1. Introduction

As a commercially available solid-state source for general light sources, GaN-based white-light light-emitting diodes (w-LEDs) are particularly desirable because of their long lifetime, small sizes, high reliability, low consumption of energy and environmentally friendly [1, 2]. In general, w-LEDs can save approximately 70% of the energy used for incandescent light bulbs [3] and do not need the hazardous mercury commonly used in luminescent tubes [2]. Generally, two ways can be employed in obtaining white light: one is mixing three monochromatic LED chips; the other is using phosphors to convert ultraviolet or blue light into white light, which is called as phosphor-converted LEDs (pc-LEDs). However, expensive cost and complicated electro-circuit operation system are needed for the first method. Thus, pc-LEDs are recognized as a powerful technique to achieve white light and regarded as a new lighting source for the next generation.

Currently, commercial pc-LEDs use blue InGaN LED chip and Y₃Al₅O₁₂:Ce³⁺ (YAG:Ce) garnet phosphor, whose efficacies can be greater than 80 lm/W for 1 W devies, higher than compact fluorescent lamps (CFLs) and comparable to linear fluorescent lamps [4]. However, YAG:Ce has a deficient red emission, leading to its bluish-cold light owing to its high color temperatures (CCTs ≈7750 K) and low color rendering index (CRI ≈70-80). But, high CCTs and low CRIs cannot meet for many sophisticated applications such as light sources in offices, schools, medicals, hospitals and hotels. To obtain lower CCTs and higher CRIs, enhanced red emission is needed in YAG:Ce phosphors. In summary, there are two ways to enrich the red emission in YAG:Ce. One is by co-doping rare earth ions which can emit red emission, such as Cr³⁺, Pr³⁺ and Sm³⁺ [5–8]. However, the red emission cannot be greatly enhanced due to weak absorption in blue region and narrow emission peaks, meanwhile the emission intensity of Ce³⁺ decreases greatly. The other is to replace Y³⁺ or Al³⁺ by other cations to change the crystalline field around Ce³⁺ and make the emission spectrum shift to longer wavelength, such as Tb³⁺, Gd³⁺, Mg²⁺-Si⁴⁺ [5, 9, 10]. This method also causes the emission intensity of Ce³⁺ decrease, while the shift of Ce³⁺ emission is very limited.

In view of recently research, Mn²⁺ was co-doped with Ce³⁺ and Eu²⁺ to enhance red spectra [11–14]. As we know, Mn²⁺ can give rise to a broad emission ranging from green to red, depending on the host for the sensitivity of the d-d transition ⁴T₁(G)-⁶A₁(G) to the crystal field [15–17]. Besides, Mn²⁺ doped samples can absorb blue and green light, which is not only good for blue LED chip but also helps to obtain orange and red light. To the best of our knowledge, rare reports have been found on the detailed photoluminescence (PL) properties of Mn²⁺ co-doped YAG:Ce phosphors [18–20]. In Ref. 18, the emission spectrum of YAG:Ce, Mn without charge compensator was reported sketchily. In Ref. 19-20, authors just investigated the performance of YAG:Mn. Until now, no one has considered the absorbing effects of Mn²⁺ and the effect of different charge compensators. Thus, we introduced Mn²⁺ into YAG:Ce to obtain color tunable phosphors, using quadrivalence ions to balance the charge difference. The emission bands of Mn²⁺ in different sites and the effect of different charge compensators on color adjustment are discussed in the present work.

2. Experiment

2.1 Materials and synthesis

All samples were synthesized using solid-state reaction. Raw materials Y₂O₃ (A.R. 99.9%), Al₂O₃ (A.R. 99.9%), CeO₂ (A.R. 99.99%), H₂SiO₃ (A.R. 99.9%) and C₄H₆MnO₄·4H₂O (A.R. 99.9%) were accurately weighed according to the composition of (Y_2.94Ce_0.06)(Al_5-2xMn_xSi_x)O₁₂ (0≤x≤0.9) (YAG:0.06Ce, xMn, xSi). A small amount of BaF₂ (A.R. 99.99%) was added to the raw materials as a flux. Y₃(Al_4.8Mn_0.1Si_0.1)O₁₂ (YAG:0.1Mn, 0.1Si), (Y_2.94Ce_0.06)(Al_4.4Mn_0.3Ge_0.3)O₁₂ (YAG:0.06Ce, 0.3Mn, 0.3Ge) and (Y_2.94Ce_0.06)(Al_4.4Mn_0.3Zr_0.3)O₁₂ (YAG:0.06Ce, 0.3Mn, 0.3Zr) samples were synthesized using GeO₂ (A.R. 99.99%) and ZrO₂ (A.R. 99.99%) as raw materials. The weighed materials were put into an agate mortar and ground to mix them thoroughly. Then the mixtures were heated up to 1500 °C for 5 h in horizontal tube furnaces. A weak reducing atmosphere with flowing H₂ (5%) and N₂ (95%) was employed in order to reduce Ce⁴⁺ to Ce³⁺ and prevent Mn²⁺ from being oxidized. We also prepared YAG:0.1Mn, 0.1Si in air as a reference.

