Open Access
Add to BibSonomyAbstract
Persistent phosphors, also called glow-in-the-dark materials, are a specific class of luminescent materials having the unique ability to emit light long after the excitation ended. For many applications in the visible spectrum, such as in emergency signage or road marks, the storage capacity of the Eu2+ based phosphors should further be increased. In this work we show that the excitation of the europium center in Sr2MgSi2O7:Eu,Dy by near-UV light not only leads to charge trapping, but also to optically stim-ulated release of previously trapped charges and subsequent luminescence (OSL). The experimental evidence for OSL at the excitation wavelength is supported by a model assuming local trapping and an additional detrapping rate proportional to the excitation intensity. In this way, the characteristics of both the charging and afterglow behaviour can be explained. The storage capacity of a persistent phosphor is thus not only controlled by the trap density and the trap depth, but also by the sensitivity to optically stimulated detrapping at the excitation wavelength.
© 2016 Optical Society of America
1. Introduction
Persistent phosphors are luminescent materials which, after ending the excitation, emit light much longer than the lifetime of the excited state of the luminescent ion [1]. This particular class of phosphors has the ability to temporarily store energy in so-called traps. In contrast to storage phosphors, which are characterized by deep traps, the ambient temperature is the main driving force leading to the release of trapped charges (i.e. detrapping), on typical time scales of seconds up to hours [2]. The field of persistent luminescence became very active over the past two decades, following the discovery of the green-emitting SrAl2O4:Eu,Dy by Matsuzawa et al., showing a long and intense afterglow [3]. Since then, many new phosphor compositions have been described, covering a wide range of emission wavelengths [2,4].
Typical applications for persistent phosphors are found in safety and emergency signage. After a sufficiently long period of illumination, the persistent phosphor remains in an equilibrium state, where trapping and detrapping are balanced. When the illumination ceases, e.g. due to a power failure, the persistent phosphor immediately provides a high emission intensity in the initial decay. Although the intensity then quickly drops, this is partly compensated by the increasing dark adaptation of the human eye [5,6]. For these applications, the persistent phosphor needs to provide high output on a relatively short timescale of about one hour.
Recently, the feasibility of using SrAl2O4:Eu,Dy persistent phosphor for road marks was assessed [7]. During daytime, the phosphor is charged, while the trap depth distribution is such that the afterglow lasts all night, for a sufficiently wide range of ambient temperatures. It turned out that the concept is not yet mature, since the emission intensity at the end of the night becomes too low, especially when temperature drops significantly during the night. This can, at least partially, be alleviated by an increase in the trapping capacity of persistent phosphors [7].
In recent years, the range of emission was also extended towards the deep red or near-infrared with persistent luminescence originating from Mn2+, Nd3+ but especially Cr3+ based phosphors [8]. These materials have been shown to be useful for in vivo bioimaging [9, 10]. When re-excitation, after introduction into the body, is not straightforward or occurs with poor efficiency [9], a larger storage capacity is again desired.
Although recently strong progress was achieved in the understanding of Cr3+ based phosphors [11], several issues remain in the case of Eu2+ based persistent phosphors, such as the chemical nature of the traps [12], the involvement of the conduction band (i.e. charge delocalization) [13] and the role of the trap distribution and possible retrapping or tunnelling recombination [14]. This leads to a trial-and-error approach in optimizing phosphor composition and synthesis conditions.
To study the storage capacity of persistent phosphors, different approaches are used. The afterglow decay curve can be recorded after charging under specific conditions. This then leads to an afterglow duration which is ideally defined as the time to decay to 0.3 mcd/m2, in the case of emission in the visible part of the spectrum [5]. Often, two or more exponentials are derived from the decay curve, which however only provide a good fit in a limited time window due to the complexity of the afterglow decay curve. Alternatively, thermoluminescence (TL) measurements are used to assess the depth, and thus the usefulness, of certain traps. Again, fairly complicated TL glow curves are often found which leads to uncertainty in assigning trap depths and the order of the trapping.
As a standard afterglow measurement is essentially the same as thermoluminescence at fixed (room) temperature, the information extracted from afterglow decay and TL glow curves should in principle be comparable. This is however often not verified, or both data sets are not compatible. Ideally, the same set of parameters should be able to predict both afterglow and TL. This was demonstrated for the case of the trap distribution found in CaAl2O4:Eu,Nd [1,15].
Furthermore, information can be extracted from monitoring the phosphor's light emission during the charging process. In the case of SrAl2O4:Eu,Dy, J. Botterman et al. designed a specific experiment to reveal the charging and decharging dynamics of a persistent phosphor [16]. It appeared there was a strong influence of the excitation wavelength used for charging, leading to TL glow curves with different shapes, from which it was concluded that trapping presumably occurs locally, close to the luminescent ion from which the electron originates. Hagemann et al. confirmed the wavelength dependent charging and the thermal barrier for charge trapping for SrAl2O4:Eu,Dy for a variety of excitation wavelengths and excitation temperatures [17].
