matrix for Er3+/Yb3+ ions to investigate the UC luminescence spectra.

Generally, there exists the problem of the inhomogeneous distribution of RE ions in single crystal resulted from crystal growth. In this study, Na5Lu9F32 single crystal was chosen as a matrix of Er3+ and Yb3+ ions in view of the same valence state and the very near ionic radii between Lu3+ (0. 861 Å) and Er3+ (0.881 Å), Lu3+ and Yb3+ (0.858 Å), which resulted in relatively homogeneous concentration of Er3+ and Yb3+ in Na5Lu9F32 single crystal. The Na5Lu9F32 single crystal with homogeneous rare earths is extremely important for the practical application in optical device.

Results and Discussions

Characteristics of Na5Lu9F32 single crystal

In order to understand the structures and characteristics of the as-grown crystals, the crystallinity was confirmed in the X-ray powder diffraction (XRD) patterns as shown in Fig. 1(b). It can be seen from the XRD spectra of Er3+/Yb3+ co-doped Na5Lu9F32 single crystal in Fig. 1(b), the structure of the single crystal was confirmed to be cubic phase, in accordance with JCPDS card (27–0725) of Na5Lu9F32 as shown in Fig. 1(c), which the diffraction peak positions of the obtained samples doped with Er3+/Yb3+ ions are matched perfectly with those of standard Na5Lu9F32, though there exists a slight shift in the range of 0.10° to 0.45° as the dopant change, indicating that this transparent crystal has pure cubic phase. Moreover, the cell parameters were calculated by the following formula18

$$d=\frac{a}{\sqrt{{h}^{2}+{k}^{2}+{l}^{2}}}$$
(1)
Figure 1
figure 1

(a) The photo of Na5Lu9F32 single crystal, the left is boule of the crystal and the right is polished slice. (b) XRD patterns of the Na5Lu9F32: Er3+/Yb3+. (c) Standard line patterns of Na5Lu9F32 (27–0725).

Accordingly, the cell parameter is calculated to be a = 0.5421 nm (space group Fm3m) for sample NFEY2 (0.99 mol% Er3+/1.97 mol% Yb3+ sample) from the measured XRD patterns.

Absorption and transmittance spectra of the Er3+/Yb3+ co-doped Na5Lu9F32 single crystal

The absorption spectra in the 400–1100 nm wavelength of the as-grown Na5Lu9F32 single crystal samples has been measured and presented in Fig. 2(a). The absorption peaks corresponding to transitions from the 4I15/2 ground state to excited states of the Er3+ ions are also assigned in Fig. 2(a). Compared the Er3+ singly-doped sample with Er3+/Yb3+ co-doped sample, there is nearly no intensity and position change for the shape of peaks in the visible region, where confirms that Er3+ ions can absorb visible light. As for the absorption center at 980 nm, Er3+/Yb3+ co-doped sample shows an obvious peak located at 980 nm corresponding to the transition from 2F7/2 → 2F5/2 of Yb3+ ion, which demonstrates that Yb3+ ions acts as an excellent sensitizer to absorb the pump light. The corresponding UV-VIS-NIR transmittance spectrum of well polished Er3+/Yb3+ co-doped Na5Lu9F32 single crystal with 1.5mm thickness also showed in Fig. 2(b). It can be confirmed that the transmittance was measured to be ~89% in the visible region (Na5Lu9F32 single crystals are highly transparent as shown in Fig. 1(a)), which proves that there exists a high transmittance in the range of 400–1000 nm.

Figure 2
figure 2

(a) Absorption spectra of Er3+/Yb3+single doped and co-doped Na5Lu9F32 single crystals. (b) Transmittance spectra of Er3+/Yb3+ co-doped Na5Lu9F32 single crystal.

