1. Introduction

Ce³⁺-doped aluminum garnet (R₃Al₅O₁₂) systems are considered suitable phosphors for use in phosphor-conversion white light emitting diodes (pc-WLEDs) pumped by blue LED chips. Among these phosphors, Y₃Al₅O₁₂:Ce³⁺ (YAG:Ce³⁺) [1–3] and Lu₃Al₅O₁₂:Ce³⁺ (LuAG:Ce³⁺) [4–8] are used predominantly as yellow- and green-emitting phosphors, respectively. In addition, aluminum garnets doped with Er³⁺ and Yb³⁺ have been investigated as upconversion (UC) phosphors, which convert near infrared (NIR, ~980 nm) to visible radiation through two-photon UC processes; these garnets include YAG:Er³⁺ [9], YAG:Yb³⁺,Er³⁺ [10,11], and Yb₃Al₅O₁₂:Er³⁺ (YbAG:Er³⁺) [12–14]. To the best of our knowledge, however, the UC luminescence of LuAG:Yb³⁺,Er³⁺ remains unexplored.

Two-photon UC may be achieved through various means including energy transfer UC (ETU), two-step absorption, cooperative sensitization, cooperative luminescence, and second-harmonic generation (SHG) [15,16]. In the case of ETU with a high UC efficiency, the Yb³⁺ ion is used as a sensitizer, and Er³⁺ or Tm³⁺ ion is co-doped as an activator. The UC photoluminescence (PL) spectra of the Er³⁺ and Yb³⁺ co-doped phosphors typically consist of green and red emission bands; the emission ratio of green to red varies with the host material and dopants. For example, the green emission of YNbO₄:Yb³⁺,Er³⁺ [17], β-NaYF₄: Yb³⁺,Er³⁺ [18,19], CaMoO₄:Yb³⁺,Er³⁺ [20], and ZrO₂:Er³⁺,Yb³⁺,Mo⁶⁺ [21], is stronger than the corresponding red emission; the reverse holds true for YAG:Yb³⁺,Er³⁺ [10,11], YbAG:Er³⁺ [12–14], α-NaYF₄:Yb³⁺,Er³⁺ [22], GdF₃:Yb³⁺,Er³⁺ [23], and LaF₃:Yb³⁺,Er³⁺ [24]. The red emission of LuAG:Yb³⁺,Er³⁺ was therefore expected to be stronger than the corresponding green emission, based on the UC luminescence of YAG:Yb³⁺,Er³⁺ and YbAG:Er³⁺. However, the UC spectra of LuAG:Yb³⁺,Er³⁺ are expected to be different from those of YAG:Yb³⁺,Er³⁺, as LuAG:Ce³⁺ differs from YAG:Ce³⁺ in the PL emission spectra. Furthermore, a change in the emission ratio of red to green was not investigated in depth in the earlier literatures.

As such, in this work, we prepared the LuAG:Yb³⁺,Er³⁺ UC phosphors and determined the effect of the ratio of the Yb³⁺ to Er³⁺ concentration on the UC emission.

2. Experiment

Lu_3-x-yYb_xEr_yAl₅O₁₂ (LuAG:xYb³⁺,yEr³⁺) powders were prepared by a flux method using Lu₂O₃ (99.99%, Grand Ceramic & Materials), Al₂O₃ (99.99%, High Purity Chemicals), Yb₂O₃ (99.99%, Grand Ceramics & Materials), and Er₂O₃ (99.9%, Grand Ceramics & Materials) as starting materials. Starting mixtures with stoichiometric compositions were ball-milled for 24 h and fired at 1700 °C for 7 h, under a nitrogen atmosphere (99.99%). A flux of 2 wt% AlF₃ (95%, Junsei) was added to facilitate the formation of the compounds through increased diffusion of the reactants. An X-ray diffractometer (XRD, Rigaku Miniflex II) using Cu_Kα radiation (λ = 1.5406 Å) was used to determine the crystal phase. The crystal parameters were obtained using the X′pert HighScore Plus software. In addition, UC PL spectra were measured at room temperature using a PL measurement system (PSI, Darsa-5000), equipped with a 200 mW IR laser diode (λ = 980 nm).

