Synchrotron radiation spectroscopy of rare earth doped persistent luminescence materials

Synchrotron radiation spectroscopy of rare earth doped persistent luminescence materials

ARTICLE IN PRESS Radiation Physics and Chemistry 78 (2009) S11–S16 Contents lists available at ScienceDirect Radiation Physics and Chemistry journal...

403KB Sizes 7 Downloads 42 Views

ARTICLE IN PRESS Radiation Physics and Chemistry 78 (2009) S11–S16

Contents lists available at ScienceDirect

Radiation Physics and Chemistry journal homepage: www.elsevier.com/locate/radphyschem

Synchrotron radiation spectroscopy of rare earth doped persistent luminescence materials ¨ a¨ a,b,, Taneli Laamanen a,c, Mika Lastusaari a,b, Marja Malkamaki ¨ a,c, Jorma Hols a,1 d Janne Niittykoski , Pavel Nova k a

Department of Chemistry, University of Turku, FI-20014 Turku, Finland Turku University Centre for Materials and Surfaces (MatSurf), Turku, Finland c Graduate School of Materials Research (GSMR), Turku, Finland d Institute of Physics, Academy of Sciences of the Czech Republic, CZ-16253 Prague 6, Czech Republic b

a r t i c l e in fo

abstract

Article history: Received 19 January 2009 Accepted 26 February 2009

The electronic structure of the polycrystalline CaAl2O4:Eu2+,Ce3+ persistent luminescence materials were studied with X-ray absorption (XANES) and UV–VUV emission and excitation spectroscopy by using synchrotron radiation. Theoretical calculations using the density functional theory (DFT) were carried out simultaneously with the experimental work. The experimental band gap energy (Eg) value of 6.7 eV agrees very well with the DFT value of 6.4 eV. From the 4f7-4f65d1 excitation bands of Eu2+, the positions of the 4f7 ground as well as the 4f65d1 excited levels were established. The excitonic fine structure which could act as trap levels close to the bottom of the conduction band could not be observed, however. The different processes contributing to the mechanism of persistent luminescence from CaAl2O4:Eu2+ were constructed and discussed. & 2009 Elsevier Ltd. All rights reserved.

Keywords: Calcium aluminate Europium Persistent luminescence Mechanism Synchrotron radiation DFT calculations

1. Introduction The persistent luminescence materials have been known and exploited for hundreds of years since the beginning of the 17th century (Newton Harvey, 1957). Despite the long history of the materials, the understanding of even the mere basics of the phenomenon itself has not been at the same level. No progress occurred earlier than the mid 1990s prior to the advent of the commercial exploitation of the modern and efficient persistent luminescence materials, the Eu2+ doped and R3+ (R: rare earth) codoped alkaline earth aluminates (MAl2O4:Eu2+,R3+; M: Ca, Sr, Ba) (Matsuzawa et al., 1996). As expected, only then serious interest was awakening to the study of the mechanisms underlying the persistent luminescence in a manner not uncommon to the breakthrough of the ceramic high Tc superconducting materials. From the rather simplistic start neither considering the energetics nor the veracity of the actual players in the phenomenon, there have been more and more elaborate models brought up as the mechanisms (e.g. Chen et al., 2006; Zhang and Wang,

 Corresponding author. Tel.: +358 2 333 6737; fax: +358 2 333 6730.

¨ a), ¨ [email protected]fi (T. Laamanen). E-mail addresses: [email protected]fi (J. Hols ¨ [email protected]fi (M. Lastusaari), [email protected]fi (M. Malkamaki). [email protected] (J. Niittykoski), [email protected] (P. Nova k). 1 Present address: OMG Kokkola Chemicals Ltd., P.O. Box 286, FI-67101 Kokkola, Finland. 0969-806X/$ - see front matter & 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.radphyschem.2009.02.012

