The influence of defects formed by Ca excess and thermal post-treatments on the persistent luminescence of CaTiO 3 :Pr

: Red emitting CaTiO 3 :Pr phosphors with a nominal composition of Ca 0.998+x Pr 0.002 TiO 3+ δ (0.02 ≤ x ≤ 0.04) were prepared by solid state reactions with different thermal post treatments and characterized by X-ray diffraction, transmission electron microscopy and photoluminescence. The Ca excess exhibited complete solubility up to 4% in the samples treated at 1400 °C but segregation in the form of Ruddlesden-Popper phases (Ca 3 Ti 2 O 7 - Ca 4 Ti 3 O 10 ) was observed in samples prepared at 1500 °C. The increase in temperature for stoichiometric samples showed a monotonic increase of decay time due to the reduction of non-radiative recombination defects. It was found that the Ca excess favored the formation of oxygen vacancies which are known to act as trap. In the samples treated at 1400 °C, 3% of Ca excess showed to be the best concentration to increase the decay time of persistent luminescence. For the samples treated at 1500 °C, the segregation of Ruddlesden-Popper phases left a constant amount of Ca soluble in all the CaTiO 3 samples. This constant concentration of Ca caused the same density of defects and, consequently, the same decay time in all samples.

Several studies were performed to understand the luminescent properties in CaTiO 3 :Pr. Zhang et al. enhanced the luminescence and afterglow with substitution of Ti by Zr due to the change in the lattice symmetry from orthorhombic to pseudo-cubic. Also, this replacement reduces the concentration of Pr 4+ which acts as a luminescence quencher because Zr has only a +4 valence, making the charge transfer impossible like in the case of Ti: Ti +4 -O-Pr +3 to Ti +3 -O-Pr +4 [14]. Yin et al. prepared CaTiO 3 :(Pr,Al) by a combustion method finding a relationship with the photoluminescence intensity and the H 3 BO 3 concentration [15]. Zhang et al. suppressed the non-radiative centers and increased the trap density by addition of Lu 2 O 3 [16].
An interesting approach to the problem was developed also by Zhang et al. who manipulated the photoluminescence and phosphorescence by changing the size of CaTiO 3 :Pr nanoparticles [17]. For nanoparticles, where the surface to volume ratio is higher than in bulk materials, the surface defects play an important role as traps or non-radiative recombination centers [18]. With thermal annealing the surface of nanoparticles is reduced and therefore also the density of surface defects. The authors found the maximum for luminescence intensity in function of the particle size where the system has the optimal balance between traps and nonradiative recombination centers.
Calcium titanate is a well-known material and the chemistry of defects has been deeply studied [18][19][20][21]. The incorporation of CaO excess in the lattice can be achieved by the formation of Ruddlesden-Popper (RP) like defects [22,23]. These defects have the form of CaO interlayers with rock salt structure perpendicular to the c-axis of perovskite blocks [24]. The distribution of these slabs in the CaTiO 3 structure is either in the form of isolated defects or grouped with irregular periodicity. These defects are soluble in the CaTiO 3 lattice up to 1480 °C. At higher temperature, the segregation of Ruddlesden-Popper phases, Ca 3 Ti 2 O 7 and Ca 4 Ti 3 O 10 , may occur [25]. Ca excess can also be introduced in the lattice by another mechanism, in this case, the Ca and O are placed in its crystallographic site and Ti and O vacancies are formed [20]. This last mechanism should have a stronger influence on the persistent luminescence because oxygen vacancies can act as traps and increase the decay time [12].
In this work we explored the introduction of defects in a controllable way. These high temperature stable defects allowed reducing the non-radiative recombination centers in the surface of the particles and increasing the decay time for the persistent luminescence. The stability of defects in function of temperature and CaO excess concentration was studied to obtain the optimal conditions.

Experimental section
Powder samples of CaTiO 3 :Pr were synthetized by solid state reaction using CaCO 3 (Riedelde Haën -puriss.), TiO 2 (Sigma Aldrich -puriss.) and Pr 2 O 3 (Alpha Aesar -99.9%). The nominal concentration of Pr was 0.2% in all the samples and the Ca concentration was calculated to give a final composition of Ca 0.998+x Pr 0.002 TiO 3+δ (0.02≤x≤0.04). The raw materials were mixed in a planetary ball mill with ethanol and ZrO 2 balls. The mixed powders were dried at 100 °C to eliminate the solvent and calcined between 1300 °C and 1500 °C during 24 hs in order to obtain the desired phase.
Powder X-ray diffraction (XRD) was used for the characterization of the samples. The XRD patterns of the products were obtained using a PANalytical X´Pert PRO θ-2θ scan system equipped with Johansson monochromator and a X'Celerator linear detector. The incident X-rays had a wavelength of 1.5406 Å (Cu-K α1 ). The diffraction patterns were scanned from 20° to 100° (2θ) with an angular step interval of 0.0167°.
The nanostructure and elemental composition were studied by transmission electron microscopy using a JEOL JEM 200 FS TEM/STEM together with energy dispersive X-ray spectroscopy (EDS). Powder samples were dispersed in ethanol and transferred to a copper grid supported with a holey carbon film.
The spectra were measured with a Fluorolog 3-22 Instrument at room temperature with a nominal resolution of 1 nm. For the persistence luminescence measurements, the samples were activated during 1 min with UV light (254 nm). The active area was standardized to 1 cm 2 with an approximate thickness of 3 mm. The decay time was measured in a self-made photometer with a photomultiplier tube (Phototec). For the comparison of the performance of the materials, persistent luminescence decay time, the limit of light perception of the darkadapted human eye was used as parameter (32 ncd/cm 2 ).