3. Results and discussion

LuAG:xYb³⁺,yEr³⁺ powders were prepared from various amounts of Yb³⁺ and Er³⁺. The ratio of Yb³⁺ to Er³⁺ (x/y = 6) was fixed, and the total amount (x + y) varied with m × (x = 0.18, y = 0.03), where m is 1, 3, 5, and 7. The XRD patterns of the synthesized powders correspond closely to those of ICSD #98-002-3846, and other impurity phases are not observed, as shown in Fig. 1(a). This indicates that a LuAG single phase was formed, and that the Yb³⁺ and Er³⁺ ions were fully incorporated into the LuAG lattice. LuAG has a cubic structure, which is described by the space group $Ia\overline{3}d$ (No. 230), in which each Lu atom is coordinated by eight oxygen atoms in a distorted cube; Al atoms occupy two non-equivalent sites, i.e., 16 octahedral and 24 tetrahedral sites [25]. For a coordination number (CN) of eight, Yb³⁺ and Er³⁺ have ionic radii of 0.985 Å and 1.004 Å, respectively, and can be readily substituted for the Lu³⁺ ions (r = 0.977 Å). The lattice constant, a, of LuAG:xYb³⁺,yEr³⁺ increases continuously [Fig. 1(b)] with an increasing amount of dopant (m), owing to these larger ionic radii (compared with that of Lu³⁺). This confirms that the Yb³⁺ and Er³⁺ ions were substituted for the Lu³⁺ ions.

 

Fig. 1 (a) XRD patterns and (b) dependence of the lattice constant of the Lu_3-x-yYb_xEr_yAl₅O₁₂ powders [prepared with m × (x = 0.18, y = 0.03)] on m.

Download Full Size | PDF

Figure 2(a) shows the UC PL spectra collected from the LuAG:xYb³⁺,yEr³⁺ powders, at varying m. The green and red regions give rise to weak and strong emission bands, respectively. These emissions originate from the Er³⁺ ions through an ET process from Yb³⁺ to Er³⁺. Several peaks are observed, in addition to the main peaks (526 and 557 nm for green; 679 nm for red). This large Stark splitting of the emission bands is attributed to the substitution of Lu³⁺ ions by Er³⁺ ions, which occupy the center of the distorted cube of LuAG. Furthermore, as Fig. 2(b) shows, the intensity of the red emission increases sharply by ∼700% with increasing m of 1 to 3, slightly at m = 5, and decreases at m = 7; this decrease results from the concentration quenching effect of the Er³⁺ ions. The intensity of the green emission increases sharply by ∼400% when m is increased from 1 to 3, saturates at m = 5, and decreases slightly at m = 7. This decrease was also observed in the case of the red emission intensity.

 

Fig. 2 (a) UC PL spectra and (b) intensities of the red and green emissions of Lu_3-x-yYb_xEr_yAl₅O₁₂ powders [prepared with m × (x = 0.18, y = 0.03)] under 980-nm radiation, as a function of m.

Download Full Size | PDF

The interaction mechanism between the Yb³⁺ and Er³⁺ ions was verified by determining the relationship between the UC emission intensity (I_em) and the input (excitation) power (P). The approximation, I_em ~Pⁿ, where n is the number of photons required for a UC process [15], was used during the calculations. Figure 3(a) shows the values corresponding to the 557 and 679 nm peaks of LuAG:0.9Yb³⁺,0.15Er³⁺ (m = 5), on a log-log scale. The corresponding slopes (n), 1.83 and 1.64, respectively, were estimated from a linear fit of the data and revealed that the emissions arose from a two-photon UC process. In addition, the estimated n values for various m, x, and y values confirmed that the two-photon UC process was involved in the emissions of LuAG:xYb³⁺,yEr³⁺. The schematic energy diagram of the two-photon UC process between Yb³⁺ and Er³⁺ is shown in Fig. 3(b). The UC emission behaviors of LuAG:xYb³⁺,yEr³⁺ can be explained by a sequential ETU process [15] as follows. Through ground-state absorption (GSA), the 980-nm radiation activates the sensitizers (Yb³⁺) to the excited state (²F_5/2). The absorbed energy at ²F_5/2 of the Yb³⁺ ions is then transferred to the Er³⁺ ions, via the two-photon ETU process. This transfer results in the multistep excitation of the Er³⁺ ions: ⁴I_15/2 → ⁴I_11/2 (GSA) and ⁴I_11/2 → ⁴F_7/2 (excited state absorption, ESA) through the ET(I) and ET(III) processes, respectively. Multiphoton relaxation then progresses sequentially from the ⁴F_7/2 to ²H_11/2, ⁴S_3/2, and ⁴F_9/2 levels of the Er³⁺ ions, as shown in Fig. 3(b).

