1. Introduction

Luminescent properties of rare-earth doped oxides, carbides and/or sulfides, under vacuum ultraviolet (VUV) excitation, have attracted a considerable attention due the possibility of their application in mercury-free fluorescent lamps, photoluminescence liquid crystal display (PLLCD) [1], field emission displays (FED) [2] and plasma display panels (PDPs) [3–5]. For PDP, for example, many of the phosphors commercially available, as the red ones, have poor chromaticity and low efficiency under VUV excitation (λ< 200 nm), mainly in the range between 146 and 173 nm, which correspond to the emission range from the inert gas plasma used for building PDPs [6,7]. Therefore, there is a current demand for the development of new materials with high efficiency, as well as a better understanding of their luminescence mechanisms when they are excited in the VUV range. Doped lithium aluminates (LiAl₅O₈ - LAO) are among the candidates for this kind of application as cited on the literature [8–15]. For example, LAO doped with Eu³⁺ has demonstrated its potential to be applied as a phosphor for use in PLLCD due to its orange red/red emission [1] and LAO: Tb³⁺ was also cited as a bluish-white phosphor able to be applied in FED, cathode X-ray tube and fluorescent devices [2].

High luminescence efficiency is a property, which can be strongly influenced for many factors such as annealing temperatures, particles size and structural composition. Several studies show that the addition of Li⁺ as co-dopant in different materials can be used to enhance their luminescence emission [16–19]. Some authors reported that Li⁺ ions improved the crystallinity, changed the grain size and also contributed to the creation of oxygen vacancies [16,19]. In this work, we report the effects of Li⁺ excess and calcination temperature on the LiAl₅O₈:Eu luminescent properties, when exposed to VUV radiation. The application of this system a potential new red VUV phosphor for PDP is discussed.

2. Experimental

LAO phosphors were doped with 3 mol% Eu, stoichiometric prepared and also with 0.5 mol% of Li excess. The chemical method for getting the samples was the sol-gel based on polyvinyl alcohol (PVA) [20,21] and the starting reactants were the metal nitrates LiNO₃ (99.99%), Al(NO₃)₃ (99.9%) and Eu(NO₃)₃(99.99%), all from Sigma-Aldrich and a PVA solution (0.1 g/ml). The raw materials, in stoichiometric amounts, were dissolved in distilled water and then 30% of PVA was added based on the final volume of the starting solution. The system was kept under stirring at constant temperature of 100 °C during 4 h. Finally, the obtained dried gels, or xerogels, were calcined in static air at 900 and 1000 °C for 2 h.

The crystal structure and phase purity of the LiAl₅O₈:Eu materials were routinely verified with the X-ray powder diffraction (XRD) measurements using a Rigaku RINT 2000/PC diffractometer with Co-Kα radiation (1,79 Å), in the Bragg-Brentano geometry. The VUV excitation of integrated luminescence and the photoluminescence (PL) emission spectra were recorded at room temperature and in ultra-high vacuum (better than 10⁻⁷ mTorr) conditions at the Toroidal Grating Monochromator (TGM) beamline (E/ΔE ~500 to 700) of the Brazilian Synchrotron Light Laboratory (LNLS) [22, 23], which was recently upgraded and now covers from 3 to 330 eV. Monochromatic photons from 6 – 10 eV (~206 to 124 nm), filtered using a MgF₂ (thickness: 200 μm) window, were used for exciting the samples and the emitted light was collected by an optical fiber (aperture: 600 μm) coupled to the chamber by a vacuum feedthrough and connected to a R928 Hamamatsu photomultiplier (PMT). In front of the PMT, a glass slide (thickness: 170 μm) was used for cutting the scattered light (λ<300 nm), in a setup prepared to collect the total photoluminescence yield (TPY) of the sample. The photoluminescence emission spectra were recorded for specific energies (6.5 and 8.5 eV – 191 and 146 nm, respectively) and the setup was composed by an optical fiber connected to an Ocean Optics QE65000 spectrometer. The spectra were corrected for the variation in the incident flux of the excitation beam using the excitation and emission spectra of sodium salicylate (C₇H₅NaO₃) as standard [24, 25]. The UV excitation spectra were recorded using a spectrofluorometer Fluorolog 3 from Horiba at the Advanced Optical Spectroscopy Multiuser Laboratory from the Institute of Chemistry from the University of Campinas (FAPESP/LMEOA/IQ/UNICAMP). The total X-ray Excited Optical Luminescence (XEOL) yield was registered, around the Al K edge and the XEOL emission spectra were recorded at the Soft X-ray Spectroscopy (SXS) beamline at the LNLS using a similar setup to the TPY measurements, but exciting the samples with X-rays (1566 eV, above the Al K edge).

