Photoluminescence and afterglow of Dy3+ doped CaAl2O4 derived via sol–gel combustion

With doping concentration varying from 0.1 to 5.0 mol%, a series of Dy3+ doped calcium aluminate (CaAl2O4:Dy3+) phosphors were synthesized via a sol–gel combustion technique. The phase, morphology, photoluminescence (PL), afterglow, and thermoluminescence (TL) glow curves of CaAl2O4:Dy3+ were investigated by means of X-ray diffractometry, scanning electron microscopy, transmission electron microscopy, PL spectroscopy, afterglow spectroscopy, and TL dosimetry, respectively. It is found that: (i) oxygen vacancies and Dy3+ work as two independent sets of luminescence centers of PL for CaAl2O4:Dy3+; (ii) Dy3+ works as the luminescence center of afterglow for CaAl2O4:Dy3+; (iii) the afterglow of CaAl2O4:Dy3+ lasts for about 115 min at the optimal doping concentration of around 0.8 mol%; and (iv) multiple traps, which are sensitive to doping concentration, are present in CaAl2O4:Dy3+. The PL and afterglow mechanisms of CaAl2O4:Dy3+ are discussed to reveal the processes of charged carrier excitation, migration, trapping, detrapping, and radiative recombination in CaAl2O4:Dy3+.


Introduction
Rare-earth doped calcium aluminate phosphors have been intensively studied due to their excellent aerglow properties, [1][2][3][4] among which Dy 3+ doped calcium aluminate (CaAl 2 O 4 :Dy 3+ ) is a well-known inorganic phosphor which emits white photoluminescence (PL) under ultraviolet excitation. 5 Apart from its white PL, CaAl 2 O 4 :Dy 3+ exhibits a white aerglow even aer the removal of ultraviolet excitation. 6 As reported by Liu et al. in 2005, Dy 3+ works as the luminescence center of aerglow for CaAl 2 O 4 :Dy 3+ and the white aerglow of solid state reaction derived CaAl 2 O 4 :Dy 3+ lasts 32 min at the optimal doping concentration of 2 at%. 6 For aerglow materials, the processes of charged carrier excitation, migration, trapping, detrapping, and radiative recombination are critically important to understand their aerglow properties. [7][8][9] For example, long aerglow can be achieved at room temperature only when traps have an appropriate activation energy somewhere around 0.65 eV whereas shallow traps (E # 0.4 eV) and deep traps (E > 2 eV) are not favorable because they can be emptied either easily or with difficulty at room temperature. 7 Up to date, only a single report exists on the aerglow of CaAl 2 O 4 :Dy 3+ , leaving the processes of charged carrier excitation, migration, trapping and detrapping not fully revealed. The lack of such knowledge hampers the comprehensive understanding of the aerglow mechanism of CaAl 2 O 4 :Dy 3+ .
With respect to solid state reaction derived CaAl 2 O 4 :Dy 3+ , sol-gel combustion derived CaAl 2 O 4 :Dy 3+ is abundant in oxygen and calcium vacancies. In CaAl 2 O 4 :Dy 3+ , oxygen vacancies are potential electron traps because they are positively charged whereas calcium vacancies are potential hole traps because they are negatively charged. 8,10 Therefore, sol-gel combustion derived CaAl 2 O 4 :Dy 3+ phosphors are suitable for exploring the aerglow mechanisms of CaAl 2 O 4 :Dy 3+ . In this paper, we report the PL and aerglow properties of sol-gel combustion derived CaAl 2 O 4 :Dy 3+ with the doping concentrations varying from 0.1 to 5.0 mol%. A picture on the PL and aerglow mechanisms of CaAl 2 O 4 :Dy 3+ is given to reveal the processes of charged carriers' excitation, migration, trapping, detrapping, and recombination in CaAl 2 O 4 :Dy 3+ .