2007; Dorenbos, 2005a; Aitasalo et al., 2004b, 2006). However, as it is the case with the photostimulated luminescence – possibly related to persistent luminescence – no completely convincing or proven mechanisms have yet appeared in the scientific literature. There seem to be two critical obstacles preventing the theoretical clearing up of the persistent luminescence mechanism(s). Despite the early work (Pedrini et al., 1986; Thiel et al., 2002) – later a great deal elaborated (e.g. Dorenbos, 2005b, 2007, 2008) – on the relationships between the energy levels of the R2+/3+/IV ions and the electronic band structure of the host lattice, the resulting energy level structures are by far too inaccurate or too painstaking to be attained in such a detail to be useful in the elucidation of persistent luminescence. In addition, not much is known about the defect energy levels in solids although it is expected that different kinds of defects may play a major role in the persistent luminescence mechanism. As the second major obstacle, the possible changes in the valence states of the dopants (e.g. Eu2+, Ce3+, Eu3+ or Tb3+) during the persistent luminescence are not known at all. The energy level  the host band structure can positions of the R2+/3+/IV ions vis-a-vis give some information about these changes though experimental verification is still lacking in most cases. Already in the very beginning of the studies on the persistent luminescence mechanisms, it was assumed that such valence changes occur (e.g. Chen et al., 2006; Zhang and Wang, 2007; Dorenbos, 2005a; Aitasalo et al., 2004b, 2006). As for the veracity of these species assumed

ARTICLE IN PRESS S12

J. H¨ ols¨ a et al. / Radiation Physics and Chemistry 78 (2009) S11–S16

to be present, it may now be stated that the occurrence of such ions as the monovalent Eu+, is not at all conceivable if common sense is used. Furthermore, such a species is not fitting to any of the energy level models referred to above. Nonetheless, the presence of tetravalent ions as Ce4+ may be plausible and fit into these models, too. In the present work, the luminescence properties of the Eu2+ doped and R3+ (especially Ce3+) co-doped CaAl2O4 persistent luminescence materials were investigated using synchrotron radiation (SR) UV–VUV excitation in order to establish at different temperatures the relationships between the (co-)dopants’ energy levels and the electronic host band structure. In addition to the experimental determination of the host band structure, a simultaneous theoretical study was carried out with the density functional theory (DFT) calculations. Further, the presence of the Eu2+/3+ and Ce3+/IV ions were probed by SR X-ray absorption methods (XANES). Eventually, a persistent luminescence mechanism was suggested based on these experimental and theoretical results.

2. Experimental 2.1. Materials preparation

2.3. Density functional theory calculations The electronic structure of the SrAl2O4 materials were calculated using the WIEN2k package (Blaha et al., 2001). WIEN2k is based on the full potential linearized augmented plane wave method, an approach which is among the most precise and reliable ways to calculate the electronic structure of solids. The generalized gradient approximation (GGA) was employed.

3. Results and discussion 3.1. Luminescence of CaAl2O4:Eu2+,R3+ The luminescence spectra of the CaAl2O4:Eu2+,R3+ materials excited to the conduction band by SR were characterized by a strong band at ca. 440 nm (22500 cm1) due to the 4f65d1-4f7 transition of the Eu2+ ion (Figs. 1 and 2). The FWHM value of this band is quite low, 2250 cm1, which indicates that the emission originates from one Eu2+ centre only despite the fact that there are three different Ca2+ sites in the monoclinic CaAl2O4 structure ¨ ¨ (Horkner and Muller-Buschbaum, 1976). However, there is only one nine-fold coordinated Ca2+ site in addition to two seven-fold coordinated ones. Only this single site has sufficient spacing to allow for the substitution by the large Eu2+ (ionic radii:

The polycrystalline CaAl2O4:Eu2+,R3+ materials were prepared with a solid state reaction between stoichiometric amounts of CaCO3, Al2O3 and rare earth oxides (Aitasalo et al., 2004a). A small amount of B2O3 (1 mol%) was used as a flux. The nominal concentration of both the Eu2+ and R3+ ions was 1 mol% of the calcium amount. The starting materials were ground to a homogeneous mixture in a ball mill. The mixtures were then annealed in a reducing (N2+10% H2) atmosphere for 1 and 4 h at 900 and 1300 1C, respectively. The structural and phase purity was confirmed by X-ray powder diffraction using a Huber 670 image ˚ No plate Guinier-camera at 295 K (CuKa1 radiation, 1.5406 A). additional phases were found in the materials.

