1. Introduction

Photostimulated persistent luminescence refers to persistent luminescence under the stimulation of a low energy photon after the excitation with high-energy radiation such as X-rays or ultraviolet (UV) radiation [1]. This phenomenon is attributed to the recombination of trapped electrons stimulated by low energy light with holes or other centers. Using this kind of phenomenon, optical information storage can be carried on, broadening the application field of long-lasting phosphors from weak light illumination and safety signage to optical information and high energy beam survey. In conventional photostimulable storage phosphors, the stored optical information is usually read out as a transient luminescence which vanishes after the excitation stops. Research upon conventional storage phosphors has mostly focused on the superior performance of the BaFX:Eu²⁺ (X = Cl, Br, I) type of photostimulated phosphors, with applications in medical, computed and screen radiography [2,3]. Unlike transient luminescence, photostimulated persistent luminescent phosphors exhibit a long lasting luminescence, making it possible to replace traditional photostimulable storage phosphors in novel storage devices, high energy beam survey as well as image and optical information storage. In recent years, research interest into new photostimulable persistent phosphors has grown.

The persistent luminescence (i.e., including thermally- and photo- stimulated luminescence of materials exhibiting delayed emission of radiation after the removal of an excitation source, usually in the visible or near infrared (NIR) regions, and longer than the luminescence lifetime of the emissive state) of Eu²⁺-doped materials has been reviewed by Van den Eeckhout et al. [4].

Liu et al. [5] reported the photostimulated NIR persistent luminescence from LiGa₅O₈:Cr³⁺ and opened research to the study of non-halogen systems. Subsequently, much effort has been directed to the development of Cr³⁺ systems for bio-imaging in the first biological window [6–8].

Strontium aluminate, SrAl₂O₄, doped with lanthanide ions such as Eu²⁺ and Dy³⁺ exhibits long persistent luminescence and these phosphors are widely used in safety signage, displays, decorations and night-vision surveillance [9,10]. Since Matsuzawa et al. [10] reported on the extremely long-persistent luminescence of Eu²⁺ and Dy³⁺ co-doped SrAl₂O₄ in 1996, these phosphors have become a focus of research. Various synthetic methods other than solid-state reaction have been adopted in the synthesis of these phosphors [11–14]. Subsequent research has mainly concentrated upon the crystal structure [15,16], extension of the long afterglow decay time and the persistent luminescence mechanisms of SED phosphors [17–20]. Interest has also been shown in the mechanoluminescence properties of these phosphors [21–23]. However, the property of photostimulated persistent luminescence of SrAl₂O₄:Eu²⁺,Dy³⁺ which has potential applications in optical storage, has not been fully investigated. Jia et al. [24] found that the lifetime of photostimulated luminescence of this system at 80 K was three times faster than that of the normal Eu²⁺ luminescence and attributed this to lattice relaxation. It is pertinent to mention the observation of a Reviewer that the photostimulated upconversion emission reported for SrAl₂O₄:Eu²⁺,Dy³⁺ [25] is more correctly described as photostimulated luminescence.

In this work, the persistent luminescence and NIR photostimulated luminescence of Eu²⁺,Dy³⁺-codoped SrAl₂O₄ have been investigated in order to gain further insight concerning the nature of the trap states. The application of this phosphor in information storage and retrieval has been investigated.

2. Experimental

2.1 Synthesis

The strontium aluminate phosphors with bright green light emission and long persistent luminescence were prepared by a solid state reaction method under reducing conditions. The raw materials SrCO₃ (AR) and Al₂O₃ (A.R: Tianjin Damou and Kemiou Chemical Reagent Companies), Eu₂O₃ (99.95%) and Dy₂O₃ (99.99%) (Shanghai Aladdin Chemical Company) and small quantities (1%) of H₃BO₃ (A.R) were mixed intimately in stoichiometric amounts with alcohol in a mortar for an hour and then annealed in a high-temperature tubular resistance furnace with 10% H₂ and 90% N₂ as a reducing atmosphere. The mixture was maintained at 1250 °C for 4 h to obtain the product powders. Pellets (0.5g, diameter 15 mm) were formed from the powder mixtures with a uniaxial loading of 30 MPa in a stainless steel mold. In this report we focus upon the sample with the composition Sr_0.97Al₂O₄:0.01Eu²⁺,0.02Dy³⁺, subsequently abbreviated to SED throughout.