In this work we evaluate the possibility that previously trapped charges are released by the excitation light intended to induce the charge trapping. This process is related to optically stimulated luminescence (OSL), which is a well-known phenomenon in storage phosphors and can be used in several dating and dosimetry methods [18, 19]. Recently, OSL has been described in several persistent phosphors as well. Typically, (near-)infrared excitation is used (1 to 1.5 eV) which can indeed lead to charge detrapping, subsequent recombination and light emission. The net effect is the release of trapped charges. As an example, Zhuang et al. showed OSL in Zn(Ga1−xAlx)2O4:Cr,Bi red persistent phosphors, with near-infrared stimulation leading to a reduced intensity on the low temperature side of the TL glow curve [20].
The energy of the stimulating light is typically somewhat larger than the trap depth (0.6 – 0.9 eV) appropriate for afterglow around room temperature. In general, it is not explained what mechanism (such as the formation of color centers) is responsible for the absorption of these photons by the trapped charges [18].
For YPO4:Ce,Sm however Bos et al. showed that for charging and decharging a redox couple Ce3+,Sm3+ ↔ Ce4+,Sm2+ is formed. Based on the wavelength dependency of the OSL, the expected absorption bands at about 1.8 and 2.6 eV of divalent Sm were found after charging [21]. Also in this case, the OSL is studied at energies lower than the excitation energy for the recombination center. Hagemann et al. however raised the possibility of optical detrapping by short wavelength ultraviolet light to explain the loading behavior of SrAl2O4:Eu,Dy [17].
In addition, there are several indications that after trapping the absorption behaviour of certain (persistent) phosphors can change significantly, even leading to prominent photochromism. Ueda et al. demonstrated photochromism in the well known CaAl2O4:Eu,Nd persistent phosphor with a reversible change from white to purple body color upon charging. It was suggested that the electron traps responsible for the photochromism may not be identical to those reponsible for the persistent luminescence [22]. Y Jin et al. observed a reversible change (from white to light gray) in the body color of europium doped Zn2GeO4, with related TL glow peaks at 341 and 353 K [23].
In this work a straightforward mathematical model - including OSL based detrapping caused by the excitation light - is presented, which is able to simulate the charging and afterglow characteristics as function of temperature and excitation intensity. For this purpose, we have chosen the blue emitting Sr2MgSi2O7:Eu,Dy as the model system, rather than the green emitting SrAl2O4:Eu,Dy. The reason for this is twofold. Firstly, SrAl2O4:Eu,Dy has two different emission bands, which have been related to two different Eu sites [16]. Significant energy transfer occurs between both sites, which would complicate the local model developped further in this work and require the introduction of too many parameters. In Sr2MgSi2O7:Eu,Dy only a single strontium site is available for substitution by Eu2+ ions [25]. Secondly, to show the occurrence of OSL at the excitation wavelength, one should be able to decouple charging and OSL. For Sr2MgSi2O7:Eu,Dy the thermal barrier for trapping is more outspoken than for the strontium aluminate, which allows us to set up experiments to separate the two effects.
2. Experimental
In this work, thin, pressed pellets of Sr2MgSi2O7:Eu,Dy (GloTech Int.) were used, having an emission peaking at 470 nm [5]. At low temperature, the lifetime of the Eu2+ center upon pulsed excitation is 603 ns. The phosphor pellets were stable, such that many cooling and heating cycles did not lead to appreciable degradation of the emission intensity.
Charging of the pellets was performed by a fiber coupled LED with a peak wavelength of 370 nm. The LED was operated at constant current mode and it was verified that the emission intensity of the LED was stable during the excitation, apart from the first 60 s when the emission intensity slightly dropped by about 2% compared to its initial level. On the detection side, the spectrometer allowed to separate the emission of the phosphor from the reflected excitation light during the charging step. Charging, afterglow and thermoluminescence were consecutively measured in the same automated, home built setup, for which technical details can be found in [16].
A specific experiment to reveal the charging and decharging dynamics of a persistent phosphor consists out of three steps [16]; (i) monitoring the luminescent emission during excitation at a certain temperature; the charging step, (ii) measurement of the luminescent emission after switching off the excitation at the same temperature, the afterglow step, and finally (iii) measurement of the thermoluminescent (TL) intensity during which the temperature is linearly increased. Charging temperatures could be varied from 213 to 353 K, while TL experiments were conducted at a heating rate of 12 K/min, up to a final temperature of 383 K, which was sufficient to empty all relevant traps. The charging step took 900 s followed by the afterglow for 1800s after which the TL measurement was started.
