Low-energy Cathodoluminescence for (Oxy)Nitride Phosphors

The luminescence is distributed not only laterally but also in-depth. Such core-surface distribution is observable with CL by changing the electron energy, since it varies the penetration depth of the incident electrons²¹. However, the penetration depth varies for each material and the correspondence between electron energy and penetration depth is not linear, and may introduce some additional effects, such as the reabsorption of higher energy photon from the deeper regions by the material itself. Thus, it may be preferable to observe directly the core-surface distribution via cross-section observation. In case of phosphor powders, such observation can be achieved by trapping the particles into a resin and polish the powder-resin composite by cross-section polisher, for instance. Since the particles are randomly distributed in the resin, the cutting direction is not controllable. However, the high amount of particles allows cutting enough particles to make such an analysis valid.

To illustrate this point, we have investigated the luminescence distribution of Si-doped AlN powder. Figure 5a shows the CL spectra of AlN powders doped with 0.0% and 1.6% of Si. The emission of undoped AlN consists of 2 bands at 350 and 380 nm, while that of AlN doped with 1.6% of a band at 350 nm with a clear shoulder at 280 nm. The 350 and 380 nm bands are attributed to Al vacancy-oxygen complexes (O_N-V_Al), while the shoulder at 280 nm with the O-purified AlN affected by Si to form SiO vapor²². Figure 5b and 5c show the CL images taken at 280 nm for as-sintered and cross-sectioned AlN powders doped with 1.6% of Si, respectively. The 280 nm emission is nonuniform along the particles from the CL image of the as-sintered sample, the brighter areas seem to be at the edges of the particles, but the morphologies of the particles and their distribution may make such observations not so obvious. However, from the CL image of the cross-sectioned sample, it appears clearly that the 280 nm emission is mainly localized at the surfaces of the particles, suggesting that AlN particles are actually coated by a Si-rich layer and surface purification may proceed.

Local composition changes in Sialon phosphors can drastically affect the luminescence properties. Thus, the same rare-earth ion in different host-lattices or in different sites may give different emissions^15,18-20. But, unfortunately, local differences during the sintering, such as a distribution of the temperature or the raw materials proportion, or the partial oxidation of the surface of the particles, are expected, resulting in changes of the composition along the particles and/or in the coexistence of several phases. Such effects may not be directly observable with structural and chemical characterization techniques. Thus, it is important to investigate the local luminescence properties of a phosphor. With the precise control of the size and position of the electron-beam in a SEM, it is possible not only to acquire a CL spectrum from a nanoscale region but also to obtain high-resolution CL images of the luminescence centers.

(La,Ce)Al(Si_6-zAl_z)(N_10-zO_z) (z~1) (JEM) is an intense blue phosphor which is suitable for general lighting. It has been found that by replacing (La, Ce) by Ca, a red shift and a broadening of CL peaks occur according to Ca-doping. It was believed that Ca was affecting the crystal field splitting of Ce³⁺. However, this explanation, based only on the luminescence spectra, is misleading, as revealed by cross sectional local analysis.¹⁵ Figure 6 shows the cross-sectional SE (CS-SE), combined CS-CL images at 300 nm (red), 430 nm (blue) and 540 nm (green) and local CL spectra taken at 5 kV for JEM phosphors doped with 0 (a, b, c) and 0.69 (d, e, f) at. % of Ca, respectively. It has to been noted that these wavelengths were selected in order to reduce the bands overlapping in case that several bands exist. For CS-CL image of Ca-undoped sample, JEM particles consist of many particles agglomerated with each other. The luminescence at 430 nm is almost uniformly distributed with some brighter area and some of localized area at 300 nm. On the other hand, the grain boundaries show darker emission. Local CL analysis reveals that the spectral shape is relatively comparable in any positions, with a shift of the bands from 430 to 450 nm and spectral intensity in a good agreement with the pictures. For CS-CL image of Ca-doped, there are significant differences between 430 and 540 nm. Submicron patches clearly appear brighter at 300 and 540 nm along a large portion of the particles with darker grain boundary regions, while the 430 nm emission is localized in another section of the particles. By local analysis, the CL spectrum taken on a 430 nm bright area (point 3) consists a band at 440 nm, comparable at the one observed for Ca-undoped sample. The bright areas at 540 nm, embedded in the same particle, (points 1 and 4) show a band at 480-490 nm. Small bright areas at 300 nm (point 2) and dark grain boundary region (point 5) show a band at 440 nm with shoulder at 480 nm, with occasionally at emission at 310 nm. Based on the literature and XRD analysis, we can attribute the band centered at around 430 nm to Ce³⁺ in JEM²² and that at 480 nm to Ce³⁺ in α-SiAlON²³. The darker broad emission is originated to Ce³⁺ in β-SiAlON, and that at 310 nm to Sialon host material. These results prove that the red shift and the broadening of CL peaks according to Ca doping may not be attributed to Ca-induced changes in the crystal field splitting of Ce³⁺ as initially thought, but more to the coexistence of different phases inside the same particles and the gradual transformation of β-SiAlON to α-SiAlON with Ca doping.