3. Results and discussion

The XRD patterns (Fig. 1) exhibit only the reflection peaks of the spinel structure of the LiAl₅O₈ with the P4₃32 (or P4₁32) space group [26]. No clear effect in the crystalline structure or secondary phase was observed with the Li⁺ excess, Eu³⁺ doping or different calcination temperatures. In LiAl₅O₈, there are one tetrahedral (8c) and two (distorted) octahedral (4b and 12d) sites. It is known that Al³⁺ occupies both tetrahedral and octahedral sites (preferentially 12d), while Li⁺ occupies only the distorted octahedral ones (preferentially 4b) [26]. The Eu³⁺ ions cannot occupy the tetrahedral site, but can substitute either the Al³⁺ or Li⁺ octahedral sites. However, the smaller ionic radius of Al³⁺ (0.535 Å) compared to Li⁺ (0.76 Å) and Eu³⁺ (0.94 Å), all with coordination number VI [27], suggests that Eu³⁺ will more probably substitute Li⁺, requiring a charge compensation. Furthermore, the average Al-O distance in the 12d site is 1.90 Å while the Li-O in 4b is 2.05, closer to the average Eu-O distance in the Eu₂O₃ sesquioxide, ranging from 2.2 to 2.7 Å [28].

 

Fig. 1 XRD patterns of the Li_1+xAl₅O₈:Eu (x = 0 and 0.05) materials compared to the standard structure proposed by Famery et al., (1979) [26].

Download Full Size | PDF

The VUV excitation spectra of the emission of LiAl₅O₈:Eu materials (Fig. 2) exhibit a broad band at ca. 191 nm and a sharp edge at ca.150 nm. The edge close to 150 nm (8.3 eV) is related to transitions from the top of the valence band (VB) to the bottom of the conduction band (CB), i.e. the host band gap or the formation of free electron-hole pair. The band gap energy (E_g) can be estimated using the minima of the first derivative of the excitation spectra, which is at 150 nm. The differences in the E_g as a function of synthesis temperature or the excess of Li⁺ (ranging from 150.0 to 150.6 nm) are smaller than the resolution of the TGM beamline (+/− 2 nm at the measurements energy range).

 

Fig. 2 Excitation spectra for LiAl₅O₈:Eu total luminescence yield scanned at the VUV range.

Download Full Size | PDF

The second band, with maxima at 190 and 205 nm, is associated to the O(2p)→Eu³⁺ charge transfer excitation of Eu³⁺ ions. This band starts at 320 (3.9 eV) in the UV region with maximum at 255 nm (Fig. 3) and continue to the VUV region. The lowest energy of this band is compatible to the Eu-O CT band in other Eu³⁺ doped oxide and aluminate hosts, which starts normally at 4-6 eV [29]. Since the VUV excitation spectra (Fig. 2) takes into account the integrated emission, it is also probable that the Eu²⁺ 4f⁷→4f⁶5d¹ excitation bands are overlapped in this region. On the other hand, it is not probable that this band arises from Eu³⁺ 4f⁶→4f⁵5d¹ excitation since the energy of these transitions should occur only at higher energies than 8.5 eV (<140 nm) [5,30], which were avoided with the use of the MgF₂ filter.

 

Fig. 3 PL excitation spectra of LiAl₅O₈:Eu, calcined at 1000°C, monitoring the Eu²⁺ (441 nm) (green curve) and Eu³⁺ (613 nm) (red curve) emission wavelength.

Download Full Size | PDF

In order to investigate the relationships between the lithium excess and calcination temperature on the luminescent properties of LiAl₅O₈:Eu under VUV excitation, the emission spectra of LiAl₅O₈:Eu and Li_1.05Al₅O₈: Eu, calcined at 900 and 1000°C, were measured under excitation below (191 nm) and above (146 nm) the band gap, as shown in Fig. 4.