In the case of OSL measurements, the experiments consist of a charging step during 2500 s at a temperature where trapping occurs, e.g. at 253 or 273 K, followed by a fast cool down to a temperature of 213 K without excitation. This cooling down step takes less than 300 s. At 213 K excitation is applied for a certain time, the OSL step, followed by a TL experiment.
The local model was fitted to the measurements using the ’optimize.curve_fit’ module from the Python SciPy package [27, 28]. The algorithm uses the Levenberg-Marquardt algorithm to find a minimum through a least square method. To check that a global minimum was found the Python ’basinhopping’ and ’brute’ routines from the ’optimize’ package were used in a large neighbourhood of the solution point and χ2-maps were calculated for each pair of parameters to check the quality of the minimum.
3. Results and discussion
3.1. A local model
Understanding the (de)trapping processes in phosphors and the associated trapping defects is far from straightforward, because it involves identification of their chemical nature, energy levels and interactions (with other the recombination centers or possibly also with other traps), which are not simultaneously accessible via a single experimental technique [25, 29]. Regarding the energy levels and interactions, thermoluminescence is often used as a probe for persistent and storage phosphors. Many models to explain the experimental observations have been around, which can generally be classified into global [30–32], local [33–35] and semi-local models [36].
In a local model it is assumed that upon excitation an electron is released from the luminescent center (i.e. Eu2+ + hν → Eu3+ + e−) and that it is subsequently trapped in the proximity of the luminescent center, whatever the trap nature may be (the trivalent co-dopant, an oxygen vacancy or a lattice defect). When that electron is thermally released from its trap, it returns to the luminescent center where it originated from. There is thus a one-to-one relation between trap and luminescent center, with many such isolated ’systems’ present within the material. Locality means that these systems have no interaction among each other, such as electron migration through the lattice upon detrapping. In general, multiple traps could be connected to a single luminescent center as shown in Fig. 1, or there may be centers without accessible traps. The dashed line in Fig. 1 isolates the center with its trap(s); in a local model only transitions within the dashed line can occur and electrons cannot cross a dashed line.
Fig. 1 One-to-one relation between luminescent center and nearby trap(s). The arrows show possible trapping routes, each route characterized by a trapping rate pij.
For the sake of simplicity, we assume that there is only one trap next to each luminescent center; this simplified model is shown in Fig. 2 where M is the total number of centers in the material (i.e. the number of Eu ions), me is the number of systems in the material where the center is in excited state (i.e. Eu2+,* (4f65d)) and m is the number of systems where the center is ionized (Eu3+) and the electron is trapped. The number of Eu2+ ions in the 4f7 ground state is then given by M − me −m. The excitation rate is pe, pm is the radiative de-excitation rate, typically in the order of 106 s−1 and pnr is the non-radiative recombination rate.
Fig. 2 The model describing the local transitions for the case each system consists of one trap only.
Locality does not prohibit that non-radiative recombination takes place via the conduction band and as such is still compatible with the quenching mechanism suggested by Dorenbos [32] but the local model requires then that all electrons promoted to the conduction band will recombine non-radiatively.
Thermally assisted trapping was shown to occur for instance for the phosphors SrAl2O4:Eu,Dy and M2Si5N8:Eu (M=Ca, Sr, Ba) [16, 37]. This is taken into account in the model with a typical expression p1 = s exp(−E1/kT). Finally p2 = s exp(−E2/kT) is the thermally assisted detrapping rate.
The dynamical equations governing the time dependencies of m and me are then:
The luminescent intensity is given by:
This is a linear system which can be solved using eigenvectors and eigenvalues of the matrix representation. If m is defined as the column vector with components me and m, the solution for charging, with the boundary condition that at t = 0 all centers are in ground state or m(0) = me(0) = 0, is:
where vi are the eigenvectors of the matrix, λi the corresponding eigenvalues, vix,viy the components of the eigenvectors and v the determinant of the matrix constructed from the components of the eigenvectors. For afterglow the equations need to be solved setting pe = 0. The eigenvectors and eigenvalues for charging are thus different from those during afterglow. However for the experiments performed, the excitation rate pe is generally very small meaning that the eigenvalues and eigenvectors for charging and afterglow are equal. Consequently, also the exponentials during charging and afterglow can be set equal. The general solution for pe = 0 is:We require the function m(t) to be continuous going from excitation to afterglow, or mch(tch) = mag(0) where tch is the total charging time, so that the constants c1 and c2 can be determined.