Emission spectra and energy transfer of as-prepared Na5Lu9F32 single crystal co-doped 1.0% Er3+ and χ% Yb3+ (χ = 0, 2, 6, 8)

The quantitative comparison of up-conversion luminescence of as-prepared Na5Lu9F32 single crystal co-doped 1.0% Er3+ and χ% Yb3+ (χ = 0, 2, 6, 8) under NIR 980 nm excitation have been shown in Fig. 3, from which three obvious bands in the range of 450–750 nm can be observed. As presented in Fig. 3, two green emissions around 525 nm and 548 nm are assigned to the radiative transition of 2H11/2 → 4I15/2 and 4S3/2 → 4I15/2, respectively. And red emission band from 638 to 688 nm is attributed to the 4F9/2 → 4I15/2 transition of Er3+ ion. Compared luminescence intensity of Er3+ single-doped sample with Er3+/Yb3+ co-doped Na5Lu9F32 one, there is a distinct improvement of up-conversion luminescence efficiency when the Yb3+ ions take part in. The relative amount of Yb3+ sensitizer ions tend to obviously improve the up-conversion luminescence (UCL) intensity in Er3+/Yb3+ phosphors, thus we have investigated the impact of various ytterbium concentrations on the intensity of the UC emission under the same experiment condition. The insert of Fig. 3 shows the integrated total intensity on Yb3+ concentrations. The UCL intensity of 668 nm emission increases rapidly as the Yb3+ concentration increases and reaches its maximum value when Yb3+ concentration is 7.97 mol% in present research, which is attributed to the aggravated cross-relaxation process of 4F7/2 + 4I11/2 − 4F9/2 + 4F9/2 with the increase of Yb3+ ions concentration. The concentration quenching of Yb3+ ions is not observed, while the maximum intensity of 548 nm and 525 nm emission is obtained when Yb3+ concentration is 1.99 mol%.

Figure 3
figure 3

Comparison of the up-conversion luminescence spectra of Er3+/Yb3+ doped Na5Lu9F32 single crystals under 980 nm excitation at room temperature; the inset shows the integrated total intensity on Yb3+ concentrations in Er3+/Yb3+ doped Na5Lu9F32 single crystals.

To reveal the origin of these radiations and the mechanism of the UC emissions, the decay profiles of Na5Lu9F32 single crystal doped Er3+ with different amount of Yb3+ ions under 980 nm excitation at 548 nm is shown in Fig. 4, and the inset is the comparison of decay times centered at 525, 548 and 668 nm for 1.0% Er3+/8% Yb3+ co-doped Na5Lu9F32 single crystal. These decay profiles were fitted with non-exponential, the decay time of these radiations centered at 525, 548 and 668 nm show the similar value of ~880 μs from the inset of the Fig. 4. Meanwhile, the lifetime for 1.0% Er3+ singly doped Na5Lu9F32 single crystal and 1.0% Er3+/χ% Yb3+ (χ = 2,6,8) co-doped Na5Lu9F32 single crystals are τm = 311.1 μs, τm = 1567.7 μs, τm = 1114.4 μs, τm = 870.4 μs respectively. It suggests that there is no concentration quenching in present doping concentration. And it can be seen that the fluorescence for the co-doped crystals (NFEY2, 6, 8) decays more quickly than that for the single-doped one (NFEY0) from Fig. 4, which indicates an obvious enhancement of up-conversion luminescence efficiency arise from dopants of Yb3+ ions. As presented in Fig. 4, the lifetime of green emission decreased apparently when the concentration of Yb3+ ions increases, which can verify that cross-relaxation (CR) process of 4F7/2 + 4I11/2 → 4F9/2 + 4F9/2 occurred. The Na5Lu9F32 single crystal shows high incorporating concentrations for Er3+ and Yb3+. Owing to the similar radii size and same valence state of between rare earth ions, Er3+ ions and Yb3+ ions could easily take place of the position of Lu3+ in Na5Lu9F32 single crystal. The homogeneity of REs in Na5Lu9F32 can be reflected from the effective segregation coefficients of REs in single crystal and they can be estimated from the measured concentration and the formula \(c=k{c}_{0}{(1-f)}^{(k-1)}\), where c and c0 express the concentrations of RE in single crystal and raw material, k is effective segregation coefficient of RE, and f is the crystallization rate19. The effective segregation coefficients for Er3+ and Yb3+ in Na5Lu9F32 were about 0.990 and 1.013, respectively, which approach to 1. It indicates that a relatively homogeneous concentrations of REs can be obtained because of the very near ionic radii between Lu3+ (0. 861 Å) and Er3+ (0.881 Å), Lu3+ and Yb3+ (0.858 Å) and results into high luminescence effect of Er3+ and Yb3+.