2.2. Characterization of samples

Powder diffraction patterns of SED phosphors were recorded by a MSAL-XD-2 X-ray diffractometer using Cu K_α radiation, λ = 1.5418 Å, at 36 kV tube voltage and 20 mA tube current. A Hitachi F-7000 Fluorescence Spectrophotometer equipped with a 150 W xenon lamp as excitation source was used to measure the optical properties of the phosphors Thermoluminescence analysis of bleached samples was carried out by a Pyroelectric Spectrograph Thermoanalyzer from room temperature to 430 °C at a heating rate of 1 °C s⁻¹.

3. Results and discussion

3.1 X-ray diffraction patterns

The composition and phase purity of samples were examined by X-ray powder diffraction patterns. The powder pattern for SED is shown in Fig. 1 and is compared with the standard file. The weak peak at 21.0° is due to Sr₄Al₁₄O₂₅ impurity. At room temperature, SrAl₂O₄ crystallizes in the monoclinic space group P2₁ (No. 4: Z = 4), and Shi et al. [26] have provided evidence that Eu²⁺ ions substitute at the two 9-coordinate Sr²⁺ sites, with preference of the tripositive lanthanide ions for the slightly smaller site.

 

Fig. 1 XRD pattern of SED phosphor synthesized by solid state method.

Download Full Size | PDF

3.2 Photoluminescence emission, excitation spectrum and persistent luminescence spectrum

The room temperature photoluminescence excitation (a), emission (b) and persistent luminescence (c) spectra of the SED phosphor were measured and are depicted in Fig. 2. The green emission in both photoluminescence and persistent luminescence is due to the 4f⁶5d¹ → ⁸S_7/2 4f⁷ transition of Eu²⁺ situated at one of the two Sr²⁺ sites [27]. The assignment of this emission to a specific site has received much attention, as summarized by Botterman et al. [27]. The low temperature emission spectra of these authors exhibit a broad feature with peak maximum at 521 nm and another weaker, sharper band with maximum intensity at 445 nm. Our 80 K emission spectra (not shown) are in agreement with this. On the basis of an empirical theory, first equating the configuration of Eu²⁺ to 5d¹, then coupling each 5d¹ Kramers level with 7 (instead of 3003) 4f levels, these authors associated the lower energy emission to that of Eu²⁺ situated at the Sr(1) site, which has larger average bond distances and a slightly larger coordination polyhedron volume. Subsequently, following the analysis of the spectra of Ce³⁺ diluted into the SrAl₂O₄ host from first principles calculations, a contrary conclusion was established: the green emission originates from Eu²⁺ situated at the smaller Sr(2) site [26].

 

Fig. 2 Photoluminescence excitation (a) and emission (b) spectra and persistent luminescence spectrum (c) of SED phosphor. The peak in (a) at 396 nm is due to Eu³⁺. Persistent luminescence was obtained using 5 min delay after excitation by 360 nm radiation.

Download Full Size | PDF

3.3 Persistent luminescence decay curve

In the dark, the green persistent luminescence (upper inset in Fig. 3) can be seen by the naked eye for one hour after stoppage of UV excitation. The decay curve in Fig. 3 depicts the persistent luminescence decay of the SED disc monitored at 510 nm following 20 min excitation at 356 nm. The tail of the curve after 4000 s indicates the background level.

 

Fig. 3 (a) Persistent luminescence decay curve of SED phosphor monitoring 510 nm emission after 356 nm irradiation for 20 min. Note the logarithmic scale. The top inset shows the persistent luminescence spectrum after 10 min delay. (b) Plot of the fitted lifetimes from biexponential decay against time of persistent luminescence. The fitting ranges are horizontally underlined together with the adjusted coefficients of determination. The fraction of the total emission intensity for the components τ₁ and τ₂ is indicated at the side of each data point.

Download Full Size | PDF

The decay curve is not well-fitted in the region from 0 to 1750 s by the equation I(t) = atⁿ, with n = −1.0 (R²_adj = 0.9845). However, the fitting improves for longer times (1020-3900 s) with n = −1.7 (R²_adj = 0.9984). Kamiyanagi et al. [28] have stated that this type of fitting indicates that the afterglow originates in the tunneling recombination within defect pairs of trapped holes and trapped electrons.