Sol-gel combustion synthesis of CaAl 2 O 4 :Dy 3+
CaAl 2 O 4 :Dy 3+ phosphors with the doping concentrations varying from 0.1 to 5.0 mol% were synthesized via the sol-gel combustion with urea as fuel. [11][12][13] All reagents were in analytical grade and provided by Sinopharm Chemical Reagents Co., Ltd (Shanghai, China). The purity of dysprosium oxide (Dy 2 O 3 ) was 99.99%. Under stirring with a magnetic bar, Ca(NO 3 ) 2 $4H 2 O (0.2 mol), Al(NO 3 ) 3 $9H 2 O (0.4 mol), urea (6.0 mol), H 3 BO 3 (0.02 mol) and stoichiometric amount of Dy 2 O 3 were dissolved in deionized water (600 mL) to form a transparent solution. A homogeneous solution was obtained aer the mixture was stirred vigorously for 60 min in a glass beaker. Aer having been aged at room temperature for two weeks, the solution was ready for the sol-gel combustion. Urea and boric acid functioned as the fuel and ux, respectively. Alumina crucibles, each with the volume capacity of 50 mL, were employed as the reaction containers. Half lled with the aged solution, the solutioncontaining alumina crucible was transferred into an air-lled box furnace for self-propagating combustion. The temperature in the furnace was preset at 780°C. Aer the solutioncontaining crucible was transferred into the furnace, the temperature in the furnace was gradually dropped to about 706°C . Heated at such high temperatures, water in the solution got boiling and the starting materials (fuels and metal nitrates) in the crucible were partially decomposed until the organic fuels were automatically ignited to initiate the exothermic reactions. The sol-gel combustion yielded voluminous gases and bright ames. In this work, the synthesis was initiated by pointheating of a small part of the mixture in the crucible, and the ignition was started at around 750°C. Once started, a wave of exothermic reactions swept through the remaining material in the crucible. This combustion synthesis lasted for about 40 s, which was exceptionally fast when compared to hightemperature solid state reactions. During the sol-gel combustion, a large amount of energy was released from the exothermic reactions, which in turn raised the temperature in the furnace up to 830°C. Measured with an infrared thermometer, the temperature in the ame was up to 1300°C. Aer the re was extinguished, the crucible was taken out of the furnace immediately. The total holding time of the crucible in the furnace was about 4 min. White powders were resulted aer the sol-gel combustion. According to molar ratio of Dy 3+ 3+ phosphors were recorded on X-ray diffractometer (D/max 2500 PC, Rigaku Corporation, Akishima, Japan) using Cu Ka radiation (l = 0.15405 nm). The scanning electron microscope (SEM) (model S-4800, Hitachi, Tokyo, Japan) was employed to analyze the morphology of the synthesized products. The SEM was coupled with a silicon dried detector as the X-ray analyzer to record the energy dispersive X-ray (EDX) spectrum of the synthesized products. The micrographs of CaAl 2 O 4 :Dy 3+ nanocrystals were recorded on a transmission electron microscope (TEM) (model JEOL JEM-2100, Japan Electronics Corp, Akishima, Japan). Samples for TEM analysis were prepared by suspending the particles in ethanol under the excitation of ultrasonication and then drying a drop of the suspension on a carbon-coated copper grid.

PL excitation and emission spectra of CaAl 2 O 4 :Dy 3+
The PL excitation spectrum of CaAl 2 O 4 :Dy 3+ was measured with the uorescence spectrometer F-7000 (Hitachi, Japan). The spectrophotometer (Tianjin Gangdong Ltd., Tianjin, China) was used to acquire the steady-state PL spectra of CaAl 2 O 4 :Dy 3+ . The excitation source of the PL spectrum was provided by a heliumcadmium laser (Kimmon Electric Co. Ltd., Tokyo, Japan). The emission wavelength and the output power of the laser radiation were 325 nm and 13 mW, respectively.