 

Fig. 3 (a) UC emission intensities (I_em) at 557 and 679 nm as a function of the input power P of LuAG:0.9Yb³⁺,0.15Er³⁺ (m = 5) powders and (b) schematic energy diagram of a two-photon ETU process for Yb³⁺ and Er³⁺.

Download Full Size | PDF

In fact, the green and red emission peaks at 526/557 and 679 nm resulted from the transitions of the respective ²H_11/2/⁴S_3/2 and ⁴F_9/2 levels to the ground state (⁴I_15/2) of the Er³⁺ ions. The population at ⁴F_9/2(Er³⁺) arises from the ⁴F_7/2(Er³⁺) → ⁴F_9/2(Er³⁺) non-radiative relaxation. However, an additional route, i.e., non-radiative relaxation from ⁴I_11/2(Er³⁺) to ⁴I_13/2(Er³⁺), has been proposed in previous studies [12,13,15,16,24]. This relaxation, which occurs via cross relaxation (CR) between the Er³⁺ ions followed by a ⁴I_13/2(Er³⁺) + ²F_5/2(Yb³⁺) → ⁴F_9/2(Er³⁺) process through ET(II), can lead to enhanced red emission.

Emission intensity ratios of red to green (R_r/g = I_{679 nm}/I_{557 nm}) of ∼0.9, 1.7, 1.9, and 2.0 [Fig. 4(a)] were obtained for powders with m = 1, 3, 5, and 7, respectively. For m = 1, the intensity of the green emission was slightly stronger than that of the red emission. When m is increased to 3, however, the intensity of the red emission surpassed that of the green emission [Fig. 2(a)]. As a result, the R_r/g value increased sharply when m increased from 1 to 3, and smoothly thereafter. The emission colors lie within the yellow range of the Commission Internationale de l´Eclairage (CIE) chromaticity diagram [Fig. 4(b)], which shifted towards the red region with increasing m value.

 

Fig. 4 (a) Emission intensity ratios of red to green (R_r/g = I_{679 nm}/I_{557 nm}) and (b) CIE chromaticity coordinates of Lu_3-x-yYb_xEr_yAl₅O₁₂ powders [prepared with m × (x = 0.18, y = 0.03)] under 980 nm radiation.

Download Full Size | PDF

As mentioned in the ‘Introduction’ section, some Yb³⁺- and Er³⁺-doped UC phosphors have R_r/g values lower than unity [17–21] whereas others have values greater than unity [10–14,22–24]. The R_r/g values exhibit, therefore, a strong dependence on the host structure. For example, there are two polymorphs of NaYF₄: α (cubic) and β (hexagonal); the green emission from β-NaYF₄:Yb³⁺,Er³⁺is considerably stronger than the red emission [18,19], while the reverse is true for α-NaYF₄:Yb³⁺,Er³⁺ [22]. In this work, however, the change in the R_r/g value is attributed to the correlation between the UC process and the concentration of Yb³⁺ /Er³⁺. As shown in Fig. 4(a), the R_r/g value increased with increasing amount of (xYb³⁺ + yEr³⁺), at a constant ratio of x/y = 6. The dependence of the R_r/g value on the concentration of Yb³⁺ and Er³⁺ was verified by preparing LuAG:xYb³⁺,yEr³⁺ powders with various x/y ratios.