Figure 4
figure 4

Decay curves of the Er3+/Yb3+ co-doped Na5Lu9F32 single crystals surveyed at 548 nm under 980 nm excitation, and the inset is the decay times centered at 525, 548 and 668 nm for 1.0% Er3+/8% Yb3+ co-doped Na5Lu9F32 single crystal.

In order to investigate the UC dynamics of Er3+/Yb3+ co-doped Na5Lu9F32 single crystal, the pump-power dependence of luminescence intensities was measured as a function of excitation power density under the 980 nm excitation. As shown in Fig. 5, the up-conversion luminescence intensities increased as excitation power increasing to the maximum ~666 mW, which may conclude that high excitation power can improve the up-conversion luminescence efficiency. It is known that the relation between emission intensity Iem and NIR excitation power IemPn in frequency up-conversion process20, where n is the number of pump photons required to excite to the emitting state. From the log–log dependence of the integrated green (~525 and ~548 nm) and red emission (~668 nm) intensities on the excitation power at 980 nm shown in the inset of Fig. 5, the graph of log (I em) versus log (P) yields a slope of n equal to approximately 1.968, 1.852, 1.721, respectively. The value of n is approximately equal to 2, which implied the lower thermal effect and saturation effect in Na5Lu9F32: Yb3+, Er3+ single crystals. Thus, the quadratic dependence indicates that blue, green and red emissions all arise from two-photon UC processes.

Figure 5
figure 5

Log–log dependence of the up-conversion intensities of 1.0 mol% Er3+ and 7.97 mol% Yb3+ co-doped Na5Lu9F32 single crystal at 525,548, 668 nm emissions as a function of the excitation power at 980 nm.