2.2. Synchrotron radiation measurements The UV–VUV excitation spectra of the CaAl2O4:Eu2+,R3+ materials were measured between 3.7 and 40 eV by using the synchrotron radiation at the SUPERLUMI beamline of HASYLAB at DESY (Hamburg, Germany) (Laasch, 2009a). The samples were mounted on the cold finger of a liquid He flow cryostat. The spectra were recorded at selected temperatures between 10 and 298 K with a 2-m primary McPherson monochromator attaining a resolution between 0.02 and 0.32 nm. The emission spectra were obtained both with a Spectra Pro 300i monochromator (200–800 nm) equipped with a conventional photomultiplier and a CCD detector (200–1050 nm). Since the SR has a pulse separation of ca. 190 ns, the time resolution of the emission spectra are somewhat impaired owing to the conventional Eu2+ decay time being close to 1 ms. The SR X-ray absorption (XAFS) spectra were measured between 5 and 9 keV at the E4 beamline of HASYLAB (Laasch, 2009b). The spectra corresponding to the LIII edge of the lanthanides were recorded at selected temperatures between 10 and 298 K with a He flow cryostat setup similar to the one described above. Due to the strong host absorption, the EXAFS/ XANES measurements were carried out in the fluorescence mode using a 7 pixel Si(Li) detector. The resolution of the experimental setup was around 1 eV in the energy range used.

Fig. 1. Schematic energy level diagram of the emission and selected excitation processes in the Eu2+ doped CaAl2O4.

Fig. 2. Synchrotron radiation excited emission spectra of both Eu2+ doped and non-doped CaAl2O4 at 10 K (SUPERLUMI, HASYLAB).

ARTICLE IN PRESS J. H¨ ols¨ a et al. / Radiation Physics and Chemistry 78 (2009) S11–S16

Ca2+(CN ¼ 7): 1.06, Ca2+(CN ¼ 9): 1.18, Eu2+(CN ¼ 7): 1.20 and Eu2+(CN ¼ 9): 1.30 A˚ (Shannon, 1976)). The single-band emission is thus consistent with the structural data. The band widths are not significantly broadened due to the introduction of the R3+ codopant ions. This observation suggests that either the defects owing to the charge compensation due to the R3+ co-dopants occupying the Ca2+ site are far away from Eu2+ or the modification of the crystal field around Eu2+ is negligible. Usually, the crystal field is strongly modified by the closely situating lattice defects and thus the former explanation seems more probable. However, further studies are needed to clarify this matter. The SR-excited emission spectra are practically identical with the conventional UV excited ones as well as with those obtained for the persistent luminescence (Aitasalo et al., 2006). The luminescence of the non-doped CaAl2O4 excited with similar radiation as the Eu2+ doped (and R3+ co-doped) CaAl2O4 is very much different from the latter. First of all, this emission is much weaker – even at 10 K – than the Eu2+ emission. In addition, the emission consists of very broad bands, the main one situating in the UV range (centred at 290 nm/34200 cm1) with a FWHM value of 8000 cm1. This kind of emission can be connected with defects occurring in materials. Despite the fact that now there are no Eu2+ or R3+ doping present, the materials were prepared in a reducing (N2+10% H2) gas sphere and thus the presence of defects is probable. The reducing annealing conditions favour the ¨ formation of the oxygen vacancies (the Kroger–Vink notation: Vdd o ) though at higher annealing temperatures the evaporation of metal oxides becomes possible (Trojan-Piegza et al., 2008) and the 00 formation of both the oxygen (Vdd o ) and metal (e.g. VCa) vacancies may occur. Presently, it should be assumed that the temperature is not high enough to produce the metal ion vacancies. The production of the oxygen vacancies is accompanied with the formation of colour (F or F+) centres with no (F) or a positive (F+ centres) net charge due to the occupation of original oxygen vacancies with electron(s). Energetically, these colour centres are located closer to the conduction band rather than to the valence band. The emission can, however, occur at different wavelengths from UV/VUV to infrared as can be observed for the non-doped CaAl2O4 (Fig. 2). In addition to the colour centre emission, emission due to the annihilation of different kinds of excitons may result in the emission of the non-doped CaAl2O4. Although more elaborate studies are needed to identify the different emitting species for the persistent luminescence, it is evident that there are defects present in the materials. To close this section, it should be noted that the defect related luminescence is either absent or too weak to be noticed along with the Eu2+ emission from the (co-)doped materials. The former indicates efficient energy transfer from the defects to the lattice.