Aerglow spectra and thermoluminescence (TL) glow curves of CaAl 2 O 4 :Dy 3+
The aerglow spectra of CaAl 2 O 4 :Dy 3+ were recorded with the PL spectrophotometer (Tianjin Gangdong Ltd, Tianjin, China) immediately aer the ultraviolet irradiation of a high-pressure mercury lamp was blocked off. The output power of the highpressure mercury lamp was 175 W. The irradiation duration of the high-pressure lamp irradiation was 3 min. Aerglow decay curve was taken by focusing the aerglows into the entrance slit of the spectrometer. The TL glow curves of CaAl 2 -O 4 :Dy 3+ were measured on a TL meter constructed according to the scheme given by Yamashita et al. 14 The phosphor was placed on an electrically heated plate, the temperature of the plate was controlled with a program in computer. As the temperature of the plate was raised linearly with time, the light output of the phosphor was recorded using a photomultiplier in a photoncounting mode. A bialkali photomultiplier tube (model H10425, Hamamatsu, Japan) was used in the TL meter for the luminescence detection, it covers the 350-600 nm wavelength range. Prior to the TL measurements, CaAl 2 O 4 :Dy 3+ phosphors were exposed to the 254 nm irradiation of a low-pressure mercury lamp for 5 min. The output power of the lowpressure mercury lamp was 32 W. The TL signals of CaAl 2 O 4 :-Dy 3+ phosphors were recorded when the CaAl 2 O 4 :Dy 3+ phosphors were heated from 10 to 200°C at a rate of 2°C s −1 .  Fig. 1 are evenly spaced aer having been shied upwards one by one, so the scale of the intensity of powder XRD in Fig. 1 is arbitrary units. As can be seen in Fig. 1 040) and (226) of monoclinic CaAl 2 O 4 , respectively. Among these peaks, the strongest one is located at 30.08°and indexed as (123). Actually, this peak is contributed jointly by two crystallographic planes (220) and (123) of monoclinic CaAl 2 O 4 , which are located at 2q = 30.11 and 30.05°, respectively. Being closely packed and nearly equal in diffraction intensity, the two peaks are not distinguishable in the XRD proles. This is the reason why some researchers indexed the peak at around 30.8°as (220). 7,8 Obviously, the diffraction peaks in Fig. 1 (Fig. S1 †). The absence of a secondary phase in the XRD proles indicates that Dy 3+ ions are successfully incorporated into CaAl 2 O 4 without introducing new phase in the crystal structure when doping concentration is lower than 5.0 mol%. For sol-gel combustion, nitrate forms of starting materials are highly necessary. Instead of Dy 2 O 3 , for example, Dy(NO 3 ) 3 is favored as the starting material. It is known that Dy 2 O 3 can be dissolved into dilute HNO 3 solution to form Dy (NO 3 ) 3 . 9 In this work, Dy 2 O 3 is used as a starting material to provide the source of Dy 3+ because a tiny amount of Dy 2 O 3 can be dissolved into the aqueous solution of Al(NO 3 ) 3 and urea aer stirring with a magnetic bar for a couple of hours at a temperature higher than 20°C. Since the synthesis process occurs at high temperatures generated by the self-propagating combustion, this method is ideally suited for the production of refractory materials including ceramics CaAl 2 O 4 .