To better understand the mechanism of UC luminescence in Er3+/Yb3+ co-doped Na5Lu9F32 single crystal, a schematic energy level diagram of Er3+/Yb3+ and the energy transfer process has been presented in Fig. 6. Following the 980 nm pumping light, the 4I11/2 level of Er3+ ion can be directly excited from ground state 4I15/2 or by energy transfer (ET) process from 2F5/2 level of Yb3+ to Er3+: 2F5/2 (Yb3+) + 4I15/2 (Er3+) → 2F7/2 (Yb3+) + 4I11/2 (Er3+). It should be noted that interactions between two Er3+ ions cannot necessarily be ignored. An NIR photon can populate an Er3+ ion to its 4I11/2 state directly. Another Er3+ ion also in the 4I11/2 state and in close proximity will transfer its energy to the initial ion, thereby exciting it to the 4F7/2 state. However, the absorption cross section of Yb3+ at ~980 nm is much larger than that of Er3+: 4I11/2, which can be seen from Fig. 2, resulting in the ET process from Yb3+ to Er3+ is dominant in the excitation of (Er3+) level. It is known that when co-doped samples are excited by high power successive laser at 980 nm, the Er3+ ion may decay non-radiatively from the 4I11/2 state to the 4I13/2 state following the initial energy transfer from the Yb3+ ion by multi-phonon relaxation in recent works. However, bridging the energy gaps of 4I11/2 → 4I13/2 (3619 cm−1) or 4S3/2 → 4F9/2 (3217 cm−1) requires at least 6−7 phonons, multi-phonon relaxation will rarely happen in Na5Lu9F32 single crystal due to its relatively low phonon energies about ~441 cm−1  21. Thus, the excitation in the 2F7/2 → 2F5/2 transition of Yb3+ is followed by a two-step ET process to neighbouring Er3+ ions as shown in Fig. 6, which brings Er3+to the 4F7/2 level. The populated 4F7/2 level of Er3+ then non-radiatively relaxes fast to the lower 2H11/2 and 4S3/2 states because of the smaller energy gaps of 4F7/2 → 2H11/2 (1162 cm−1) and 2H11/2 → 4S3/2 (794 cm−1), and two green emissions then be observed. Above process generates two 2H11/2 → 4I15/2 and 4S3/2 → 4I15/2 transitions centered at 525 and 548 nm, respectively. Meanwhile, red emission band is attributed to the 4F9/2 → 4I15/2 de-excitation process. And the population on 4F9/2 should come from the CR process of 4F7/2 + 4I11/2 → 4F9/2 + 4F9/2, which can be evidenced by the Yb3+ concentration-dependent UCL spectra of Na5Lu9F32: Yb3+, Er3+ single crystals (Fig. 3) as well as the decay curves of the Er3+/Yb3+ co-doped Na5Lu9F32 single crystals surveyed at 548 nm (Fig. 4). As shown in the Fig. 3, When the concentration of Yb3+ increases to about 7.97 mol%, probably due to concentration quenching between Er3+ and Yb3+ ions, the energy transfer process of Er3+ (4I11/2) + Yb3+ (2F5/2) → Er3+ (4F7/2) + Yb3+ (2F7/2) reduced accordingly, which leads to the reduction of the population of 4F7/2 state of Er3+ ions and results in the decrease of 525 nm and 548 nm emissions. Thus the strongest emissions at 525 nm and 548 nm in this study is observed when the Yb3+ is about 1.99 mol%. However, the process of two photon UC process 22F5/2 (Yb3+) + 4I15/2 (Er3+) → 22F7/2 (Yb3+) + 4F7/2 (Er3+) continues to increase luminescence intensify as the Yb3+ concentration increases. The concentration quenching between Yb3+ and Yb3+ ions has not been observed in this study.

Figure 6
figure 6

Energy level diagram of Er3+/Yb3+ and energy transfer progress.

UC external quantum efficiency of Na5Lu9F32: 1% Er3+, 7.9% Yb3+

Having analyzed all the results above, pump-power dependent of UC external quantum efficiency (EQE) of Na5Lu9F32: 1 mol% Er3+/7.9%mol Yb3+ under the excitation of NIR ~980 nm light was measured by an integrating sphere and displayed in Fig. 7(a). Following the 980 nm excitation, it should be noted that the EQE increases with the excitation power density, and the optimum EQE was about 6.80% under 5.5 Wcm−2 980 nm light excitation. And once the power density overpasses 5.5 Wcm−2, the EQE of UCL hardly changed. In order to verify this point further, we calculated the theoretical EQE based on steady-state rate equations. On account of the simplified model shown in Fig. 7(a) above, the UC external quantum efficiency (ɳ IQE) can be estimated by following equations:

$$\begin{array}{c}{c}_{0}{N}_{0}{N}_{Yb1}-{c}_{1}{N}_{1}{N}_{Yb2}-{W}_{1}{N}_{1}=0\\ {c}_{1}{N}_{1}{N}_{Yb2}-{W}_{2}{N}_{2}=0\\ {N}_{0}+{N}_{1}+{N}_{2}={N}_{er}\\ {N}_{Yb1}+{N}_{Yb2}={N}_{Yb}\\ {N}_{Yb2}={\rho }_{P}\sigma {N}_{Yb1}\end{array}$$
(2)
$${\rho }_{P}=\frac{{\lambda }_{P}{I}_{P}}{hc\pi {\omega }_{p}\wedge 2}=\frac{{\lambda }_{p}}{hc}P$$
(3)
$${N}_{2}=\frac{{c}_{o}{c}_{1}{N}_{c}{N}_{Yb}\wedge 2{\rho }_{P}\wedge 2\sigma \wedge 2}{[{W}_{1}{W}_{2}+{c}_{0}{c}_{1}{N}_{Yb}\wedge 2+({c}_{0}+{c}_{1}){N}_{Yb}]\rho \wedge 2\sigma \wedge 2+[({c}_{0}+{c}_{1}){N}_{Yb}{W}_{2}+2{W}_{1}{W}_{2}]\rho \sigma +{W}_{1}{W}_{2}}$$
(4)
$$\eta =\frac{{W}_{2}{N}_{2}}{{\rho }_{P}\sigma {N}_{Yb0}}=\frac{4.38P}{1+0.51P} \% $$
(5)