3.2. Excitation of CaAl2O4:Eu2+,R3+ The excitation of the Eu2+ luminescence in the CaAl2O4:Eu2+,R3+ materials can be a very complicated process (Fig. 1). In addition to the interconfigurational 4f7-4f65d1 transitions (processes within Eu2+), the excitation can take place over the band gap or, finally, via the defects which include also the co-dopant R3+ ions. At 10 K, the temperature at which the persistent luminescence is not yet smearing out the details in the excitation spectrum, the 4f7-4f65d1 transitions observed as broad bands span over the energy range from the emission at 440 nm (ca. 23,000 cm1) and excitation at 315 (32,000), 250 (40,000) and finally at the weak band at 210 nm (ca. 48,000 cm1). To the energies given above, the relaxation should be added/ subtracted: the energy of the emitting level is increased while those of the other levels are decreased. The total span of the

S13

crystal field split 4f65d1 level of Eu2+ in halide hosts is some 20,000 cm1 (Dorenbos, 2003), but probably closer to 25,000 cm1 in oxides as CaAl2O4 due to a stronger crystal field. Accordingly, it may be concluded that the weak band at 210 nm corresponds to the highest 4f65d1 level of Eu2+. In principle, the 4f65d1 electron configuration (the 2D level) should be split into five components by the crystal field of low symmetry (as C1 in CaAl2O4). The remaining part of the 4f65d1 configuration (i.e. 4f6) should be visible as a fine structure composed of seven (7FJ; J ¼ 0-6) sharper bands (van Haecke, 2007). In the present case, not all of these five main components nor the fine structure can be observed. The absence of the former may be due to two facts: the missing component of the 2D level may be overlapping with another stronger absorption or the band is just too weak to be observed. The weakness of the band at 210 nm is due to the great mobility of the electron lost to the conduction band when it is first promoted into the 4f65d1 levels within the conduction band. Instead, the excitation to the 4f65d1 levels below the conduction band can be observed as neat bands at lower energies in the excitation spectrum of Eu2+. To establish the first relationship between the Eu2+ energy level scheme and the host band structure, the weak intensity of the band at 210 nm is used: it gives the upper limit to the energy difference between the 4f7 ground level of Eu2+ and the bottom of the conduction band. The lower limit is given by the well-resolved band at 315 nm. Together with the correction due to the relaxation within the 4f65d1 level, the 4f7 ground level of Eu2+ lies some 30,000–27,000 cm1 below the bottom of the conduction band. On the other hand, the charge transfer band of Eu3+ (not shown here), corresponding approximately to the energy difference between the 4f7 ground level of Eu2+ and the top of the valence band, is at 35,000 cm1 (Aitasalo et al., 2006) (and with the relaxation, ca. 32,000 cm1). According to the conventional wisdom, the band gap energy of CaAl2O4 is then given as the sum of these two energies (Fig. 1), thus ca. 59,000 cm1. In addition to the interconfigurational 4f7-4f65d1 transitions of Eu2+, the excitation of Eu2+ can occur via the host lattice. The band gap energy of CaAl2O4 has previously been estimated to 46,800 cm1 (5.8 eV) (Jia and Yen, 2003b). As the most facile excitation to be determined experimentally—provided a suitable excitation source is available, of course, the valence to conduction band excitation could be observed as an abrupt increase in the Eu2+ emission intensity with excitation at 181.2 nm (6.84 eV, 55,000 cm1). This energy gives the host valence to conduction band edge energy calculated approximately in the preceding chapter. The two values, 59,000 and 55,000 cm1, are in a rather good agreement with each other though the difference underlines the importance of the relaxation within the excited levels following the excitation—either by the charge transfer to Eu3+ or the 4f7-4f65d1 transitions of Eu2+. In fact, the relaxation of some 5000 cm1 seems to be insufficient though the inaccuracy in reading the exact transition energy may play a role, as well. In any case, the present work gives important information about the relaxation energies which are otherwise difficult to obtain. The valence to conduction band excitation for the Eu2+ doped and the non-doped CaAl2O4 are slightly different as shown by the 1st derivate curves (insert of Fig. 3). Since there is no sign of the excitonic features frequently associated to the absorption edge in either excitation spectra, it should be concluded that the excitons are situated very close to the absorption edge and cannot be resolved even at 10 K. Alternatively, these excitons are not stable enough even at 10 K. The edge energies differ only by 0.1 eV which is much less than the trap energies previously calculated from the thermoluminescence glow curves (Aitasalo et al., 2006). Otherwise the two excitation spectra are very similar including the fine structure above the absorption edge. Further studies are

ARTICLE IN PRESS S14

J. H¨ ols¨ a et al. / Radiation Physics and Chemistry 78 (2009) S11–S16

2+

Fig. 3. UV–VUV synchrotron radiation excitation spectra of both Eu doped and non-doped CaAl2O4 at 10 K (SUPERLUMI, HASYLAB). Insert: the 1st derivate curve of the excitation spectrum.