2.2 Measurements and characterization

The crystal structure of the synthesized samples was identified by using a Rigaku D/Max-2400 X-ray diffractometer (XRD) with Nifiltered Cu Kα radiation operating at 40 kV and 60 mA. The photoluminescence (PL) and PL excitation (PLE) spectra measurements were made on powders pressed into an aluminum plaque using a FLS-920T fluorescence spectrophotometer equipped with a 450W Xe light source and double excitation monochromators. The quantum efficiency (QE) was measured by a Fluorlog-3 spectrofluorometer equipped with 450 W xenon lamp (Horiba Jobin Yvon). All the above measurements were performed at room temperature. High temperature luminescence intensity measurements were carried out by using an aluminum plaque with cartridge heaters, the temperature measured by thermocouples inside the plaque and controlled by a standard TAP-02 high temperature fluorescence controller.

3. Results and discussion

All synthesized YAG:0.06Ce, xMn, xSi (0≤x≤0.3) samples are assigned to pure YAG, according to JCPDS file 33-0040. For 0.3<x<0.9, primary phase of YAG with a few by-products of Ce₂SiO₅ (JCPDS file 48-0054) was obtained, which can be explained as parts of Mn²⁺ ions occupy Y³⁺ sites, resulting in the surplus of Ce³⁺. Fortunately, the content of impurity (<5%) is low enough to neglect its effect, shown in Fig. 1 .

 

Fig. 1 XRD patterns of YAG:0.06Ce, xMn, xSi (0≤x≤0.9) samples.

Download Full Size | PDF

Figure 2(a) shows the PL and PLE spectra for YAG:0.06Ce. The PL spectrum of YAG:0.06Ce shows a typical yellow band peaked at 560 nm, originating from the transitions form 5d to 4f of Ce³⁺ [5–7]. The PLE spectrum for the yellow emission appears as an intense blue band around 460 nm with a weak band around 340 nm. The PL spectrum of YAG:0.1Mn, 0.1Si shown in Fig. 2(b) exhibits three emission bands at around 528 nm, 593 nm and 745 nm, named as Mn²⁺ (I), Mn²⁺ (II) and Mn²⁺ (III), respectively under blue light excitation. When YAG:0.1Mn, 0.1Si was synthesized in air, no visible luminescence band was seen under 460 nm excitation. Thus, these bands should be attributed to the spin-forbidden ⁴T₁(4G)-⁶A₁(6G) transition [14–16] of Mn²⁺ into three kinds of sites, related to the crystal structure of YAG host [21]. Considering that the position of Mn²⁺ emission bands depends strongly on the host lattices, if the crystal field around Mn²⁺ is weak, the splitting of the excited d energy levels will be small, resulting in Mn²⁺ emission with higher energy [17]. Thus, the green, yellow and deep red emission bands in YAG:0.1Mn, 0.1Si emission spectrum can be attributed to Mn²⁺ ions occupying Y³⁺, Al³⁺(I) (with six-coordination numbers) and Al³⁺(II) (with four-coordination numbers) sites, respectively. Simultaneously, from Fig. 2(b), the yellow emission intensity is much stronger than others, indicating Mn²⁺ ions are more inclined to occupy Al³⁺(I) sites. The PLE spectra of YAG:0.1Mn, 0.1Si consists of several weak peaks centering at around 422, 466 and 508 nm corresponding to the transitions from ⁶A₁ to ⁴E(⁴G), ⁴T₂(⁴G) and ⁴T₁(⁴G), respectively [15, 16]. When monitored at 593 nm and 745 nm, the relative intensities of excitation peaks are different, which further proves that Mn²⁺ ions occupying different sites. The comparison between the PL spectrum of YAG:0.06Ce and the PLE spectrum of YAG:0.1Mn, 0.1Si in Fig. 2(c) reveals a great spectral overlap. Accordingly, efficient resonance-type energy transfer (ET) from Ce³⁺ to Mn²⁺ is expected.

 

Fig. 2 PL and PLE spectra of YAG:0.06Ce (a) and YAG:0.1Mn, 0.1Si (b), the comparison between the PL spectrum of YAG:0.06Ce and the PLE spectrum of YAG:0.1Mn, 0.1Si (c).