 

Fig. 4 PL emission spectra of LiAl₅O₈:Eu and Li_1.05Al₅O₈:Eu samples excited at 191 nm (left) and 146 nm (right). All spectra were normalized to the Eu^{3+ 5}D₀ → ⁷F₂ transition intensity.

Download Full Size | PDF

The emission spectra under excitation at the Eu³⁺ CT band (191 nm) of LiAl₅O₈:Eu and Li_1.05Al₅O₈:Eu samples exhibit two broad bands centered at 300 and 480 nm and peaks at 612 and 700 nm (Fig. 4, left). All the spectra were normalized in intensity to the ⁵D₀→⁷F₂ transition of Eu³⁺ at 612 nm. The band at 300 nm probably arises from defects’ emission. Several defects can be present in this host as intrinsic Schottky ones generated via evaporation of the volatile Li₂O [31] or those created by the charge compensation needed when doping a trivalent ion (Eu³⁺) in a monovalent site (Li⁺).

The peaks at the range of 585 to 750 nm are due to Eu³⁺ emission from the excited state ⁵D₀ to the ⁷F_J (J: 0-4). Finally, the 480 nm band arises probably from the allowed 4f⁶5d¹→4f⁷ emission of Eu²⁺ ions, formed due to the high temperature synthesis. Eu²⁺ can be stable at very high Fermi energy materials, which seems to be the case of LAO. The Fermi energy is normally half of the gap energy [32], in this case ca. 4.1 eV. The energy of Eu²⁺ ground state can be estimated using the energy of the Ligand to Metal Charge Transfer (LMCT) O(2p)→Eu³⁺ transition [33] which starts at 3.9 eV in this host. Eu²⁺ is, thus, stable in this host since its ground state is below the Fermi energy. Albeit the Eu²⁺ concentration being very small, its allowed transition facilitates its identification by the analysis of the emission spectra, even if not detectable by X-ray absorption spectroscopy. Even if the emission probability of Eu²⁺ (Laporte allowed) is much higher than Eu³⁺ one (Laporte forbidden), under very low Eu²⁺ concentrations a comparison between the emission ratio of Eu²⁺/Eu³⁺ in different spectra can qualitatively indicate an increase/decrease in the reduction process. Thus, the higher Eu²⁺/Eu³⁺ emission ratio increase under higher calcination temperatures indicate that the reduction is a temperature-dependent reaction.

The emission spectra measured with excitation above the band gap at 146 nm (Fig. 4, right) exhibit the same bands of the spectra of LMCT excited, with an extra peak at 710 nm. This peak is related to Cr³⁺ ruby like ²E→⁴A₂ emission. Most of Al reactants contain Cr³⁺ as an impurity and this peak is similar to the one observed by Singh et. al. for the Cr³⁺ doped lithium aluminate¹¹. This result show that there is no energy transfer from Eu³⁺ to Cr³⁺, since, when excited in the Eu³⁺ LMCT energy (Fig. 4, left), no Cr³⁺ emission is observed.

Differently from the CT excited, the emission spectra excited above band gap energy (Fig. 4, right) exhibit more intense Eu²⁺ bands, allowing a better understanding of the reduction mechanism. Most of the red emitting PDP materials found in literature have an emission at only the red region [34], thus we show a new possibility of PDP materials that may be tuned from red to blue emission depending on excitation wavelength and synthesis parameters.

One can observe that the ratio of Eu²⁺/Eu³⁺ emissions increases with temperature, and decreases with excess of Li⁺. As already pointed out, Li₂O is a volatile oxide, so the effect of temperature might be related to the lack of Li⁺ in the system. Besides, the changes in the optical properties of the solids with different Li concentration cannot be related to Li⁺ itself since Li⁺ (1s²) is optically inactive and thus the only optical properties changes caused by its excess are related to the changes on Eu^2+/3+ stabilization.