Results and discussions
Monoclinic CaAl 2 O 4 is known to have the stuffed tridymite structure with the space group P2 1 /n and Z = 12.  17 In monoclinic CaAl 2 O 4 , a threedimensional network is formed by corner-sharing [AlO 4 ] tetrahedra, the Ca 2+ ions are located along the channels in b-direction. Fig. S2(b) † illustrates the schematic view of the monoclinic CaAl 2 O 4 along the b-direction. Such a crystal structure makes it easy for Dy 3+ ions to substitute the Ca 2+ sites in CaAl 2 O 4 . The effective ionic radii of cations depend on their coordination numbers (CN). In the case of site occupancy, the effective ionic radii of Ca 2+ are 0.100 nm (CN = 6) and 0.118 nm (CN = 9), the effective ionic radii of Dy 3+ are 0.091 nm (CN = 6) and 0.108 nm (CN = 9). 18 Due to the comparable ionic sizes, Dy 3+ ions tend to replace the Ca 2+ ions in the lattice of CaAl 2 O 4 . Fig. 2(a) displays the SEM micrograph of sol-gel combustion derived CaAl 2 O 4 :Dy 3+ (5.0 mol%). As shown in Fig. 2(a), the solgel combustion derived CaAl 2 O 4 :Dy 3+ phosphors are in the form of irregular aggregates with their dimensions up to 20 mm. Solgel combustion derived aluminate nanocrystals are prone to forming large aggregates. Micrometer-sized pores are randomly distributed in the aggregates due to the production of a large amount of gases during the reaction. 3,8,[11][12][13]19,20 Fig. 2(b) shows the SEM micrographs of CaAl 2 O 4 :Dy 3+ to display the channels formed in the aggregates, each scale bar represents 2 mm. It is evident that channels with different diameters are present in the aggregate. Fig. 2(c) shows the TEM micrograph of the CaAl 2 O 4 :Dy 3+ . Obviously, each aggregate consists of a large number of nanocrystals. Previous work has evidenced that one aggregate consists of a number of nanocrystals whose sizes vary from 10 nm to 80 nm. 4,16,17,21 As a contrast, solid state reaction derived CaAl 2 O 4 :Dy 3+ exhibits quite different morphology.  . 6 Due to the features of solid state reaction, the solid state reaction derived CaAl 2 O 4 :Dy 3+ phosphors are different from the sol-gel combustion derived ones in the aspects of population density of intrinsic defects. Especially in the sol-gel combustion method, addition of H 3 BO 3 lowers reaction temperature, accelerates a diffusion of raw materials, and eventually inuences the aerglow properties. Detailed discussions are given by Takeuchi       xed at 574 nm for the PL excitation measurement. As can be seen in Fig. 5 20 The characteristic emissions of Dy 3+ activator are represented by two narrowband emissions peaking at around 482 and 574 nm, respectively, which are due to the 4 F 9/2 / 6 H 15/2 and the 4 F 9/2 / 6 H 13/2 transitions of Dy 3+ activator. [2][3][4]6,23 As documented in the literature, the transition 4 F 9/2 / 6 H 15/2 (DL = 2 and DJ = 3) is magnetic dipole allowed whereas the transition 4 F 9/2 / 6 H 13/2 is identied as a hypersensitive electric dipole transition (DL = 2 and DJ = 2) of Dy 3+ . The reason why the characteristic emissions of Dy 3+ activator are hardly discernible in the PL spectrum of CaAl 2 O 4 :Dy 3+ (0.1 mol%) rests on the fact that the characteristic emissions of Dy 3+ are too weak when compared to the strong emissions from the intrinsic defects in the host. As the doping concentration is increased to 0.2 and 0.4 mol%, the narrowband emissions of Dy 3+ at 574 nm become discernible in the PL spectra of CaAl 2 O 4 :Dy 3+ (0.2 and 0.4 mol%), as shown by the PL spectra (b) and (c). As the doping concentration is elevated further to 0.6 mol%, the narrowband emissions of Dy 3+ can be identied clearly in the PL spectrum of CaAl 2 O 4 :Dy 3+ , as shown by the PL spectrum (d). It is apparent in Fig. 6 that the intensity of the characteristic emissions of Dy 3+ increases monotonically with the doping concentration. Consequently, the PL spectrum of CaAl 2 O 4 :Dy 3+ indicates that the intrinsic defects in CaAl 2 O 4 and the doping species Dy 3+ act as two independent sets of luminescence center of PL in CaAl 2 O 4 :Dy 3+ . The insight reason for the change in the PL intensity with doping of Dy 3+ is due to the changes in the population density of the luminescence center of PL. The substitution of Dy 3+ for Ca 2+ in the lattice of CaAl 2 O 4 promotes the production of two kinds of luminescence center of PL, as shown in eqn (1). The rst kind of luminescence center of PL is the Dy 3+ in Ca 2+ site while the second kind of luminescence center of PL is the oxygen vacancy.