The decay can be fitted in the range from 0 to 4000 s by a biexponential function with lifetimes of 9.9 ± 0.06 s and 74.7 ± 0.6 s (R²_adj = 0.9972). As pointed out by Korthout et al. [17], the origin of the two components is not known but presumably corresponds to different types of trap center, different levels of the same trap species and/or retrapping.

From 0 to 400 s after stopping excitation, the decay curve in Fig. 3(a) is poorly fitted by a monoexponential decay function (R²_adj = 0.9567) and more accurately fitted by a bi-exponential decay function (R²_adj = 0.9972) with the two lifetimes of 9.91 ± 0.06 s and 74.9 ± 0.6 s. From 400 s to 2000 s delay after excitation, the monoexponential fit is much poorer and the biexponential fit (R²_adj = 1.0000) gives two different lifetimes of 148.5 ± 0.5 s and 497.5 ± 1.6 s. The lifetimes therefore increase with time but the ratio of the normalized amplitudes of the lifetimes changes. In fact, when fitting the persistent luminescence decay from 2000 s to 3950 s delay after excitation with biexponential decay (R²_adj = 0.9992), the lifetimes become even longer: 217 ± 44 s and 988 ± 26 s. Biexponential fittings of the lifetimes τ₁ and τ₂ of the persistent luminescence decay over various time periods, as indicated by horizontal bars with the corresponding R²_adj values above, are displayed graphically in Fig. 3(b). The fraction of the total emission intensity for the components τ₁ and τ₂ (as defined in [27]) is indicated at the side of each data point. The fractional amplitude of the shorter lifetime increases with time, whereas that of the longer lifetime decreases. The plots of τ₁ and τ₂ against time are each fitted by exponential growth functions as shown by solid lines:

3.4. Persistent luminescence excitation spectrum

To evaluate the effectiveness of different excitation wavelengths for persistent luminescence, the excitation wavelength-persistent luminescence intensity of the SED disc was investigated to discern the most efficient excitation wavelength. Figure 4 displays the persistent luminescence intensity at 510 nm for different exciting wavelengths, after ceasing illumination for 10 s, normalized by the intensity of the exciting xenon lamp radiation. First, the persistent luminescence was not detected at the range of excitation wavelengths longer than 479 nm, which corresponds to the absence of absorption by Eu²⁺. The maximum intensity of persistent luminescence is observed under excitation at 335-365 nm, corresponding to the maximum absorption of Eu²⁺ at both sites. From the low temperature excitation spectrum (Fig. 3 in [27]), the excitation wavelength range from 432 nm to 470 nm (Fig. 4) is associated only with absorption by only the Sr(2) site. Aitasalo et al. [29] have commented that the excitation processes for the photoluminescence and persistent luminescence of SrAl₂O₄:Eu²⁺ are significantly different but the result in Fig. 4 shows distinct similarity between the two processes in the wavelength region that we have investigated.

 

Fig. 4 Normalized 510 nm persistent luminescence intensity at 10s after excitation (I_10s) plotted against the excitation wavelength. Some relevant wavelengths are marked. All data were collected from the bleached SED sample.

Download Full Size | PDF

3.5 Thermoluminescence

Figure 5 depicts the thermoluminescence spectrum over the temperature region from 40 °C to 430 °C with different waiting times of 10 s, 10 min and 40 hours respectively after pre-irradiation by ultraviolet radiation for 2 min. The three curves all contain two broad bands at ~130 °C and 380 °C. However, the lower temperature peak comprises several unresolved components. The shift of the peaks with varying delay time is small. When the waiting time increased, the intensity of the peak at ~130 °C decreased whereas that of the peak at 380 °C did not change. This indicates the preferential release of electrons in shallow traps.

 

Fig. 5 Thermoluminescence spectra of the SED phosphor recorded at different delay times after pre-irradiation by ultraviolet radiation at 356 nm for 2 min. The peak temperatures are indicated.

Download Full Size | PDF

Aitasalo et al. [29] reported two peaks within the temperature range from 40 to 240 °C: located at 80 °C and 160 °C in the thermoluminescence spectrum of SrAl₂O₄:Eu²⁺ after pre-irradiation and delay of 3 min. The initial rise method showed that the lower temperature peak corresponded to at least three traps, with depths of 0.55-0.65 eV. The relative intensities of the peaks in Fig. 5 differ considerably from those in Fig. 4 of [29] – notably that the 80 °C peak is much weaker herein and the ~130 °C peak appears. By contrast, Chang et al. [13] observed a broad band with maximum at 103 °C in the thermoluminescence spectrum of SED:xEu²⁺,yDy³⁺. Takasaki et al. [20] calculated trap depths of 0.3 eV and 0.6 eV. The results differ not only because of differences in synthesis but also instrumental parameters. However it is clear that trap depths of 0.3-0.5 eV occur.