Fig. 4. Calculated density of states (DOS) for the CaAl2O4 host material with the DFT (GGA) method. Comparison is made with the UV–VUV excitation spectrum of CaAl2O4:Eu2+ at 10 K.

thus needed to clarify the effect of defects to the UV–VUV excitation of the CaAl2O4:Eu2+ persistent luminescence materials. As an intermediary conclusion it can be stated that, in case of CaAl2O4:Eu2+, the synchrotron radiation excitation spectroscopy at 10 K is not able to reveal the energy level structure of defects which are supposed to situate close to the bottom of the conduction band acting as traps to store the energy for further use as persistent luminescence. 3.3. Electronic band structure of CaAl2O4 by DFT calculations There exists no previous calculation of the electronic band structure of CaAl2O4. Thus, the first and basic objective for the DFT calculations was the estimation of the band gap energy of the Eu2+ doped and non-doped CaAl2O4 materials. As the outcome from the DFT calculations employing the GGA method, the Eg value of 6.4 eV (Fig. 4) was obtained which value is very close to the experimental one (6.7 eV) (Fig. 3). The agreement between the experimental and calculated values is very good since the difference of 0.3 eV can be considered insignificant. However, both the experimental and calculated values are somewhat uncertain since, in both cases, either very weak excitation or close to non-zero density of states (DOS) can be observed above ca. 5.5 eV up to the main shifts at 6.7 and 6.4 eV, respectively. According to the DFT calculations, the valence band has mainly the O(2p) character, whereas the bottom of the conduction band consists mostly of the Ca states (Fig. 4). Most of the Al DOS is located deep in the conduction band. The structure of the valence and conduction bands as well as the Eg values did not change significantly whether the optimized or non-optimized crystal structures of the Eu2+ doped or non-doped CaAl2O4 material were used as an input. The calculated fine structure of the conduction band is quite different from the measured excitation spectrum, but this is due to the fact that the excitation to anywhere in the conduction band can take place from anywhere in the valence band provided the DOS has a non-zero value. In fact, only the absorption edge is a reliable value and has the most importance. 3.4. Valence state of Eu2+ and Ce3+ in CaAl2O4 As most of the persistent luminescence mechanisms presented so far have suggested (e.g. Jia et al., 2000; Kodama et al., 2002; Lin et al., 2003; Jia and Yen, 2003a; Clabau et al., 2005), there will be a change in the valence state of Eu2+. Whether this is to Eu+ or to Eu3+, there exist differing opinions. If one uses common sense, the

Fig. 5. Synchrotron radiation XANES spectra for the Eu2+ doped CaAl2O4 at 10 and 296 K (E4 beamline, HASYLAB).

monovalent ion could easily be ruled out but using a scientific approach, one should have a definite proof to reject any solution despite how irrational it may sound. The change of the (co)dopants’ valence state should occur when they are irradiated and thus the X-ray absorption methods offer both the means of irradiation and the observation of the change. As a target of the present studies, the Eu2+ doped and Ce3+ co-doped CaAl2O4 material was chosen because in this system both the Eu2+ dopant and the Ce3+ co-dopant can undertake easily another valence, Eu3+ and Ce4+, respectively. The temperature of measurement should also play an important role since at low temperatures as 10 K, the persistent luminescence should not occur because the thermal energy available is too low and only the storage of the energy takes place. As a result, the number of species in their temporary valence state should increase and facilitate their observation. It seems, however, that both at 10 K and at room temperature, the divalent europium is the only species present in the Eu2+ doped CaAl2O4 materials since only one signal at 6980 eV (6988 eV for Eu3+) can be observed (Fig. 5). Neither the measurement temperature nor the time does play any role. The measurement time should increase the number of the Eu3+ species. As for the Eu3+ species, not even a trace of the monovalent Eu+ species could be observed at either temperature. The presence of these two simple species was thus ruled out in the persistent luminescence mechanisms. However, instead of the outright formation of the Eu3+ species, one can

ARTICLE IN PRESS J. H¨ ols¨ a et al. / Radiation Physics and Chemistry 78 (2009) S11–S16

S15

Fig. 7. Persistent luminescence mechanism for the CaAl2O4:Eu2+,R3+ materials.