Download Full Size | PDF

Mn²⁺ co-doped YAG:0.06Ce samples will lead the charge mismatch. Thus, quadrivalence ions are needed as charge compensators. Besides, Anant A. Setlur et al [4] reports that co-doped N^3- in YAG:Ce samples using Si₃N₄ as N source gave an additional Ce³⁺ absorption bands, whereas using AlN as N source, no additional bands can be detected, but they didn’t further investigate it. Accordingly, it is reasonable to infer that the presence of charge compensators may play an important role. Thus, samples without charge compensator and with different charge compensators are synthesized as control group. Figure 3 shows the PL spectra of a typical concentration of Mn²⁺ co-doped YAG:0.06Ce samples with different charge compensators (the red line), using YAG:0.06Ce (dot line) as a reference. It is observed from Fig. 3 that the effect of different quadrivalence ions on color adjustment through controlling the Mn²⁺ emissions in YAG:0.06Ce, xMn followed the order Si⁴⁺>Zr⁴⁺>no compensator>Ge⁴⁺. The emission intensities are also following this order (not given in this paper). But Si⁴⁺ singly doped one shows blue shift, and no excess band can be seen.

 

Fig. 3 PL spectra of a typical concentration of Mn²⁺ co-doped YAG:0.06Ce samples with different charge compensators (the red line) using YAG:0.06Ce (the dot line) as a reference.

Download Full Size | PDF

The above phenomenon can be explained as follows. According to ionic radii and the PL spectra, it is assumed that most of the Mn²⁺ ions come into Al³⁺(I) sites and a small quantity enter into Al³⁺(II) sites, which makes the octahedron and tetrahedron expand. As the radius of Si⁴⁺ is close to that of Al³⁺, Si⁴⁺ ions are more tend to come into Al³⁺ (II) sites, resulting in the tetrahedron shrink. Whereas, the radius of Zr⁴⁺ is more close to that of Y³⁺, so Zr⁴⁺ ions come into Y³⁺ sites, which makes the dodecahedron shrink. So both Si⁴⁺ and Zr⁴⁺ as charge compensators doped into YAG:0.06Ce, Mn samples will result in minimum volume change. That means that Si⁴⁺ and Zr⁴⁺ act not only as charge compensators but also as volume compensators. However, due to the almost same ionic radii between Ge⁴⁺ and Al³⁺, Ge⁴⁺ are easier to occupy both Al³⁺ sites than Mn²⁺, resulting in decreasing the doping concentration of Mn²⁺ into Al³⁺ (II) and Al³⁺ (III) sites. In summary, different quadrivalence ions as compensators can adjust the number of Mn²⁺ doped into Al³⁺ (II) and Al³⁺ (III) sites, which can be further proved by XRD patterns. Almost all samples are single-phase compared with JCPDS No. 33-0040, except for Y_2.94Ce_0.06Al_4.4Mn_0.3Ge_0.3O₁₂, shown in Fig. 4 . When Mn²⁺ or Si⁴⁺ singly doped into YAG:0.06Ce samples, the diffraction peaks shifting to a larger angle compared with standard diffraction, indicating part of Mn²⁺ ions doped into Y³⁺ sites, agreed with the spectral phenomena. For Si⁴⁺ and Zr⁴⁺ co-doped YAG:0.06Ce, xMn samples, the shift of diffraction peaks is smaller than that of Mn²⁺ or Si⁴⁺ singly doped into YAG:0.06Ce samples, and without any impurity phase. However, for Y_2.94Ce_0.06Al_4.4Mn_0.3Ge_0.3O₁₂ sample, no shift is seen, but there is little impurity phase exist. These phenomena agree with above description.

 

Fig. 4 XRD patterns of Mn²⁺ co-doped YAG:0.06Ce samples with different charge compensators.

Download Full Size | PDF

In order to confirm the concentration of Mn²⁺, the PL spectra of a series of samples were shown in Fig. 5 . The PL spectra of all samples upon Ce³⁺ excitation at 460 nm not only exhibits the Ce³⁺ emission at 560 nm but also the Mn²⁺ emission at 593 nm (Mn²⁺(II)) and 745 nm (Mn²⁺ (III)). The Mn²⁺(I) emission band cannot been resolved clearly owing to overlap with Ce³⁺ emission band, and also Mn²⁺(I) band is located at green light region, it is not discussed in the present work. When x≤0.1, there are mainly two bands, the Ce³⁺ emission band and the Mn²⁺(II) band. When 0.1<x≤0.3, Mn²⁺(III) band is dominating. The emission intensity of Mn²⁺ increases, while the emission intensity of Ce³⁺ decreases with increasing its concentration x. These results give strong evidence of the effective Ce³⁺-Mn²⁺ ET, and also further prove that Mn²⁺ ions are inclined to occupy Al³⁺(I) sites, until saturation, then Al³⁺(II) sites are occupied. Besides, when x>0.3, the emission intensity of three bands (both Ce³⁺ and Mn²⁺) decrease, which is due to the concentration quenching effect.