It is easy to show that there is always a fast eigenvalue |λ1| ≥ pm corresponding to the photoluminescent decay time (in this case the 4f65d-4f7 transition in Eu2+ without the involvement of trapping). In (de)charging experiments one typically monitors the emission intensity at a sampling rate of 1 to 10 spectra per second, meaning that the exponential related to this fast eigenvalue is observed as a stepwise increase at the beginning of the excitation or drop at the end of excitation. Eq. (3) shows that the luminescent intensity scales linearly with pe or this stepwise increase at the beginning of the excitation equals the emission intensity drop at the end of excitation.
Three important conclusions can be drawn from this solution and these immediately allow to check if a persistent phosphor obeys the local model as described. (i) The eigenvalues and eigenvectors for charging and afterglow are equal within the experimental accuracy and consequently the exponential behavior found during charging and afterglow should be equal. (ii) The jump in emission intensity at the start of excitation equals the emission intensity drop at the end of excitation. (iii) The dynamics of charging and afterglow do not depend on the excitation intensity pe and a simple linear scaling is expected as function of the excitation intensity, as long as pe remains small.
3.2. Problems with the simple local model
Figure 3 shows charging, afterglow and TL measurements at a temperature of 273 K with varying excitation intensity. The measurement shows that the local model is not applicable as there are several problems. (i) The exponential behaviour during charging and afterglow is different. (ii) The jump in intensity at the start of excitation differs strongly from the drop after excitation. For example the initial intensity jump at t = 60 s for 14 mA driving current for the excitation source is approximately 2.7 times smaller than the intensity drop after charging at t = 960 s. (iii) The dynamics of charging heavily depend on the excitation intensity. The inset in Fig. 3 of the normalized emission intensities clearly shows that the curvature increases with increasing excitation intensity.
Fig. 3 Influence of the excitation intensity on charging, afterglow and TL measurement for Sr2MgSi2O7:Eu,Dy, at a charging temperature of 273 K. The TL measurement started at t = 2700 s.
Clearly if a local model would apply, there is something missing. Although it is not a priori correct to assume a local model for Sr2MgSi2O7:Eu,Dy it can be refined as shown below and a much improved result is obtained.
In the local model described by Eq. (1) excitation photons can only be absorbed by Eu2+ centers in the ground state. Their total number at any time is M − me −m, a number which decreases while the phosphor is being charged as the number of trapped charges m increases in the course of time. This implies that absorption is not a constant value but decreases over time and consequently that the reflected excitation intensity should increase.
The detected excitation intensity is the sum of the optically reflected intensity, due to refractive index mismatch and scattering, and the non-absorbed intensity inside the phosphor. The measured integrated excitation is shown in Fig. 4 where the excitation intensity was varied at a constant temperature. It clearly shows an opposite behaviour, with a decrease of the excitation intensity. This means that absorption is increasing during charging, opposite to what is expected from the local model.
Fig. 4 Reflected intensity of the excitation light as function of time for Sr2MgSi2O7:Eu,Dy at 273 K for different excitation intensities.
3.3. Local model with OSL
Taking the absorption behaviour into account, the most obvious correction to the local model is to add detrapping caused by the excitation source. The trap release rate is of course proportional to the incident excitation intensity; this comes down to introducing a proportionality constant α. The trap release rate is then α pe where α is the ratio between the trap release cross-section and the excitation cross-section for Eu2+. A simple modification of the model incorporating the release of trapped charges by the excitation source itself is shown in Fig. 5. To maintain locality it is assumed that the electron released from the trap goes back locally to its luminescent center. The result is that the rate α pe is added to the thermally assisted detrapping rate p2.
Fig. 5 The local model with OSL added and for which locality is also assumed for the OSL trap release.
The dynamical equations are changed to:
3.4. Effects of adding the OSL term and fitting to Sr2MgSi2O7 : Eu,Dy
Using the local OSL model equations Eq. (5) a general fit was tried to the charging and afterglow measurements of Sr2MgSi2O7:Eu,Dy for different temperatures. The TL measurements were not used in the fit as it is clear that by using a single trap only, the literature data [24] and measurements, which are indicative of a trap depth distribution, cannot be fully reproduced. However, when using the parameter set derived from charging and afterglow only, the TL glow curves are qualitatively reproduced.
The temperature dependent rates p1, p2 and pnr were allowed to vary independently for each temperature. This is especially required for the trap depth as a trap depth distribution is expected [24]. The energies were then calculated from these rates using a frequency or attempt-to-escape factor of s = 1014 s−1 which is of the order of the lattice vibration frequency [38], so the parameter s is not directly used as a fitting parameter. A lower frequency factor, e.g. s = 1012 s−1, only introduces a small shift in the calculated energies by an amount of 0.11 eV. In addition, no temperature dependency for s was taken into account. The photoluminescent decay rate was set fixed to the measured value of 1.66 106 s−1 for the lifetime of ≈600 ns at low temperature. The results are shown in Table 1 and Fig. 6. The excitation rate pe is indeed found to be very small and the OSL factor α is high.