Fig. 6. Synchrotron radiation XANES spectra for the Eu2+ doped and Ce3+ co-doped CaAl2O4 at 10 and 296 K (E4 beamline, HASYLAB).

assume the formation of a Eu2+–h+ pair similar to the Tb3+–h+ pair in the Tb3+ doped Lu2O3 persistent luminescence material (TrojanPiegza et al., 2008). This pair should, however, distort the lattice and could be possibly seen in the EXAFS analysis of the XAS spectra to be carried out in the near future. The co-doping of the basic Eu2þ doped CaAl2O4 material with the R3þ ions will in most cases change dramatically the persistent luminescence properties of the material although the effect is not always a positive one. In addition to Nd3+ which drastically improves the persistent luminescence shown by CaAl2O4:Eu2+, also the Ce3+ co-doping has a positive effect though to a lesser extent (Aitasalo et al., 2006). The co-doping with Ce3+ changes the XANES spectrum of the CaAl2O4:Eu2+ material by introducing the signal corresponding to the Eu3+ species at both temperatures (Fig. 6). At the same time, there is only the signal for the Ce3+ visible. One could interpret these results as follows: the formation of Eu3+ means the photoionization of Eu2+ and subsequent energy storage in the form of a trapped electron. However, this could require the simultaneous reduction of Ce3+ to Ce2+. The latter species can probably be found at extreme conditions though they are not present in these aluminates and, to be sure, the XANES spectra give no evidence for the Ce2+ species. Accordingly, the analysis of the XANES spectra must be carried out in a different manner. First clue is given by the form of the cerium added into the original oxide mixture: cerium was added in the tetravalent rather than in the trivalent form. As a result, the Eu2+ ion formed from the Eu3+ originally employed was oxidized back to Eu3+ and CeIV reduced to Ce3+. Initially, the nominal cerium and europium concentrations were about the same but since the reduction/ oxidation of cerium is easier that those of europium, there is still much Eu2+ left at the end of the process. Evidently, also the reducing N2+H2 gas sphere helps. As a conclusion, there was no evidence obtained for the formation of the Eu3+ species owing to the persistent luminescence mechanism and thus the Eu2+–h+ pair formation is to be preferred. A complementary issue arisen from the XANES measurements on the CaAl2O4:Eu2+,Ce3+ material, is the question how the Ce3+ co-doping could improve the persistent luminescence of the CaAl2O4:Eu2+ material if a considerable amount of Eu2+ is transformed to Eu3+? During this process one loses much of the Eu2+ luminescence recombination centres—which were below the optimum concentration, anyhow. The solution to this problem is

that the increase in the Eu3+ concentration increases the number of lattice defects which act as the centres to store the excitation energy. It must be realized that the performance of the persistent luminescence material is a trade-off between the energy storage capacity and the luminescence intensity. Unfortunately, these two requirements affect usually opposite to each other rather than parallel. 3.5. Persistent luminescence mechanism The final aim of the work carried out above was to develop a credible persistent luminescence mechanism. Since the late 1990s, numerous reports have been made available on persistent luminescence (e.g. Aitasalo et al., 2006; Trojan-Piegza et al., 2008; Jia and Yen, 2003a, 2003b; Jia et al., 2000; Kodama et al., 2002; Lin et al., 2003; Clabau et al., 2005) but only a few mechanisms have survived the progress made in the understanding of the complex experimental data. To avoid most of the issues contradicting the common chemical and physical sense, one can put together the wealth of information presented above as well as available in the scientific literature as the following mechanism for the CaAl2O4:Eu2+, R3+ materials (Fig. 7): the irradiation of the material by blue light (or UV radiation) involves the photoexcitation of Eu2+ via the 4f7-4f65d1 transitions which, as stated above, overlap with the conduction band of CaAl2O4. The capture of the excited electron to the conduction band together with its rapid movement therein, leads to the capture of the electron by traps close to the bottom of the conduction band. Since the photoionization of Eu2+ to Eu3+ is not proven, one should assume instead the creation of a Eu2+–h+ pair. The reverse process, the actual persistent luminescence, involves the gradual release of the trapped electrons – controlled by the thermal energy via the Boltzmann distribution – followed by the migration of electrons to the Eu2+ luminescence recombination centre. No evident pitfalls exist in this mechanism though proving this is hard due to the ubiquitous uncertainty about the thermally controlled mechanism.