As the doping concentration of Dy 3+ increases, the population densities of the two kinds of luminescence center of PL are increased, which in turn lead to the enhancement in the PL intensities. It is noted that the PL intensity of each spectrum in Fig. 6 is normalized at 400 nm.
To check the inuences of excitation wavelength on the emissions of CaAl 2 O4:Dy 3+ , we measured the PL spectra of CaAl 2 O 4 :Dy 3+ (5 mol%) under different excitation wavelengths. The emission spectra of CaAl 2 O 4 :Dy 3+ under the excitations of 326 nm and 350 nm are shown in Fig. S3. † As can be seen in Fig.  S3, † the structures of the two PL spectra are nearly identical to each other regardless the variation in the excitation wavelength. However, the emission intensity of CaAl 2 O 4 :Dy 3+ (5 mol%) is apparently sensitive to the excitation wavelength. For example, the characteristic emissions of Dy 3+ are stronger under the excitation of 350 nm than those under the excitation of 326 nm. One of the reasons rests on the fact that Dy 3+ exhibits stronger absorption at 350 nm than at 326 nm, as endorsed by Fig. 5.
Due to the combined contributions of defect emissions from the host and the characteristic emissions of the dopant, the color coordinates of the emissions of CaAl 2 O 4 :Dy 3+ varies with the doping concentration. Color coordinates of luminescent materials, which can be calculated from their PL spectral data, are important parameters to quantitatively describe the emission color for luminescent materials. 32,33 The CIE chromaticity coordinates of CaAl 2 O 4 :Dy 3+ are given in Table S1 † for different doping concentrations. As shown in Table S1 The assignment of the broadband emissions in Fig. 6 to intrinsic defects in CaAl 2 O 4 gains further support from the PL spectrum of solid state reaction derived CaAl 2 O 4 :Dy 3+ . The solid state reaction route involves chemical decomposition and reactions at much high temperatures (oen from 1000 to 1500°C ) to produce a new solid composition. At such high temperatures, oxygen atoms migrate into the crystal lattice to repair the defects with the result of decreased population density of oxygen vacancies in CaAl 2 O 4 . Therefore, the broadband emissions should be weakened if the density of oxygen vacancies in CaAl 2 O 4 is reduced at high temperature. Fig. 7 depicts the normalized PL spectra of solid state reaction derived The emission spectrum of a high-pressure mercury lamp contains a pronounced spectral line at 365.4 nm, which is close to an absorption peak of CaAl 2 O 4 :Dy 3+ . Aer exposure to the illumination of a high-pressure mercury lamp (175 W) for 3 min, CaAl 2 O 4 :Dy 3+ exhibits intense aerglow aer the ultraviolet excitation is blocked off. Fig. S6 † depicts the aerglow photos of CaAl 2 O 4 :Dy 3+ (0.8 mol%) taken at different times aer the irradiation of the high-pressure mercury lamp is blocked off. As shown in Fig. S6, † the white aerglow of CaAl 2 O 4 :Dy 3+ (0.8 mol%) can last more than 60 min. The initial luminance of the aerglow is found to depend on the doping concentration of Dy 3+ . Fig. 8(a) shows the plot of the integrated aerglow intensity of sol-gel combustion derived CaAl 2 O 4 :Dy 3+ versus the doping concentration of Dy 3+ . As can be seen in Fig. 8(a), the optimal doping concentration is around 0.8 mol%. It is noted that the aerglow gets quenched when the doping concentration is high (i.e., 5 mol%). Just like the case of PL quenching at high doping concentration, the non-radiative interaction between dopants is one of the reason of the concentration induced aerglow quenching. 