Table 1. Parameter fit results for a local OSL model fit to Sr2MgSi2O7:Eu,Dy.
Fig. 6 Fitting results of the thermal barrier for trapping, thermal quenching barrier and trap depth as function of temperature.
The trap depth energies (Fig. 6), and more specifically the presence of a shallower and a deeper trap, are consistent with the behavior found by Laamanen et al. [24]. Their glow curve analysis suggested that the main TL peak could consist of two traps at 0.60 and 0.79 eV and that the shallow trap is emptied almost completely before the deeper one starts to be bleached to any extent.
Values for the thermal quenching barrier are relatively stable as function of temperature and agree with the values referenced by Dorenbos between 0.3 and 0.8 eV below the conduction band [39]. The variation could also be due to some influence of s not being temperature independent, which is then compensated via the thermal quenching energy. Note that above 260 K the thermal barrier for trapping is systematically lower than the barrier for thermal quenching, effectively allowing charge trapping. At low temperatures, fit errors on the energy barriers are expected to be high as the effect of a thermal barrier becomes small which could explain the peculiar behavior at low temperatures.
By fitting the local equations - including the OSL term - to charging and afterglow, all effects seem to be predicted up to the correct order as shown in Figs. 7 and 8. The predicted factor α is high but reasonable for Eu2+. Assume that the trap behaves as an F-center, then the ratio of the absorption cross section of an F-center to the excitation cross section of an Eu3+ is of the order of 106 [40] while the ratio of the oscillator strength of Eu2+/Eu3+ is of the order 104 [41] bringing the ratio α to an order of 100. When setting α = 0 no satisfying fit can be found as in this case the initial intensity jump at start of charging and the drop after charging makes a reasonable fit impossible. The measured afterglow is extremely small (Fig. 3), which is also the case for the model (Fig. 7). Due to the fact that a fixed trap depth is used instead of a trap depth distribution, differences between the model and the measurements are noticeable, although the general trend is well reproduced.
Fig. 7 Fit results (dashed lines) based on α = 228, compared to measurements (full lines) as function of charging temperature. The small peak on the emission intensity at the begin of charging, clearly seen at low temperatures, is due to the thermal relaxation of the LED excitation source after switching on.
Fig. 8 Using the fit results from Table 1, a calculation is done for variation of the excitation intensity pe at a charging temperature of 273 K. The simulated dynamic behaviour is similar to what is observed in the experiment of Fig. 3.
Addition of the OSL term explains (i) the difference in initial jump at charging and drop after charging. The eigenvalues and eigenvectors will be different for charging and afterglow as the term α pe can no longer be neglected. Therefore (ii) the dynamics of charging will depend in a more pronounced way on the excitation intensity due to the OSL term α pe. (iii) The absorption increases during charging instead of decreasing. The absorption rate is pe(M − me −m)+α pem ≈ peM +(α − 1)pem as the lifetime of the excited state is very small. If α > 1 absorption will increase as function of time as the number of trapped electrons m increases. And finally (iv), the maximum number of trapped charges will be limited. This qualitatively matches with XANES results where the change in the ratio Eu3+/Eu2+ upon illumination remains rather limited even with high illumination intensities [16]. It is clear that OSL is a very good candidate to explain several counterintuitive effects in the charging behaviour.
3.5. Experimental verification of OSL by the excitation light for Sr2MgSi2O7:Eu,Dy
From a mathematical point of view, it seems that in order to find a reasonable fit to Sr2MgSi2O7:Eu,Dy measurements, OSL by the excitation needs to be added. In addition, the increase in absorption during charging - although rather subtle - is a further convincing argument to support the OSL hypothesis.
An explicit experimental verification of the occurrence of OSL for Sr2MgSi2O7:Eu,Dy would put the theory on firm foot. However it is not easy to separate the effects of OSL and charging, as both are occurring simultaneously. In the specific case of Sr2MgSi2O7:Eu,Dy, we can make use of the thermal barrier for charging which prevents charging below 220 K. In this case, excitation of the phosphor does not lead to charge trapping, yet luminescence (and thus absorption) is occurring. The existence of a thermal barrier for trapping is clearly shown in Fig. 9 (bottom, blue curve) where excitation of the phosphor during 2500 s at 213K does not lead to an appreciable TL signal.
Fig. 9 (top) Emission intensity of Sr2MgSi2O7:Eu,Dy for TL following the charging at 253 K (green curve) and with additional excitation for 2500 s at 213 K between charging and TL (red curve). (middle) Temperature profiles for both experiments and (bottom) TL glow curves as a function of temperature. When charging the phosphor at 213 K no appreciable TL glow curve is noticed (blue curve), showing that no trapping occurs at this temperature.