4. Conclusions The synchrotron radiation in the UV–VUV and X-ray absorption spectroscopies was employed to reveal the persistent luminescence mechanism in the CaAl2O4:Eu2+, Ce3+ materials. As a result, the forbidden energy gap from the valence band to the conduction band in the CaAl2O4 host could be established but no excitonic

ARTICLE IN PRESS S16

J. H¨ ols¨ a et al. / Radiation Physics and Chemistry 78 (2009) S11–S16

features just below the conduction band could be observed. The experimental band gap value of CaAl2O4 determined from the VUV excitation of Eu2+ could be reproduced by the density functional theory calculations with a very good agreement. The energy level scheme of Eu2+ in the band gap could be established from the intensities of the 4f7-4f65d1 transitions. Based on the present results, a credible mechanism for the persistent luminescence in CaAl2O4:Eu2+, Ce3+ was exposed but more experimental and theoretical work is needed to ensure the correctness of the mechanism. The present results of the use of the DFT calculations are very encouraging, and the next step would be to treat the energy levels of the dopants and co-dopants, Eu2+ and R3+, respectively, in the band structure of CaAl2O4. Eventually, the defect energy levels obtained with the thermoluminescence and VUV spectroscopies should be calculated with the DFT method, too.

Acknowledgements Financial support is acknowledged from the Turku University Foundation, Jenny and Antti Wihuri Foundation (Finland) and the Academy of Finland (Contract #117057/2006). The synchrotron radiation study (HASYLAB, Germany) was supported by the European Community-Research Infrastructure Action under the FP6 Structuring the European Research Area Programme, RII3-CT2004-506008 (IA-SFS). The help of Dr. Gregory Stryganyuk, Dr. Edmund Welter and Dr. Dariusz Zajac (HASYLAB) during the synchrotron measurements as well as that of Mrs. L. Pihlgren in measuring the UV–VUV spectra is acknowledged, too. The theoretical DFT study was supported by the research mobility agreements (112816/2006/JH and 116142/2006/JH, 123976/2007/ TL) between the Academy of Finland and the Academy of Sciences of the Czech Republic. The DFT calculations were carried out using the supercomputing resources of the CSC IT Center for Science (Espoo, Finland). References ¨ a, ¨ J., Lastusaari, M., Niittykoski, J., Suchocki, A., 2004a. Aitasalo, T., Durygin, A., Hols Low temperature thermoluminescence properties of Eu2+ and R3+ doped CaAl2O4. J. Alloys Compd. 380 (1–2), 4–8. ¨ a, ¨ J., Jungner, H., Krupa, J.-C., Lastusaari, M., Legendziewicz, J., Aitasalo, T., Hols Niittykoski, J., 2004b. Effect of temperature on the luminescence processes of SrAl2O4:Eu2+. Radia. Meas. 38 (4–6), 727–730. ¨ a, ¨ J., Jungner, H., Lastusaari, M., Niittykoski, J., 2006. ThermoAitasalo, T., Hols luminescence study of persistent luminescence materials: Eu2+- and R3+ doped calcium aluminates, CaAl2O4:Eu2+,R3+. J. Phys. Chem. B 110 (10), 4589–4598.