3.6 NIR photostimulated luminescence

The inset of Fig. 6 shows the excitation spectrum of photostimulated persistent luminescence when monitoring the emission at 510 nm and stimulating the SED disc with near infrared (NIR) radiation after 30 mins of ceasing pre-irradiation of the sample with 420 nm UV excitation for 5 min. The maximum emission intensity occurs for the wavelength range between 740 and 780 nm. This photostimulation corresponds to the elevation of an electron from a filled trap level to the conduction band, from whence recombination occurs at a Eu³⁺ ion, leading to 4f⁶5d¹ → 4f⁷ emission of Eu²⁺. From the figure, the smallest trap depth in the measurement range is then less than 1.4 eV and the traps form a continuum, with maximum density at 1.6 eV. These experimental results agree with the report of Ji et al. [25]. However, energy labeled E₂ at 2.6 eV in [25], assigned to the removal of an electron from the valence band to the ground state of Dy³⁺, in fact corresponds to the commencement of the 4f⁷ → 4f⁶5d absorption spectrum of Eu²⁺.

 

Fig. 6 The NIR photo-stimulated persistent luminescence decay curve of the SED disc. The decay curve was recorded at 1 min following the 420 nm excitation for 1 min, and then stimulation by 760 nm NIR radiation was employed at “On” and switched off at “Off”. The inset shows the excitation spectrum when monitoring 510 nm emission using NIR photostimulation at 10 min after the stoppage of UV excitation for 1 min.

Download Full Size | PDF

Although it is contended that Dy³⁺ is not the trap [17], the current model of persistent luminescence involves the ionization of the Eu²⁺ 5d electron to conduction band states and the subsequent trapping by Dy³⁺ [30,31], The model is partly based upon the vacuum referred binding energy (VRBE) diagram with the electronic ground state of Dy²⁺ situated slightly below the bottom of the conduction band. Firstly, the band gap of SrAl₂O₄ (or the co-doped system) has been determined to be 4.98 eV [25], 6.5 eV [32], 6.6 eV [19,33] and 7.28 eV [26] thus making the trap location below the conduction band uncertain. Secondly, the charge transfer energy of Eu³⁺ in SrAl₂O₄ has been given as 4.96 eV [34], 4.68 eV [35] and as 4.77 eV in CaAl₂O₄:Eu³⁺,Li⁺ [36]. The electronic ground state of Dy²⁺ is then located at 2.2-2.4 eV higher than this value above the valence band maximum. Hence there are some uncertainties in the construction of this diagram. Furthermore, the location of the Dy²⁺ ground state is affected by the environment of this ion – for example being lowered with respect to the conduction band when adjacent to an oxygen ion vacancy. If we accept the VRBE explanation [30,31] then the pseudo-continuous trap distribution highlighted in the present study must be accounted for. The 4f¹⁰ Dy²⁺ ion has a similar electronic structure to Ho³⁺, although the larger size of Dy²⁺ causes the inter-electronic repulsion to be less so that the terms are closer in energy. In fact, Dy²⁺ has many (>30) energy levels extending from the electronic ground state up to ~0.7 eV and the de-trapping rates from these (intermediate) levels may differ, thereby accounting to some extent for the change in decay lifetimes with time of persistent luminescence decay. The ⁵I₈ levels will be in thermal equilibrium. The energy separation of Dy^{2+ 5}I₈-⁵I₇ trap states is estimated to be ~0.6 eV by comparison with 4f¹⁰ Ho³⁺. Moreover, the Dy²⁺ ground state energy at a specific site is subject to the particular environment, which is expected to vary due to the absence of charge compensation and the presence of random charges. First principles density functional calculations of the electron trap(s) would be helpful in identification, as demonstrated in [37].