Although the observation of the different emission centers and their distribution is possible by using low-energy CL, it may not be enough to fully understand the nature of the luminescence centers. In such cases, it is necessary to combine the CL measurements with other techniques. Since the incident electrons can generate other signals beside CL, it is possible to directly correlate the light emission with electrical, chemical or structural properties by investigating the same area with the different electron-beam techniques. Thus, the correlation of CL with high-resolution TEM (HRTEM) and EBIC has been used to characterize defects, such as dislocations or stacking faults. As for the variation of concentration/composition, the combination of CL with TEM, EDS or Auger spectroscopy can result in a better understanding of the origin of the luminescence.

Here, we illustrate this aspect by investigating the emission of Si-doped AlN powder. Figure 7 shows the CS-CL and CS-EDS images (a, b) and local spectra (c, d) of AlN particle doped with 4.0% Si doping. The CS-CL image was taken at 350 nm, while the CS-EDS image consists of the superposition of Si and Al distribution. The CS-CL image shows darker elongated structure in the center of the particles. Local CL spectra taken in the bright region consist of a strong peak at 350 nm with shoulders at 280, 380, and 460 nm. However, there are clear changes in the ratios between these different bands with the position. Areas showing a brighter emission at 350 nm (point 1) shows a higher 280 nm emission and smaller 460 nm emission compared to the main band at 350 nm, while the darker elongated patch (point 2) shows a smaller 280 nm emission and higher 460 nm emission compared to the main band at 350 nm. The 460 nm is originated from Si-accommodating defects in AlN²⁴. EDS images and local spectra reveal that the darker elongated area show a smaller Al and higher Si composition compared to the rest of the particles. Compared with the results observed in Figure 5, we can assume that by increasing the amount of Si into AlN, a secondary reaction occurs between Si and AlN, which induces the formation of SiAlON phases.

The two important parameters for a material used in devices are materials high performance and stability under stress. Indeed, a degradation of the material properties under stress will reduce its lifetime, which is not industrially viable. Thus, for electron-beam stimulated devices, such as cathode ray tubes (CRTs) and FEDs, it is necessary to develop electron-beam irradiation resistant phosphors and/or to understand the electron-beam induced mechanisms in order to prevent or reduce such effects. The luminescence degradation can occur via different mechanisms, such adsorption/desorption or charging at the surfaces, creation or activation of defects, etc.^25-27. Although these intensity variations complicate the quantitative analysis of CL results, they can be used to investigate the lifetime of optoelectronic devices.

