Complex study on photoluminescence properties of YAG:Ce,Gd phosphors

Luminescence characteristics of gadolinium co-doped yttrium aluminium garnet doped with cerium phosphors were studied. In this work, powder X-ray diffraction (XRD) spectra, elemental composition analyses, excitation and emission spectra, conversion efficiency of emission phosphor, corresponding (CIE) chromaticity colour coordinates and pulsed photoluminescence decay kinetic curves were investigated, all the measurements were performed at room temperature. The properties of the phosphors were studied by comparing the composition of the phosphors and their luminescent properties.


Introduction
Yellow colour phosphors based on yttrium aluminum garnet (YAG, Y 3 Al 5 O 12 ) are promising materials because of their good chemical and physical properties: excellent chemical stability at elevated temperature and under radiation stimulation, good emission characteristics under excitation.YAG phosphor is used to convert electrons to visible light in displays [1][2][3][4][5]. Ce 3+ -activated phosphor Y 3−x Al 5 O 12 :xCe 3+ (YAG:Ce) as a well-known efficient phosphor material is used to convert ultraviolet and blue LED radiation into yellow emission. It is assumed that the excitation of Ce 3+ ion causes maximum absorbance in the blue region resulting in converting InGaN based blue LED radiation into a very broad intense yellow emission band and then combing the yellow emission with the nonabsorbed blue emission from the blue LED to generate white light. However, it shows a low color rendering index (CRI). Thus in order to improve the quality of white light YAG:Ce can be doped with other rare earth ions, such as Gd 3+ and Tb 3+ [6]. Gd 3+ (Gd 3+ i.r. = 0.938 Å) has been added to substitute at the Y 3+ (0.9 Å) site, which attributes to lattice distortion in the region of Ce 3+ and shift of the emission band to a longer wavelength.
In the present study, industrial YAG phosphors with different concentration of Gd3+ were investigated.

Experimental details
Industrial YAG phosphors of SDL4000, SDL2700 and SDL3500 with different ratio of elemental composition were used for the present research. Information on the test samples is shown in Table 1. The elemental composition analyses of the samples were performed by a x-ray photoelectron spectroscopy (XPS) ESCALAB 250 with a PHI 5000Versaprobe-Scanning ESC Amicroprobe. The measurements were carried out by a monochromatic Al Ka X-Ray source (500 µm, 60 W, 10 kV) with photon energy of 1486.6 eV. Survey scans (binding energies ranging from 0 to 1400 eV) were performed using a 1 eV/step, while for higher resolution spectra of the Gd 4d, Ce 3d,O 1s, Al 2p,C ls, and Y 2d peaks, the hemispherical analyzer pass energy was maintained with 0.05 eV/step. The phosphors of SDL2700 and SDL3500 contain Gd and Ce except for Y, Al and O. Gd and Ce in phosphor SDL4000 were not detected because of its small content. We emphasize that the used method allows us to determine the composition of the sample on its surface.
The crystal structure of the measured phosphors was identified by using the powder X-ray diffraction (XRD) Rigaku smart lab with Cu, Ka radiation generated at 40 kv/30 mA. The recorded XRD spectra were covered from 10° to 60° with increments of 0.02°.
The photoluminescence spectra of the samples were recorded by FLS980 spectrometer (Ediburgh instruments) with a xenon lamp (250-1000 nm) as an excitation source. CIE color coordinates in the yellow light region were determined by luminescence spectra of the investigated phosphors.  Figure 1 shows the XRD patterns of tested phosphors. For comparison, a standard YAG phase (PDF 09-1316 for Y 3 Al 5 O 12 ) was also inserted. From the inset in figure 1, the diffraction peaks are in line with that of the standard YAG phase. It is clear that no impurity phases were formed, which are possible due to the dopant Ce 3+ and Gd 3+ ions. It can be found that the diffraction peak of (420) moves to the lower angle side with the increase of Gd 3+ content relative to the position in the crystal. It is assumed [7] that the shift of the peak (420) is caused by lattice distortion and substitution of ions Y 3 + (0.92Å) by ions Gd 3 + (0.938Å).   Figure 2(а) shows excitation spectra of tested phosphors at the fixed emission wavelength of 560 nm, two bands can be seen in the region between 300 and 500 nm and these are peaked at 344 and 454 nm.