 

Fig. 5 PL spectra of YAG:0.06Ce, xMn, xSi (0≤x≤0.9) samples.

Download Full Size | PDF

Figure 6 shows the Commission International de L’Eclairage (CIE) chromaticity coordinates and appearances of YAG:0.06Ce, xMn, xSi samples. From the photograph in Fig. 6, we can see that the color of synthesized samples changed from yellow to orange, even to deep red. Energy transfer between Ce³⁺ to Mn²⁺ makes it possible to obtain the greenish-yellow emission band of Ce³⁺, the orange and deep red emission bands of Mn²⁺ in single YAG host. But the CIE chromaticity coordinates shift to orange range with increasing the Mn²⁺ concentration until x = 0.3. Red color does not reflect in the CIE chromaticity coordinates. The reason is that the red emission in visible part of the spectrum is not enriched, because of Mn²⁺ related red emission peak at 750 nm, but unfortunately, 690 nm is nearly the cut-off wavelength for the CIE 1931 color-matching function x. Human eyes do not really effectively perceive anything longer than 690 nm. As a consequence, the improvement of the PL intensity, which all happens for wavelengths longer than 690 nm, contribute to neither the calculation of CIE 1931 chromaticity coordinates x and y, nor the calculation of the luminous flux. The obtained red color from eye should due to absorbing the blue and green emission faster than absorbing the orange and red emission by increasing Mn²⁺ concentration, which can be easily understood by Fig. 2(b). Therefore, emission color of YAG:0.06Ce, xMn, xSi samples can be adjusted by controlling the content of Mn²⁺ to meet different needs of illumination applications from two aspects.

 

Fig. 6 The CIE chromaticity diagram and photograph of YAG:0.06Ce, xMn, xSi (0≤x≤0.9) samples.

Download Full Size | PDF

As for practical LEDs application, the proposing phosphors should meet the requirements below: When measured at 150 °C, the change of efficiency should be lower than 10% and CIE changes (both △CIE x and △CIE y) should be lower than 0.015. Also the quantum efficiency of phosphors is higher than 70%. Thus, we measured the temperature quenching characteristics and the quantum efficiency. For YAG:0.06Ce, 0.04Mn, 0.04Si sample, the integral emission intensities from 480 nm to 800 nm can reach as high as 101% compared with commercial YAG:Ce phosphor (com-YAG:Ce) at the same excitation condition. As for com-YAG:Ce and YAG:0.06Ce, 0.04Mn, 0.04Si samples, when measured at 150 °C, the emission intensities are 80.2% and 77.6%, the △CIE x are 0.002 and 0.007, the △CIE y are both 0.003 (shown in Fig. 7 ), which are all lower than 0.015. And the quantum efficiencies of are 89.7% and 70.2%, respectively. All these characterizes show that YAG:0.06Ce, 0.04Mn, 0.04Si sample can be used as candidate to obtain warm white light.

 

Fig. 7 Temperature-dependent emission of YAG:0.06Ce,0.04Mn, 0.04Si sample, the inset shows temperature-dependent emission intensities of Com-YAG:Ce and YAG:0.06Ce,0.04Mn, 0.04Si samples (a), the △CIE x and △CIE y and temperature characteristics for YAG:0.06Ce, 0.04Mn, 0.04Si sample and com-YAG:Ce phosphor (b).

Download Full Size | PDF

4. Conclusions

Tunable luminescence YAG:0.06Ce, xMn phosphors with different charge compensators are prepared by solid state reaction. The emission spectra of Mn²⁺ singly doped YAG samples show three bands, and the orange band is the strongest one, which means Mn²⁺ ions not only occupy two kinds of Al³⁺ sites, but also occupy Y³⁺ sites in YAG host, and the most inclined sites are Al³⁺(I) sites. Quadrivalence ions including Zr⁴⁺, Ge⁴⁺ and Si⁴⁺ are introduced to balance the charge difference between Mn²⁺ and Al³⁺. Among them, Si⁴⁺ as charge compensator exhibits the best tunable effect on controlling the Mn²⁺ emissions in YAG:Ce, xMn, the reason is due to its suitable ionic radius of Si⁴⁺. For Y₃Al₅O₁₂:0.06Ce, 0.04Mn, 0.04Si sample, the integral emission intensity is 101% of that of com-YAG:Ce phosphor and it has better thermal quenching characteristics than com-YAG:Ce phosphor. The Commission International de L’Eclairage (CIE) chromaticity coordinate is (0.453, 0.526), which is redder than that of YAG:0.06Ce phosphor (0.436, 0.540). Therefore, warm white LEDs are expected using the single Y₃Al₅O₁₂:0.06Ce, 0.04Mn, 0.04Si phosphor.