where c0 and c1 are the ET coefficients for the UC processes between the donor and the acceptor in 4I15/2 and 4I9/2 states, respectively. N2, N1, and N0 are the population 2H11/2/4S3/2 4,I9/2, and 4I15/2 densities of the levels of the Er3+ ions. And the NYb1 and NYb2 represent the 2F7/2 and 2F5/2 levels of Yb3+ ions, respectively. NYb and Ner are the concentrations of the Yb3+ ions and the Er3+ ions. W1 denotes the non-radiative relaxation rate from level 4I11/2 to lower state of Er3+, W2 is the radiative decay rate from level 2H11/2/4S3/2 to ground state of Er3+, and σ is the Yb3+ absorption cross section at the pumping wavelength. The symbol ρP denotes the excitation power variable, given by equation (3). Here, IP is the incident pump power, λP and ωP are the pump wavelength and beam radius, respectively, h is Planck’s constant, and c is the speed of light. P is the incident pump power density. From the above equations, we can obtain the population of 2H11/2/4S3/2 level using equation (4). Thus, the UC external quantum efficiency (EQE) can be deduced by equation (5). The calculated results were also plotted a line in Fig. 7(a). It can be noted that the theoretical measurement in accordance with the experimental dates. Generally, there exists the thermal effect in the nano-crystals21, 22, which results into reduction of the quantum efficiency. It can be confirmed that the rare earth ion doped Na5Lu9F32 single crystal has advantage of achieving high quantum efficiency and high thermal stability compared with its nano-crystal. The inset of the Fig. 7(a) shows the integrated external quantum efficiency on Yb3+ concentrations. The relative amount of Yb3+ ions tends to improve the EQE in Er3+/Yb3+ co-doped Na5Lu9F32 single crystals obviously, thus we have investigated the impact of various ytterbium concentrations on the EQE of the up-conversion progress under the same experiment condition. It is shown in the inset of Fig. 7(a) that EQE of up-conversion increases as the Yb3+ concentration increases monofonically and reaches its maximum value at ~6.80% when Yb3+ concentration is 7.97 mol%. To testify the good UCL effect in Er3+/Yb3+ co-doped Na5Lu9F32 single crystals, we also compare them with Er3+/Yb3+ co-doped LiYF4 samples in Fig. 7(b) whose maximum EQE of up-conversion is about ~5.8%23. It is known that β- NaYF4 is widely accepted as the most efficient up-conversion host in the world. The Er3+/Yb3+(2 mol%/18 mol%) doped β- NaYF4 nano-crystal reaches its maximum QY at ~4.8%24. It should be noticed that Er3+/Yb3+ co-doped Na5Lu9F32 single crystals can achieve higher external quantum efficiency at the lower power density, which indicates the Er3+/Yb3+ co-doped Na5Lu9F32 single crystals own a better UCL effect and worth of further study.

Figure 7
figure 7

(a) Power-dependent UC external quantum efficiency at 980 nm light excitation in Er3+/Yb3+ doped Na5Lu9F32 single crystals, the inset shows the external quantum efficiency on Yb3+ concentrations; (b) Comparison of UC external quantum efficiency of Er3+/Yb3+ co-doped Na5Lu9F32/LiYF4 single crystals under 980 nm excitation.