The phosphor was excited at a temperature of 253 K, where trapping does occur, followed by a fast cool down to 213 K. The fraction of afterglow is then limited and a TL glow curve is recorded (Fig. 9, green curve). We are therefore able to bring the phosphor into a state which contains a significant amount of trapped charges.
Now the phosphor is excited at 213 K in between the charging at higher temperature (253 K) and the TL glow curve (Fig. 9, red curve). Note that the intensity and duration of excitation is identical during charging and excitation at low temperature. The much higher emission intensity at 213 K (as compared to the situation at 253 K) is due to the rather strong thermal quenching in this particular compound. The effect is a reduction of the area under the glow curve to 71% of the area in the case without excitation pulse at 213 K, unambiguously showing that the excitation light can lead to significant detrapping. In addition, the shape of the TL glow curve changed, with the major reduction noticed on the low temperature side. Experiments showed that by increasing the duration of the excitation pulse up to 12500 s at low temperature, the area of the TL peak further decreases down to 62% of its original value. Obviously, a reduction of the TL glow peak could also result from fading. It was explicitly verified that delaying the TL measurement for 2500 s when the sample is kept at low temperature does not change the shape and only reduces the integrated intensity of the glow curve by 3.5% showing that fading is not very relevant on these time scales (Fig. 10) and that the effect induced by the OSL is approximately one order of magnitude larger.
Fig. 10 Fading experiment showing that only 3% of the trapped charges are released at a temperature of 213 K, when the duration at low temperature is increased by 2500 s (red curve compared to green curve). Excitation was on during the initial 2500 s and then remained off for both experiments.
4. Conclusions and perspectives
We showed experimentally that the excitation source is detrapping previously trapped charges for Sr2MgSi2O7:Eu,Dy. This was achieved by making use of the thermal barrier for trapping in this particular persistent phosphor. TL experiments showed that the number of filled traps decreases while exciting the phosphor at 213 K, from which we can indeed conclude that the excitation source is releasing previously trapped charges. For this particular phosphor and excitation wavelength, OSL is preferentially clearing shallow traps. Consequently, TL spectra might be deformed depending on the charging temperature and the excitation wavelength, which could explain deviating results in literature obtained on similar materials. However, the reduction on the low temperature side of the TL glow curve is also reminiscent of higher order kinetics, which should be investigated in more detail.
A local model for persistent luminescence was developed where detrapping results from optical stimulation by the excitation light. The model predicts many aspects of the dynamics of charging and detrapping. It explains (i) the difference in exponentials during charging and afterglow, (ii) the difference of the initial jump of the luminescent intensity at the start of charging and the drop when charging is switched off, (iii) the dependency of the dynamics on the excitation intensity, and (iv) the increase in absorption during charging. The local OSL model fits the thermoluminescence experiments well, although the model is still very simple: no trap depth distribution and only one type of local ’systems’ was used. Further improvements to the model are expected by introducing a trap depth distribution. From IR-OSL [42, 43] measurements, it is known that OSL cross-sections are not only wavelength but also temperature dependent. Adding this dependency should further improve the model. We expect that charging dynamics will depend on the excitation wavelength and consequently this will result into different trap filling for different wavelengths.
Detrapping by the excitation source limits trap filling of a phosphor; there is a high trapping efficiency but it is countered by OSL trap release. Reducing the OSL cross-section, e.g. by appropriate choice of the trapping defects, would lead to an improved charging capacity of a persistent phosphor. Finding traps having a low OSL cross-section would thus definitely improve persistent luminescent materials.
Acknowledgments
The authors are grateful to Ghent University for financial support via the Enclose project (BOF-GOA). The authors would like to thank Jonas Botterman for technical assistance.