34 However, the traps generated by the dopant in the lattice of CaAl 2 O 4 should be the key factor to be responsible for the aerglow quenching. As described in eqn (1), doping CaAl 2 O 4 with Dy 3+ results in oxygen vacancies. These positively charged oxygen vacancies can act as electron traps. The increase in the population density as well as the change in the trap depth of these the positively charged oxygen vacancies generate signicant effects on the aerglow duration of the phosphor. For example, the aerglow duration is very short when the trap depth is shallow (E < 0.6 eV), 8,15,16,35 and no aerglow can be observed when the trap depth is too deep (E > 2.0 eV). To observe aerglow at room temperature, the traps should have an appropriate activation energy somewhere between these two extremes, a trap depth around 0.65 eV is considered to be optimal. 7 Fig. 8(b) depicts the aerglow spectrum of the sol-gel combustion derived CaAl 2 O 4 :Dy 3+ at the doping concentration of 0.8 mol%. As can be seen in Fig. 8(b), the aerglow spectrum consists of two narrow emission bands of Dy 3+ , which is distinctly different from the broadband aerglow of CaAl 2 O 4 :Eu 2+ . 1,21,36,37 Thus Fig. 8(b)    comparison, the aerglow spectrum of solid state reaction derived CaAl 2 O 4 :Dy 3+ (0.8 mol%) is given in the ESI as Fig. S8. †

Aerglow decay prole of CaAl 2 O 4 :Dy 3+
The duration of aerglow is a straightforward and standardized parameter to evaluate the properties of an aerglow material, whereby 0.32 mcd m −2 is oen used as threshold for dening the duration of an aerglow. Fig. 9 depicts the aerglow decay prole of sol-gel combustion derived CaAl 2 O 4 :Dy 3+ (0.8 mol%). The phosphor was exposed to illumination of a high-pressure mercury lamp irradiation (175 W) for 3 min before the measurement of aerglow decay curve. As shown by the raw data in Fig. 9, the aerglow duration of CaAl 2 O 4 :Dy 3+ is determined to be about 115 min. It is found that the decay curves in Fig. 9 could be best tted to the tri-exponential function according to eqn (2) where I(t) is the aerglow intensity at time t aer blocking the laser excitation, I i is the prefactor of the exponential component whose lifetime decay constant is s i (i = 1-3). The red solid line in Fig. 9 represents the t of the experimental signals by eqn (2). The tting parameters are tabulated in the gure. Clearly, the CaAl 2 O 4 :Dy 3+ has three largely different decay components with constants of s 1 = 2.51 min, s 2 = 11.31 min and s 3 = 89.29 min.
We can see that s 2 and s 3 are much longer than s 1 , suggesting the presence of deeper traps in the phosphor. The luminance reading of the rst data point in Fig. 9 is 8.3 mcd m −2 at 3.67 min. In order to derive the luminance at t = 0, it is necessary to extrapolate the data, yielding the value of the luminance to be 22.35 mcd m −2 at t = 0. It is worth of noting that the luminance of the phosphor at the initial moment is different from the luminance of the rst component at the initial moment (i.e., I 1 = 16.56 mcd m −2 ). Actually, the luminance of the phosphor at the initial moment is the summation of the luminance of the three components at the initial moment, that is, the sum of I 1 , I 2 and I 3 . Moreover, the aerglow duration of the phosphor is different from the longest lifetime decay constant s 3 (i.e., 89.29 min). According to eqn (1), the aerglow duration of the phosphor is dened by the variable t when I(t) reaches the threshold 0.32 mcd m −2 . The luminance decreases to 0.3205 mcd m −2 when t = 111 min. Consequently, the aerglow duration derived from eqn (2) (about 111 min) is very close the aerglow duration derived from raw data (about 115 min). Being distinctly different from PL decay, the aerglow phenomenon is a particular case of thermostimulated luminescence and is a defect dependent phenomenon.