CIE chromaticity coordinates of Er3+/Yb3+ co-doped Na5Lu9F32 single crystal samples

Since the observed UC emissions are located in visible wavelength area, the exact UC luminescence CIE chromaticity coordinates of all the Er3+/Yb3+ co-doped Na5Lu9F32 single crystal samples are shown in Fig. 8 to obtain the true color of the UC emissions. The CIE coordinates (x, y) are (0.3254, 0.6612), (0.3422, 0.6450), (0.3666, 0.6202), and (0.4721, 0.5188) respectively, for Er3+/Yb3+ co-doped samples with 1.0Er/0Yb, 1.0Er/1.97Yb, 1.0Er/5.99Yb, and 1.0Er/7.97Yb doping concentrations. From the CIE chromaticity coordinates, it is clear that the combination of the UC emission is yellowish green color to which human eyes are sensitive. This characteristic is favorable for applications in UC displays for electro-optical devices.

Figure 8
figure 8

CIE chromaticity coordinates of Er3+/Yb3+ co-doped Na5Lu9F32 single crystals.

Conclusions

Following excitation with a 980 nm diode laser, the Er3+/Yb3+ co-doped Na5Lu9F32 single crystals can be grown by a vertical Bridgman method and an enhanced up-conversion green and red lights can be obtained. Study on the pump power dependent UC spectra shows that the UC emissions of the green and red lights arise from two-photon process from excited Yb3+ to Er3+ energy transfer. The combination of the UC green and red lights yields a yellowish green light to which human eyes are very sensitive. The up-conversion quantum efficiency of Er3+/Yb3+ co-doped Na5Lu9F32 single crystal under 5.5 Wcm−2 980 nm light excitation was as high as 6.80%. Such Er3+/Yb3+ co-doped Na5Lu9F32 single crystals may have potential applications in the fields of biosensors, infrared pumped visible eye-safe lasers, optical telecommunication and UC displaying devices due to its low phonon energy, high luminous efficiency, high transmittance for light and stability in chemical-physical properties.

Experimental

Materials

Er3+/Yb3+ co-doped samples were grown by an improved Bridgman method according to the formula 30NaF-18KF-(52-χ-γ)YF3-χErF3-γYbF3 (χ = 1.0, γ = 0, 2 6, 8), respectively, with 99.99% purity raw materials of NaF, LuF3, ErF3, and YbF3.

Synthesis of Na5Lu9F32: Er3+/Yb3+ single crystals

Then these mixture samples were ground thoroughly in a mortar for about 0.5 h. In order to remove the moisture and the oxide in the raw materials, the mixtures were sintered with anhydrous HF at 750 °C for 8 h. The seed crystals were oriented along a-axis. The seeding temperature was about 770–820 °C and the temperature gradient cross solid-liquid interface was 70–90 °C/cm. The growing process was carried out by lowering the crucible at a rate of 0.05–0.06 mm/h. The detailed process was similar to that reported elsewhere25. Figure 1(a) shows the boule of as-grown single crystal and the polished slice of crystal with 1.50mm thickness. The single crystal appears high transparency and pink.

Characterization

X-ray diffraction (XRD) of the samples was measured using a D8 Advance diffractometer (BRUKER, German). The absorption spectra were recorded with a Cary 5000 UV/VIS/NIR spectrophotometer (Agilent Co., America). The emission spectra were obtained with a FLSP 920 type spectrometer (Edinburgh Co., England). The external quantum efficiency were measured by a fluorescence spectrometer (FLS 980) of Edinburgh instruments combined with an integrating sphere. All the measurements were measured at room temperature. The actual concentrations of Er3+ and Yb3+ ions in all the samples were measured by an inductively coupled plasma atomic emission spectroscopy (ICP-AES, Perk in Elmer Inc., Optima 3000). The measured Er3+ and Yb3+ concentrations in all crystals are presented in Table 1.

Table 1 The measured molar concentrations and the number of Er3+ and Yb3+ in Na5Lu9F32 crystal.