The intensities of the red/green emissions and the R_r/g value of LuAG:xYb³⁺,0.15Er³⁺ powders are shown in Figs. 5(a) and (b), respectively. As the figure shows, the intensities of both emissions increase apparently with increasing x of 0.8 to 0.9, and decreases sharply thereafter. This decrease results possibly from the modifications to the host crystal structure, owing to the high Yb³⁺ concentration. In contrast, the R_r/g value increases slightly with increasing x of 0.8 to 1.0 [Fig. 5(b)]. This behavior can be explained as follows. As previously mentioned [Fig. 3(b)], the population at ⁴F_9/2(Er³⁺) for the red emission results from ⁴F_7/2(Er³⁺) → ⁴F_9/2(Er³⁺) and ⁴I_13/2(Er³⁺) → ⁴F_9/2(Er³⁺) processes. These processes occur via ET(III) + multiphoton relaxation and CR + ET(II), respectively. Therefore, the populations at ⁴F_7/2(Er³⁺) and ⁴I_13/2(Er³⁺) are determined primarily by the ET from Yb³⁺ to Er³⁺ and the ⁴I_11/2(Er³⁺) level. The number of Yb³⁺ ions at the ²F_5/2 level has, accordingly, an effect on the intensity of the red emission of the synthesized powders. Moreover, ET(I) and ET(III) have an effect on both the green and red emissions and hence the increase in the Yb concentration preferentially enhanced ET(II) and CR (⁴I_11/2 → ⁴I_13/2) between the Er³⁺ ions. Previous studies also reported that the R_r/g value increased with increasing Yb³⁺ concentration [12,13,26,27].

 

Fig. 5 (a) Intensities of the red and green emissions and (b) emission intensity ratios of red to green (R_r/g = I_{679 nm}/I_{557 nm}) of the LuAG:xYb³⁺,0.15Er³⁺ powders.

Download Full Size | PDF

Figure 6(a) shows the intensities of the red and green emissions of the LuAG:0.9Yb³⁺,yEr³⁺ powders. These intensities increase sharply by ∼170 and 230%, respectively with increasing y of 0.1–0.15, owing to the increase in the concentration of the activator (Er³⁺). On the other hand, both of the intensities decrease with further increases in y (0.15−0.18). However, in contrast to those of the LuAG:xYb³⁺,0.15Er³⁺ powders, the R_r/g values decreased gradually [Fig. 6(b)] from 2.6 to 1.7 with increasing y of 0.1–0.18. These results indicate that the high concentration of Er³⁺ resulted in a concentration quenching effect and the back ET from Er³⁺ to Yb³⁺. This back ET resulted in a decrease in ET(II) and CR(⁴I_11/2 → ⁴I_13/2) between the Er³⁺ ions. The population at the ⁴F_9/2(Er³⁺) level, which arose via ⁴I_11/2(Er³⁺) → ⁴I_13/2(Er³⁺) → ⁴F_9/2(Er³⁺) transitions, was therefore significantly reduced, and most of the Er³⁺ ions, in turn, populated the ⁴F_7/2(Er³⁺) level. As such, non-radiative relaxation to the ²H_11/2(Er³⁺) and ⁴S_3/2(Er³⁺) levels exceeded relaxation to the ⁴F_9/2(Er³⁺) level, thereby leading to a decrease in the R_r/g values.

 

Fig. 6 (a) Intensities of the red and green emissions and (b) emission intensity ratios of red to green (R_r/g = I_{679 nm}/I_{557 nm}) of the LuAG:0.9Yb³⁺,yEr³⁺ powders.

Download Full Size | PDF

4. Conclusion

The Yb³⁺ and Er³⁺ ions were fully substituted for the Lu³⁺ ions, leading to an increase in the lattice constant, a, with increasing amounts of Yb³⁺ and Er³⁺. The UC emission spectra of the synthesized powders consisted of green and red emission bands that resulted from a two-photon ETU process from Yb³⁺ to Er³⁺. The green emissions were attributed to the ²H_11/2(Er³⁺)/⁴S_3/2(Er³⁺) → ⁴I_15/2(Er³⁺) transitions. However, the red emissions occurred via two routes, (I) and (II): (I) ⁴I_11/2(Er³⁺) → ⁴I_13/2(Er³⁺) → ⁴F_9/2(Er³⁺) → ⁴I_15/2(Er³⁺) and (II) ⁴I_11/2(Er³⁺) → ⁴F_7/2(Er³⁺) → ⁴F_9/2(Er³⁺) → ⁴I_15/2(Er³⁺). Route (I) was preferentially enhanced with increasing x, leading to an increase in the R_r/g value. In contrast, R_r/g decreased with increasing y owing to the diminution of route (I). The corresponding CIE chromaticity coordinates of the synthesized powders were all located within the yellow region and shifted towards the green or red regions, depending on the R_r/g value.