The main panel of Fig. 6 shows the effect of photostimulation by 760 nm radiation upon the persistent luminescence decay of the SED disc. When the radiation is switched on (“On”) the persistent luminescence intensity increases rapidly and then reaches saturation. As noted by Ji et al. [25], the function of the NIR radiation is to empty filled traps and not to fill new traps. The rise portions of Fig. 6 can be fitted by biexponential functions with lifetimes of 29.3 ± 0.5 s and 4.7 ± 0.4 s. The rise curves in Fig. 6 might be explained by a partial (re)trapping in shallower traps. For the “On” phase, the deep trap release by the NIR radiation does not yield immediate recombination, but partial retrapping in shallower traps, leading to a certain extent a redistribution in trap depth occupation. For the “Off” phase, the persistent luminescence intensity increases compared with a normal persistent luminescence decay curve due to retrapping in shallower traps after NIR photostimulation.

When turning off the NIR excitation source, the green luminescence decays slowly, exhibiting a long-lasting luminescence. After many times of turning on and off NIR photostimulation, the sample still shows good photostimulated persistent luminescence. This means that the SED sample can repeatedly read information using NIR light after one period of UV excitation, indicating its potential application in optical information storage.

3.7 The write-in and read-out properties of the SED phosphor

The SED phosphor has a potential application in optical storage. The ranges of write-in and read-out wavelengths for this phosphor are summarized in Figs. 7(a),(b), respectively. Figures 7(a) and (b) are similar to Fig. 4 and the inset of Fig. 6, respectively, but the experiments utilized to construct these figures were different. The method of construction of the Figs. 7(a),(b) was as follows. For the write-in Fig. 7(a), the sample was irradiated for 2 min by different monochromatic ultraviolet wavelengths between 240 nm to 490 nm. Then the pre-irradiation was stopped for 30 s. Immediately, radiation with the NIR wavelength of 760 nm was employed to stimulate the sample for 70 s. The emission intensity at 510 nm after 10 s of ceasing NIR stimulation was then plotted against the ultraviolet irradiation wavelength. The NIR photostimulation raises electrons from deep traps populated by pre-irradiation. Wavelengths beyond 430 nm up to 490 nm are not efficient in populating these traps and the most efficient write-in wavelengths are between 330 and 360 nm. Figure 7(c) shows the persistent luminescence decay at 510 nm as above when using the different write-in wavelengths of 290 nm, 350 nm and 450 nm and ceasing the pre-irradiation, prior to photostimulation. Fitting these decays by biexponential functions over the same time duration shows that the short decay lifetimes are the same within experimental error, whilst the long decay lifetime is rather longer for 450 nm charging. Figure 7(b) shows the read-outs (also quantified by emission intensity at 510 nm) following charging by 430 nm radiation for 2 min, decay for 30 s, then stimulation by NIR radiation between 600 and 900 nm (separately, using 10 nm steps) for a further 70 s, and taking the intensity reading after the NIR stimulation stopped for 10 s. The phosphor has good read-out performance when using wavelengths between 630 and 850 nm, with the optimum value at 760 nm. Figure 7(d) depicts the NIR photostimulation part of the read-out experiment, with duration of 70 s. Based upon the NIR photostimulated persistent luminescence phenomena, the write-in and read-out capabilities of the phosphor can be applied in optical information storage.

 

Fig. 7 The write-in (a) and read-out (b) wavelength ranges for SED with normalization of the NIR photostimulation intensity. The decay following various write-in wavelengths (c) and the rise for NIR photostimulation at selected wavelengths after charging at 420 nm (d). The emission at 510 nm was monitored in all cases. Refer to the text for a detailed explanation of these experiments.

Download Full Size | PDF

4. Conclusions

Numerous reports have been previously made concerning the persistent luminescence phosphor SrAl₂O₄:Eu²⁺,Dy³⁺. In the present work the phosphor SED has been synthesized by a high temperature solid state reaction and the persistent- and photo- luminescence and excitation spectra, thermoluminescence and photostimulated properties have been investigated. The major new findings are that the persistent luminescence decay curve, as fitted by a biexponential function, exhibits a change in lifetimes with time of persistent luminescence. The fraction of the total emission intensity increases for the shorter lifetime component with increasing time. The change in lifetime has been associated with the emptying of a pseudo-continuum of trap states, either from different traps or from different levels of the same trap species. Secondly, photostimulation of the phosphor can take place over a wide range of NIR wavelengths, which is consistent with this type of distribution of traps over a wide energy range. However, as argued above, these facts do not exclude the postulate that Dy³⁺ is the electron trap species. Finally, the use of the phosphor in information storage and retrieval has been emphasized.