Both excitations (190 and 147 nm) promote Eu²⁺ as well as Eu³⁺ to excited states, allowing both Laporte allowed and forbidden radiative emissions, respectively. The 190 excitation allows Eu²⁺ emission because in this region, 4f⁷→4f⁶5d¹ transition is overlapped to the Eu³⁺ LMCT band, as it can be observed in conventional excitation spectroscopy in the UV region (Fig. 3). On the other hand, the reason that 147 nm excitation allows both emissions is due to the formation of electron-hole pair followed by an energy transfer for both ions. To confirm this mechanism for excitation at high energy, the total XEOL yield was registered (Fig. 5, left) showing that the luminescence increases as a function of excitation energy and around the Al K edge it presents a positive edge. It suggests there is very efficient energy transference originated from the recombination of electron-holes pairs that excites the material’s optical channels [35]. The XEOL emission spectra (Fig. 5, right) exhibited the same profile as the band gap excited spectra, presenting defects and Eu²⁺ bands, as well as the Eu³⁺ and Cr³⁺ peaks. This result confirms that once the electrons are excited to very high energy states, the energy is transferred to all emitting centers.

 

Fig. 5 Total XEOL yield and excitation spectra (left) and XEOL emission spectra (right) of LiAl₅O₈:Eu and Li_1.05Al₅O₈:Eu samples.

Download Full Size | PDF

A model that explains how the two valences of Eu ions are actually accommodated inside the LiAl₅O₈ structure is based on the charge compensation defects, which are created in the aliovalent substitution of Li⁺ by Eu³⁺. The trivalent Eu³⁺ ions doped into LiAl₅O₁₂ will replace the Li⁺ ions due to size similarity, forming the positive defect ${\text{Eu}}_{\text{Li}}^{\text{••}}$ and interstitial oxide ions ${\text{O}}_{\text{i}}^{\text{'}\text{'}}$(Eq. (1)). It is well known that interstitial oxide ions can be oxidized to O₂(g) at high temperatures reducing other species like Cu²⁺ in superconductors [36,37] or Eu³⁺ in persistent luminescence materials [24] (Eq. (2)). This oxidation will be more spontaneous at higher temperatures, which explains the higher amount of Eu²⁺ in the 1000 °C heated samples.

The electrons released in the interstitial oxide oxidation reduce the ${\text{Eu}}_{\text{Li}}^{\text{••}}$ species to ${\text{Eu}}_{\text{Li}}^{\text{•}}$ (Eq. (3)). Since Eu²⁺ has a larger ionic radius (1.17Å) compared to Eu³⁺ (0.95Å) and Li⁺ (0.76Å), it will be better accommodated if there is more space for distortions, like in the presence of Li vacancies. It means that the samples heated at 1000 °C will accommodate better the Eu²⁺ ions due to the evaporation of Li₂O be more effective at higher temperatures. The lack of Li will also be higher in the stoichiometric samples corroborating with the results from emission spectroscopy. The presence of defects (e.g. ${\text{V}}_{\text{Li}}^{\text{'}}$) was demonstrated by the use of synchrotron radiation, both in the VUV and X-ray around de Al K edge for exciting the samples, which resulted in the presence of luminescence in the UV region at the emission spectra. Unfortunately, the exact origin of these defects is hard to know.

Based on our data, it is difficult to give a quantitative amount of Eu²⁺ and Eu³⁺ ions from the emission intensities of the sample because the two luminescence centers have their own characteristic excitation wavelengths. For example, Eu²⁺ has a Laporte-allowed, spin-forbidden transition with lifetime of the order of μs, while Eu³⁺ has a Laporte-forbidden transition with lifetime of the order of ms. This difference allows energy transfer from Eu²⁺ to Eu³⁺ that can increase Eu³⁺ emission instead of its expected decrease as it is reduced to Eu²⁺. The determination of their concentrations could be better accomplished with X-ray absorption or even paramagnetic susceptibility [38] measurements, that will be done in future works.

4. Conclusions

The influence of lithium excess and calcination temperature on the luminescence properties of the LiAl₅O₈:Eu under VUV excitation is related to the higher Eu³⁺→Eu²⁺ reduction in the stoichiometric and at higher calcination temperatures. This reduction occurs due to the necessary charge compensation when Eu³⁺ substitutes the monovalent Li⁺ ion. The emission of defects under VUV and X-ray excitation corroborates the presence of these charge compensation defects. The presence of divalent europium is proved with both VUV-excited emission and UV excitation spectra which are composed by Eu²⁺ and Eu³⁺ ions transitions. The VUV excitation measurements yielded the value of the host band gap (8.3 eV), indicating that the Fermi level of the system is above Eu²⁺ ground state. The tuning of luminescence of this material may be done by changing the amount of Li and calcination temperature, making this a promising system for application as green/red VUV phosphor.