References and links
1. P. F. Smet, K. Van den Eeckhout, O. Q. De Clercq, and D. Poelman, “Chapter 274 - Persistent Phosphors,” in Handbook on the Physics and Chemistry of Rare Earths, J.-C. Bünzli and K. P. Vitalij, eds. (Elsevier, 2015). [CrossRef]
2. K. Van den Eeckhout, P. F. Smet, and D. Poelman, “Persistent luminescence in Eu2+-doped compounds: a review,” Materials 3, 2536–2566 (2010). [CrossRef]
3. T. Matsuzawa, Y. Aoki, N. Takeuchi, and Y. Muramaya, “A new long phosphorescent phosphor with high brightness, SrAl2O4:Eu2+Dy3+,” J. Electrochem. Soc. 143, 2670 (1996). [CrossRef]
4. K. Van den Eeckhout, D. Poelman, and P. F. Smet, “Persistent luminescence in non-Eu2+-doped compounds: A Review,” Materials 6, 2789–2818 (2013). [CrossRef]
5. D. Poelman, N. Avci, and P. F. Smet, “Measured luminance and visual appearance of multi-color persistent phosphors,” Opt. Express 17, 358–364 (2009). [CrossRef] [PubMed]
6. D. Poelman and P. F. Smet, “Photometry in the dark: time dependent visibility of low intensity light sources,” Opt. Express 18, 26293–26299 (2010) [CrossRef] [PubMed]
7. J. Botterman and P. F. Smet, “Persistent phosphor SrAl2O4:Eu,Dy in outdoor conditions: saved by the trap distribution,” Opt. Express 23, A868–A881 (2015). [CrossRef] [PubMed]
8. Y. Zhuang, Y. Katayama, J. Ueda, and S. Tanabe, “A brief review on red to near-infrared persistent luminescence in transition-metal-activated phosphors,” Opt. Mater. 36, 1907–1912 (2014). [CrossRef]
9. T. Maldiney, A. Bessiere, J. Seguin, E. Teston, S. K. Sharma, B. Viana, A. J. J. Bos, P. Dorenbos, M. Bessodes, D. Gourier, D. Scherman, and C. Richard, “The in vivo activation of persistent nanophosphors for optical imaging of vascularization, tumours and grafted cells,” Nat. Mater. 13, 418–426 (2014). [CrossRef] [PubMed]
10. Q. le Masne de Chermont, C. Chanac, J. Seguin, F. Pelle, S. Maitrejean, J.-P. Jolivet, D. Gourier, M. Bessodes, and D. Scherman, “Nanoprobes with near-infrared persistent luminescence for in vivo imaging,” Proc. Natl. Acad. Sci. U. S. A. 104, 9266–9271 (2007). [CrossRef] [PubMed]
11. D. Gourier, A. Bessiere, S. K. Sharma, L. Binet, B. Viana, N. Basavaraju, and K. R. Priolkar, “Origin of the visible light induced persistent luminescence of Cr3+-doped zinc gallate,” J. Phys. Chem. Solids 75, 826–837 (2014). [CrossRef]
12. K. Korthout, K. Van den Eeckhout, J. Botterman, S. Nikitenko, D. Poelman, and P. F. Smet, “Luminescence and x-ray absorption measurements of persistent SrAl2O4:Eu,Dy powders: Evidence for valence state changes,” Phys. Rev. B 84, 085140 (2011). [CrossRef]
13. P. Dorenbos, “Thermal quenching of Eu2+ 5d-4f luminescence in inorganic compounds,” J. Phys.: Condens. Matter 17, 8103–8111 (2005).
14. A. Dobrowolska, A. J. J. Bos, and P. Dorenbos, “Electron tunnelling phenomena in YPO4:Ce, Ln (Ln = Er, Ho, Nd, Dy),” J. Phys. D-Appl. Phys. 47, 335301 (2014). [CrossRef]
15. K. Van den Eeckhout, A. J. J. Bos, D. Poelman, and P. F. Smet, “Revealing trap depth distributions in persistent phosphors,” Phys. Rev. B 87, 045126 (2013). [CrossRef]
16. J. Botterman, J. J. Joos, and P. F. Smet, “Trapping and detrapping in SrAl2O4:Eu,Dy persistent phosphors: Influence of excitation wavelength and temperature,” Phys. Rev. B 90, 085147 (2014). [CrossRef]
17. H. Hagemann, D. Lovy, S. Yoon, S. Pokrant, N. Gartmann, B. Walfort, and J. Bierwagen, “Wavelength dependent loading of traps in the persistent phosphor SrAl2O4:Eu2+, Dy3+,” J. Lumines. 170, 299–304 (2016). [CrossRef]
18. E. G. Yukihara, S. W. S. McKeever, and M. S. Akselrod, “State of art: optically stimulated luminescence dosimetry - Frontiers of future research,” Radiat. Meas. 71, 15–24 (2014). [CrossRef]
19. P. Leblans, D. Vandenbroucke, and P. Willems, “Storage phosphors for medical imaging,” Materials 4, 1034–1086 (2011). [CrossRef]
20. Y. Zhuang, J. Ueda, and S. Tanabe, “Tunable trap depth in Zn(Ga1−x Alx)2O4:Cr,Bi red persistent phosphors: considerations of high-temperature persistent luminescence and photostimulated persistent luminescence,” J. Mater. Chem. C 1, 7849–7855 (2013). [CrossRef]