Possible PL and aerglow mechanisms of CaAl 2 O 4 :Dy 3+
CaAl 2 O 4 is an insulator with a bandgap of around 6.7 eV. 39,40 Oxygen and calcium vacancies are intrinsic defects in CaAl 2 -O 4 :Dy 3+ . On one hand, these intrinsic vacancies act as luminescence center of PL with the result of a broadband PL spectrum peaking at about 400 nm. 15,16 On the other hand, these intrinsic vacancies work as traps for charged carriers. For example, oxygen vacancies in CaAl 2 O 4 are proposed to work as electron traps because they are positively charged whereas calcium vacancies are potential hole traps because they are positively charged. 10,39 Dopant Dy 3+ belongs to extrinsic defect in CaAl 2 O 4 :Dy 3+ . Aer Dy 3+ ions are incorporated into the host, a series of defect energy levels are introduced into the bandgap of CaAl 2 O 4 . The lowest energy level of the excited state of Dy 3+ is known as 4 F 9/2 while the energy levels of the ground state of Dy 3+ are denoted as 6 H J (J = 15/2-5/2). As evidenced by the characteristic emissions in Fig. 6, 7 and S3, † dopant Dy 3+ ions act as luminescent center of PL in CaAl 2 O 4 :Dy 3+ to yield narrowband emissions peaking at 482 and 574 nm, respectively. Furthermore, this extrinsic defect can work as electron trap because it is positively charged.  Aer non-radiative relaxations, the hot electrons are captured by either the oxygen vacancies (process ②) or the electron traps (process ③). The subsequent radiative recombination of electrons captured at oxygen vacancies with holes in the valence band of CaAl 2 O 4 yields the broadband PL peaking at about 400 nm (process ④). Since the transition 6 H 15/2 / 6 P 3/2 of Dy 3+ (326 nm) matches well with the 325 nm excitation of the laser, some incident photons are absorbed by Dy 3+ activator in CaAl 2 O 4 (process ⑤). It is noted that several energy levels are located between the excited state 6 P 3/2 and the lowest excited state 6 H 15/2 , among which include 4 I 13/2 , 4 P 5/2 and 6 P 7/2 . With the assistance of phonons, a portion of hot electrons at the excited state 6 P 3/2 are relaxed to the lowest excited state 4 F 9/2 . Then electron transitions 4 F 9/2 / 6 H 15/2 and 4 F 9/2 / 6 H 13/2 of Dy 3+ lead to the characteristic emissions of Dy 3+ (process ⑥). Apart from the radiation recombination, some hot electrons at excited states of Dy 3+ can be captured by the electron traps via non-radiative relaxations (process ⑦). In the light of the proposed mechanisms in Fig. 10 Once the ultraviolet excitation is ceased, processes ①-⑦ are stopped immediately. Under thermal activation, electrons can be released from the electron traps via processes ⑧ and ⑨. Subsequently, aerglow with characteristic emissions of Dy 3+ is resulted when the electrons released from the excited states of Dy 3+ recombine radiatively with holes via process ⑥. Theoretically speaking, aerglow with broadband emissions peaking at about 400 nm should be observed when thermally detrapped electrons recombine radiatively with holes via process ④. In practice, such a broadband aerglow is negligible because its intensity is several orders of magnitude weaker than that of Dy 3+ related aerglows. Therefore, the aerglow spectrum of the CaAl 2 O 4 :Dy 3+ consists of two narrowband emissions peaking at 482 and 574 nm, respectively.