Blaha, P., Schwarz, K., Madsen, G.K.H., Kvasnicka, D., Luitz, J., 2001. In: Schwarz, K. (Ed.), WIEN2k, An Augmented Plane Wave+Local Orbitals Program for Calculating Crystal Properties. Vienna University of Technology, Austria. Chen, Y., Liu, B., Kirm, M., Qi, Z., Shi, C., True, M., Vielhauer, S., Zimmerer, G., 2006. Luminescent properties of blue-emitting long afterglow phosphors Sr2xCaxMgSi2O7:Eu2+, Dy3+ (x ¼ 0, 1). J. Lumin. 118 (1), 70–78. Clabau, F., Rocquefelte, X., Jobic, S., Deniard, P., Whangbo, M.-H., Garcia, A., Le Mercier, T., 2005. Mechanism of phosphorescence appropriate for the longlasting phosphors Eu2+-doped SrAl2O4 with codopants Dy3+ and B3+. Chem. Mater. 17 (15), 3904–3912. Dorenbos, P., 2003. Relation between Eu2+; and Ce3+ f 2 d-transition energies in inorganic compounds. J. Phys.: Condens. Matter 15 (27), 4797–4807. Dorenbos, P., 2005a. Mechanism of persistent luminescence in Eu2+ and Dy3+ codoped aluminate and silicate compounds. J. Electrochem. Soc. 152 (7), H107–H110. Dorenbos, P., 2005b. Mechanism of persistent luminescence in Sr2MgSi2O7:Eu2+; Dy3+. Phys. Status Solidi b 242 (1), R7–R9. Dorenbos, P., 2007. Absolute location of lanthanide energy levels and the performance of phosphors. J. Lumin. 122–123, 315–317. Dorenbos, P., 2008. Energy of the Eu2+ 5d state relative to the conduction band in compounds. J. Lumin. 128 (4), 578–582. ¨ ¨ Horkner, W., Muller-Buschbaum, H.K., 1976. Zur kristallstruktur von CaAl2O4. J. Inorg. Nucl. Chem. 38, 983–984. Jia, D., Yen, W.M., 2003a. Enhanced V3+ K center afterglow in MgAl2O4 by doping with Ce3+. J. Lumin. 101 (1–2), 115–121. Jia, D., Yen, W.M., 2003b. Trapping mechanism associated with electron delocalization and tunneling of CaAl2O4:Ce3+, a persistent phosphor. J. Electrochem. Soc. 150 (3), H61–H65. Jia, D., Zhu, J., Wu, B., 2000. Improvement of persistent phosphorescence of Ca0.9Sr0.1S:Bi3+ by codoping Tm3+. J. Lumin. 91 (1–2), 59–65. Kodama, N., Sasaki, N., Yamaga, M., Masui, Y., 2002. Mechanism of long-lasting phosphorescence of Eu2+ in melilite. Radiat. Eff. Defects Solids 157 (6–12), 915–919. Laasch, W., 2009a. Beamline: I (SUPERLUMI), HASYLAB web pages. /http:// hasylab.desy.de/facilities/doris_iii/beamlines/i_superlumiS. Laasch, W., 2009b. Beamline: E4 (EXAFS II), HASYLAB web pages. /http://hasylab. desy.de/facilities/doris_iii/beamlines/e4_exafs_iiS. Lin, Y., Tang, Z., Zhang, Z., Nan, C., 2003. Influence of co-doping different rare earth ions on the luminescence of CaAl2O4-based phosphors. J. Eur. Ceram. Soc. 23 (1), 175–178. Matsuzawa, T., Aoki, Y., Takeuchi, N., Murayama, Y., 1996. New long phosphorescent phosphor with high brightness, SrAl2O4:Eu2+, Dy3+. J. Electrochem. Soc. 143 (8), 2670–2673. Newton Harvey, E., 1957. A History of Luminescence: From the Earliest Times Until 1900. Amer. Phil. Soc. Philadelphia PA, U.S.A., pp. 305–365. Pedrini, C., Rogemond, F., McClure, D.S., 1986. Photoionization thresholds of rareearth impurity ions; Eu2+: CaF2, Ce3+: YAG, and Sm2+: CaF2. J. Appl. Phys. 59 (4), 1196–1201. Shannon, R.D., 1976. Revised effective ionic-radii and systematic studies of interatomic distances in halides and chalcogenides. Acta Crystallogr. A 32, 751–767. Thiel, C.W., Sun, Y., Cone, R.L., 2002. Progress in relating rare-earth ion 4f and 5d energy levels to host bands in optical materials for hole burning, quantum information and phosphors. J. Mod. Opt. 49 (14–15), 2399–2411. ¨ a, ¨ J., Zych, E., 2008. Thermoluminescence and Trojan-Piegza, J., Niittykoski, J., Hols kinetics of persistent luminescence of vacuum-sintered Tb3+-doped and Tb3+,Ca2+-codoped Lu2O3 materials. Chem. Mater. 20 (6), 2252–2261. van Haecke, J., 2007. Europium doped sulfide phosphors for photo- and electroluminescence. PhD Thesis, Universiteit Gent, Gent, Belgium. Zhang, Z., Wang, Y., 2007. UV–VUV excitation luminescence properties of Eu2+ doped Ba2MSi2O7 (M ¼ Mg, Zn). J. Electrochem. Soc. 154 (2), J62–J64.