To illustrate this point, we have the CL spectra and evolutions of two blue-emitting phosphors, Ce-doped La₅Si₃O₁₂N and Si/Eu-doped AlN. Figure 8a shows the CL spectra for Ce-doped La₅Si₃O₁₂N and Si/Eu-doped AlN after 20 sec of irradiation at 5 kV. Both samples show an intense blue emission: the band position and intensity for Ce-doped La₅Si₃O₁₂N are 456 nm and 3,270 cps, respectively, while those for Si/Eu-codoped AlN are 466 nm and 3,100 cps. A priori, the main difference between these 2 samples is the broadness of the emission, since the emission for Ce-doped La₅Si₃O₁₂N is larger due to the coexistence of several bands. Thus, it seems that both materials are suitable as blue-emitting phosphors for FEDs, and that we have to consider criteria as fabrication cost, the compatibility with the other phosphors or the stability of the luminescence properties under electron-beam irradiation, to determine the most suitable. Figure 8b shows the evolutions of CL intensity of Ce-doped La₅Si₃O₁₂N and Si/Eu-doped AlN during electron-beam irradiation at 5 kV. For Ce-doped La₅Si₃O₁₂N, the intensity decreases from 3,270 to 450 cps in 5 min and to 95 cps in 60 min. Namely, under 3,600 sec of 5-kV irradiation, the intensity decreases more than 95% of the initial intensity. For Si/Eu-doped AlN, the intensity decreases 3,100 to 2,500 cps in 60 min, namely a decrease of 20% of this initial intensity. These results clearly show that the Si/Eu-doped AlN is much better candidate than Ce-doped La₅Si₃O₁₂N is due to its higher stability.

Figure 1: Luminescence of rare-earth doped SiAlON phosphors. Pictures of different phosphors under visible (a) and ultraviolet (b) light. (c) Normalized CL spectra of Eu²⁺ in different host lattices. Please click here to view a larger version of this figure.

Figure 2: Setup of CL. (a) Photograph of the SEM with CL system, with inset a photograph of the ellipsoidal mirror. (b) Schematic image of light detection system. Please click here to view a larger version of this figure.

Figure 3: Preparation of SiAlON phosphors. (a) Determination of the starting powders, weight and sintering conditions; (b) Mixing of the raw powders; (c) Sintering of the powder mixture; (d) Sintered powders before and after crushing. Please click here to view a larger version of this figure.

Figure 4: Cross-section preparation. (a) Mixing of the powders with resin and hardener, and evaporating air in the mixture. (b) Pouring into a silicon mold and heating. (c) Polishing of the chips by handy-lap and Ar-ion cross-section polisher. (d) Measuring cross-sectional polished area. Please click here to view a larger version of this figure.

Figure 5: Core-shell distribution in Si-doped AlN. (a) CL spectra of AlN powders doped with 1.6% of Si at 5 kV. (b–c) CL images taken at 5 kV and 280 nm for as-sintered (b) and cross-sectioned (c) AlN powders doped with 1.6% of Si, respectively. Please click here to view a larger version of this figure.

Figure 6: Local analysis of Ca-doped JEM phosphor. CS-SE, combined CS-CL images at 300 nm (red), 430 nm (blue) and 540 nm (green) and local CL spectra taken at 5 kV for JEM phosphors doped with 0 (a, b, c) and 0.69 (d, e, f) at. % of Ca, respectively. Please click here to view a larger version of this figure.

Figure 7: CL and EDS comparison of Si-doped AlN. CS-CL and CS-EDS images (a, c) and local spectra (b, d) of AlN particle doped with 4.0% Si doping. Please click here to view a larger version of this figure.

Figure 8: Luminescence stability of two blue phosphors. (a) CL spectra of Ce-doped La₅Si₃O₁₂N and Si/Eu-doped AlN after 20 sec of irradiation at 5 kV. (b) Evolutions of CL intensity of Ce-doped La₅Si₃O₁₂N and Si/Eu-doped AlN during electron-beam irradiation at 5 kV. Please click here to view a larger version of this figure.