In the light of the PL and aerglow mechanisms of CaAl 2 -O 4 :Dy 3+ in Fig. 10, electrons in the electron traps contribute to the PL under photoexcitation, too. Upon the photoexcitation, some electrons in the electron traps are detrapped via processes ⑧ and ⑨. Broadband emissions peaking at around 400 nm can be expected when the detrapped electrons recombine radiatively with holes via the radiative process ④, and narrowband emissions of Dy 3+ can be resulted when the detrapped electrons recombine radiatively with holes via the radiative process ⑥. In most cases, however, such contributions are negligible because they are many times weaker than the emissions resulted directly from the ultraviolet photoexcitation. Only when the aerglow is sufficiently strong, such contributions to the PL are discernible. For example, the PL spectrum of Dy 3+ doped SrAl 2 O 4 is the result of superposition of the broadband emissions of the host peaking at about 400 nm and another broadband emissions peaking at about 520 nm. 12,13 3.9 TL glow curves of CaAl 2 O 4 :Dy 3+ TL is an important tool to determine the activation energies (i.e., trap depths) of trapping levels in crystals. 37,39,41,42 Fig. 11 represents the TL glow curves of sol-gel combustion derived CaAl 2 O 4 :Dy 3+ with different doping concentrations. Each CaAl 2 O 4 :Dy 3+ phosphor was exposed to ultraviolet light of 254 nm for 5 min before the TL measurements. The temperature rising rate was 2 K min −1 . It can be seen that each TL glow curve in Fig. 11 exhibits an extremely broad and asymmetric prole, suggesting the presence of multiple trap levels in CaAl 2 O 4 :Dy 3+ . 43 Moreover, both the prole and the peak temperature of the TL glow curve are sensitive to the doping concentration: (i) the TL glow curve consists of a primary peak at around 350 K and a secondary peak at about 310 K when the doping concentration increases from 0.1 to 0.4 mol%; (ii) the TL glow curve has only one peak, which gradually shis to higher temperature (from 350 to 375 K) as the doping concentration increases from 0.6 to 0.8 mol%; and (iii) the TL glow curve has only one peak, which gradually shis to lower temperature (from 375 to 350 K) as the doping concentration increases further from 1.0 to 5.0 mol%. The vertical dash line in Fig. 11 marks the position of 350 K. With the vertical line as a guideline, the evolution of the peak temperature with the doping concentration can be identied clearly. Interestingly, the peak temperature of the TL glow curve of CaAl 2 O 4 :Dy 3+ (0.8 mol%) is the highest (around 375 K) among the 10 phosphors under test, which coincides with the best aerglow performance of CaAl 2 -O 4 :Dy 3+ at optimal doping concentration of 0.8 mol%.
To understand the TL behavior of CaAl 2 O 4 :Dy 3+ , it is necessary to deconvolute each TL glow curve and evaluate the trapping parameters. The TL curve based on the general order function is given by the following equation: where I is the TL intensity at temperature T, s is the preexponential factor with the unit of s −1 , n 0 is the concentration of trapped charges at time t = 0, E is the trap depth, k is the Boltzmann constant, b is the order of kinetics, and b is the heating rate. In order to determine the kinetic parameters of the multiple traps, computerized glow curve deconvolution of the TL glow curves is carried out with general order kinetics using a computer program given by Chung et al. 44 Fig. 12 depicts the computerized glow curve deconvolution of the TL glow curve of CaAl 2 O 4 :Dy 3+ at the doping concentration of 0.8 mol%. It is found that this TL glow curve can be described satisfactorily by using the general order kinetics to model 5 traps in CaAl 2 O 4 :-Dy 3+ . The gure-of-merit (FOM) of the deconvolution is 2.071%. The kinetic parameters and the electron lifetime at room temperature (s 300 ) are summarized in Table 1 for each trap in the CaAl 2 O 4 :Dy 3+ (0.8 mol%). As can be seen in Table 1 In addition to the TL glow curve shown in Fig. 12, we have also deconvoluted the TL glow curves of CaAl 2 O 4 :Dy 3+ with the doping concentrations of 0.1, 0.2, 0.4, 0.6, 1.0, 2.0, 3.0, 4.0 and 5.0 mol%, respectively. For the sake of brevity, the computerized glow curve deconvolutions of the TL glow curves of CaAl 2 O 4 :Dy 3+ are shown in ESI as Fig. S9-S17. † The kinetic parameters and the electron lifetime at room temperature are listed in Table S2 † for each trap in the CaAl 2 O 4 :Dy 3+ . Apparently, the parameters of electron traps in CaAl 2 O 4 :Dy 3+ can be effectively tuned via the control of doping concentration, which in turn can be exploited to modify the brightness and duration of the aerglow of CaAl 2 O 4 :Dy 3+ .   Table 1 Kinetic parameters of the computerized glow curve deconvolution of the TL glow curve of CaAl 2 O 4 :Dy 3+ (0.8 mol%). T m represents the peak temperature, E is the trap-depth, s is the frequency factor, b is the order of kinetics, and s 300 is the room temperature electron lifetime in the trap

Conflicts of interest
There are no conicts to declare.