21. A. J. J. Bos, N. R. J. Poolton, J. Wallinga, A. Bessiere, and P. Dorenbos, “Energy levels in YPO4:Ce3+,Sm3+ studied by thermally and optically stimulated luminescence,” Radiat. Meas. 45, 343–346 (2010). [CrossRef]
22. J. Ueda, T. Shinoda, and S. Tanabe, “Photochromism and near-infrared persistent luminescence in Eu2+-Nd3+-co-doped CaAl2O4 ceramics,” Opt. Mater. Express 3, 787–793 (2013). [CrossRef]
23. Y. Jin, Y. Hu, Y. Fu, Z. Mu, and G. Ju, “Reversible white and light gray photochromism in europium doped Zn2GeO4,” Mater. Lett. 134, 187–189 (2014). [CrossRef]
24. H. F. Brito, J. Holsa, H. Jungner, T. Laamanen, M. Lastusaari, M. Malkamki, and L. C. V. Rodrigues, “Persistent luminescence fading in Sr2MgSi2O7:Eu2+,R3+ materials: a thermoluminescence study,” Opt. Mater. Express 2, 287–293 (2012). [CrossRef]
25. H. Duan, Y. Z. Dong, Y. Huang, Y. H. Hu, and X. S. Chen, “First-principles study of intrinsic vacancy defects in Sr2MgSi2O7 phosphorescent host material,” J. Phys. D-Appl. Phys. 49, 025304 (2016). [CrossRef]
26. Z. Pan, Y.-Y. Lu, and F. Liu, “Sunlight-activated long-persistent luminescence in the near-infrared from Cr3+-doped zinc gallogermanates.,” Nat. Mater. 11, 58–63 (2012). [CrossRef]
27. S. van der Walt, S.C. Colbert, and G. Varoquaux, “The NumPy array: a structure for efficient numerical computation,” Comput. Sci. Eng. 13, 22–30 (2011). [CrossRef]
28. J. D. Hunter, “Matplotlib: a 2D graphics environment,” Comput. Sci. Eng. 9, 90–95 (2007). [CrossRef]
29. H. F. Brito, J. Holsa, T. Laamanen, M. Lastusaari, M. Malkamaki, and L. C. V. Rodrigues, “Persistent luminescence mechanisms: human imagination at work,” Opt. Mater. Express 2, 371–381 (2012). [CrossRef]
30. S. W. S. McKeever and R. Chen, “Luminescence models,” Radiat. Meas. 27, 625–661 (1997). [CrossRef]
31. R. Chen, J. L. Lawless, and V. Pagonis, “Two-stage thermal stimulation of thermoluminescence,” Radiat. Meas. 47, 1–5 (2012). [CrossRef]
32. P. Dorenbos, “Mechanism of persistent luminescence in Eu2+ and Dy3+ codoped aluminate and silicate compounds,” J. Electrochem. Soc. 152, H107–H110 (2005). [CrossRef]
33. D. Jia, “Charging curves and excitation spectrum of long persistent phosphor SrAl2O4:Eu2+,Dy3+,” Opt. Mater. 22, 65–69 (2003). [CrossRef]
34. M. Jain, B. Guralnik, and M. T. Andersen, “Stimulated luminescence emission from localized recombination in randomly distributed defects,” J. Phys.: Condens. Matter 24, 385402 (2012).
35. F. Clabau, X. Rocquefelte, S. Jobic, P. Dienard, M. H. Whangbo, A. Garcia, and T. Le Mercier, “Mechanism of phosphorescence appropriate for the long-lasting phopshors Eu2+-doped SrAl2O4 with codopants Dy3+ and B3+,” Chem. Mater. 17, 3904–3912, (2005) [CrossRef]
36. V. Pagonis, L. Blohm, M. Brengle, G. Mayonado, and P. Woglam, “Anomalous heating rate effect in thermoluminescence intensity using a simplified semi-local transition (SLT) model”, Radiat. Meas. 51, 40–47 (2013). [CrossRef]
37. P. F. Smet, K. Van den Eeckhout, A. J. J. Bos, E. van der Kolk, and P. Dorenbos, “Temperature and wavelength dependent trap filling in M2Si5N8:Eu (M = Ca, Sr, Ba) persistent phosphors,” J. Lumines. 132, 682–689 (2012). [CrossRef]
38. A.J.J. Bos, “Theory of thermoluminescence,” Radiat. Meas. 41, S45–S56 (2007). [CrossRef]
39. P. Dorenbos, “Mechanism of persistent luminescence in Sr2MgSi2O7:Eu2+,Dy3+,” Phys. Status Solidi B 242, R7–R9 (2005). [CrossRef]
40. C. R. Ronda, Luminescence, from Theory to Applications (Wiley, 2007)
41. S. Shionoya, Phosphor Handbook (CRC Press, 2006)
42. S.W.S. McKeever, L. Bøtter-Jensen, N. A. Larsen, and G. A. T. Duller, “Temperature dependence of OSL decay curves: experimental and theoretical aspects,” Radiat. Meas. 27, 161–170 (1997). [CrossRef]
43. N.A. Spooner, “On the optical dating signal from quartz,” Radiat. Meas. 23, 593–600 (1994). [CrossRef]