Photoluminescence spectra and excitation spectra
The peak wavelength at 454 nm matches well with that of the light emitted by InGaN based blue LED; therefore, YAG phosphors can efficiently absorb blue light of the chip and convert it into a higher wavelength region. Figure 2(b) (c) shows the photoluminescence (PL) emission spectra of YAG and YAG:Ce codoped with Gd at a fixed excitation wavelength of 340 nm and 460 nm. The PL emission spectra show a broad emission band, which is supposed to be caused by Ce 3 + ions with emission maxima occurring at 560 nm. It is known that Ce 3+ is sensitive to the local crystalline field environment; therefore, lattice distortion by Gd ion substituting the Y 3+ ion causes the shift in the emission band. It was observed from Figure 2   The measurement results obtained for the spectral characteristics of the studied phosphors are summarized in Table 2. The positions of the luminescence bands do not depend on the area of the luminescence spectrum excited and the type of excitation. We can detect the following changes in the luminescence and excitation spectra.

Evaluation limiting values of light output at conversion of spectra in LED
The limiting values of the light output during spectrum conversion in phosphor were evaluated. The evaluation was carried out under the following conditions. We only take into account the energy losses in the process of converting chip radiation into luminescence in phosphor. In the spectrum of "white" light of LED is considered the proportion of "blue" radiation of chip that involved in the formation of the spectrum. The calculation method is described in [8], and the calculations were performed for two variants: the luminescence of phosphor is excited under λ ex =344 nm in the UV radiation region and in the blue radiation region under λ ex =454 nm. In this work, it was assumed that the value of emission intensity at peak wavelength λ=454 nm equals to the peak luminescence of phosphors at λ em =560 nm under λ ex =454 nm. The results are presented in Table 3. As follows from the calculated results, the values of energy losses in the process of converting chip radiation under λ ex =344 nm into luminescence of phosphor in the region of λ em =560 nm are not less than 44%, the luminous efficiency of the LED cannot be more than 280 lm/W. Accordingly, under λ ex = 454nm, the values of energy losses are more than 27%, the luminous efficiency of the LED cannot be more than 350 lm / W. Table 4 shows the chromaticity coordinate of phosphor emission. With the increase in the content of Gd 3 + emission spectrum is approaching the ideal range of white light with x = 0.33 and y = 0.33.

The light output of LED with phosphors
The integrated spectral efficiency was measured by using an integrating sphere and calibrated spectrophotometer AvaSpec-ULS3648. The energy output η is the radiation flux ratio of phosphor to absorb the excitation flux. Two types of chips were used for excitation, which generate radiation with a stream of 173 µW/ cm 2 in the band with a maximum at 460 nm and 16 730 µW/ cm 2 at 447 nm. Power of LED are almost 100 times different. As follows from the results, radiation energy output η almost independent on the power of the exciting radiation during the conversion process of SDL4000 phosphor. The change does not exceed 2.5%. However, he energy output of SDL2700 phosphor by 22% and in the SDL3500 phosphor by 18% is reduced, with increase power of LED in 100 times. This difference in energy output can be attributed to the fact that phosphors are sensitive to temperature variation.

Conclusions
Complex research on a group of phosphors based on YAG with different content of Ce 3 + activator and co-activator Gd 3+ was conducted. These impurities are intentionally not introduced to the SDL4000 phosphor, in which Ce 3+ and Gd 3+ were not detected by using XPS method. All the studied phosphors have a crystal structure in accordance with a typical body centered cubic structure of YAG (Y 3 Al 5 O 12 ); however, lattice distortion causes the shift of the peak (420). It is assumed that the lattice distortion is caused by substitution of ions Y 3 + (0.92Å) by ions Gd 3 + (0.938Å). This explanation is well correlated with the measurement results obtained for XRD patterns of the studied phosphors. However, the peak shift in the SDL4000 phosphor relative to the peak in the crystal is evidently due to the other factors. Perhaps crystal lattice distorted in the SDL4000 phosphor is formed in the synthesis of intrinsic defects lattice.
Upon excitation in the region of 344 nm and 454 nm, the position of luminescence bands of all investigated phosphors respectively at 580 nm of SDL2700, 560 nm of SDL3500 and 540 nm of SDL4000 phosphor. With the increase of Gd 3+ content, the emission wavelength shows a great red shift, which is due to the crystal lattice distortion formed in the region of the luminous center. It is assumed that the emission center of phosphors is Ce 3+ ion. Apparently, luminous center of the SDL4000 phosphor has the same nature, though Ce 3+ is not intentionally introduced during synthesis. It is interesting to observe that the luminescence band of SDL4000 phosphor under λ ex = 344 nm is broader as compared to that under λ ex = 454nm. This difference can be explained either by the presence of two different crystal structures of luminous centers, or one type of luminous centers with different environments. One of the centers is dominant in SDL2700, 3500 phosphors.
It can be considered that the luminescence characteristics depend not only on the composition of the doping activators and co-activators, but also on the presence of high concentrations of native defects in the crystal which were included during synthesis. It is difficult to ensure compliance with stoichiometry during formation of the crystal with a complex structure. Therefore, native defects are introduced in the crystal during synthesis. Native defects together with activators and co-activators can form complex defects which are called nanodefects [9].