Results and discussion
All the samples treated between 1300 °C and 1500 °C exhibited diffraction patterns with CaTiO 3 as main phase (Fig. 1). No detectable impurities like CaO or Ca(OH) 2 were found by XRD in the samples treated at 1300 °C and 1400 °C, indicating the complete solubility of CaO in CaTiO 3 . Samples treated at 1500 °C revealed small impurities which can be identified as Ca 3 Ti 2 O 7 and/or Ca 4 Ti 3 O 10 . The low intensity and the overlap of the reflexes with the ones corresponding to CaTiO 3 made it difficult to clearly identify the phase. The CaO excess can be incorporated in the crystal structure in the form of local defects (Ti and O vacancies) and/or CaO layers (according to the mechanisms mentioned above) when the samples are prepared at 1300 °C and 1400 °C. The existence of Ca 3 Ti 2 O 7 in the samples treated at 1500 °C is in good agreement with the phase diagram reported by Roth [25] in which the formation of the Ca rich phase is observed at 1480 °C.
The local composition of CaTiO 3 :Pr samples treated at 1400 °C was studied by TEM and EDS. EDS measurements of all samples confirmed the existence of CaTiO 3 grains with a Ca:Ti ratio of 1:1. Figure 2 shows a line scan across the particle prepared with 3% Ca excess. The Ca:Ti ratio is close to one, confirming the homogeneous distribution of Ca excess in the CaTiO 3 particles. The excitation and emission spectra of samples with different Ca excess and thermal treatments can be observed in Fig. 3. The emission spectra are dominated by the red emission at 612 nm, which corresponds to the transition for 1 D 2 to 3 H 4 ground state. The excitation spectra present three bands, which are in agreement with previously reports in bibliography [12]. The A band (~330 nm) presents no changes in the intensity, while B (~360 nm) and C (~280 nm) exhibit clear changes in the relative intensities. Band B was reported to change the intensity with thermal treatment, which is in agreement with the observations in Fig. 3. The persistent luminescence of the stoichiometric CaTiO 3 :Pr samples treated at different temperatures is shown in Fig. 4. The persistent luminescence decay time exhibits an increase with the temperature of thermal treatment. This can be related to the reduction of nonradiative recombination centers and the efficient charge transport from traps to the emission centers as previously reported for nanoparticles treated at different temperatures [17]. The persistent luminescence decay of the samples with Ca excess treated at 1400 °C is shown in Fig. 5 and illustrates that all samples with Ca excess exhibit longer decay time than the stoichiometric one. As previously explained, the incorporation of CaO defects in the CaTiO 3 structure can induce the formation of oxygen vacancies [20] which can act as electron traps and increase the decay time in the persistent luminescence [12]. Remarkable is the maximum decay time at 3% of Ca excess. It was reported that the substitution of Ba in BaTiO 3 :Pr with trivalent ions (e.g. Al, Ga, Sc) has a strong influence on the luminescence intensity [26]. According to the mechanism presented, the substitution of trivalent ions located in the Ti sites close to Pr compensates the charge imbalance caused by the substitution of Ca by Pr. This indicates that a charge imbalance close to Pr ions reduces the luminescence intensity. In our case, the introduction of CaO excess would generate oxygen vacancies [20]. Low concentration of these defects act as traps, increasing the persistent luminescence decay time, but as the density of oxygen vacancies increases, these defects generate a charge imbalance and decrease the decay time.
The comparison of samples treated at 1500 °C show a different behavior in contrast to the group of samples treated at 1400 °C (Fig. 6). According to the observations in the XRD patterns ( Fig. 1(c)), the Ca excess was segregated in the form of RP phases. The different decay time between the stoichiometric and non-stoichiometric samples suggest that some small amounts of Ca remains soluble in the matrix. This remaining amount of Ca in the matrix should be constant in the non-stoichiometric samples due to the precipitation of RP phases. This is reflected in the persistent luminescence decay time measurements (Fig. 6). Samples with non-stoichiometric composition have all the same behavior, which is clearly different from the samples with stoichiometric composition. The samples treated at 1500 °C showed the longest persistent luminescence of all samples despite that the remaining Ca excess in these samples is lower compared to samples treated at lower temperature. This indicates that a long persistent luminescence decay time is the consequence of more than one factor, which are in our case the higher thermal treatment and a Ca excess in the lattice. The time required to achieve the limit of light perception of the darkadapted human eye of all the samples is summarized in Table 1. To understand of the mechanism for luminescence decay, curves were plotted in log-log (Figs. 4-6) and hyperbolic. The lack of linearity of the curves indicates that there is not a simple single mechanism for de-excitation of charges in the system. To understand this mechanism, further characterizations should be performed.

Conclusions
Luminescent Ca 0.998 + x Pr 0.002 TiO 3+δ (0.02≤x≤0.04) samples with different afterglow properties were successfully prepared by solid state reactions with different thermal post treatments. Variations of the Ca content and thermal treatments allowed us to modify the defects which can act as traps and increase the decay time in persistent luminescence. The Ca excess showed to be soluble up to 4% in the form of local defects or isolated CaO slabs for the samples prepared at 1400 °C. The samples treated at 1500 °C revealed a defined concentration of Ca in the structure. A Ca content exceeding this concentration was segregated in form of RP phases. The Ca excess in the samples treated at 1400 °C showed that 3% is the best concentration for long time persistent luminescence. For the samples treated at 1500 °C, the constant Ca concentration in the structure caused a similar decay time in the persistent luminescence. The samples treated at higher temperature showed the longest decay time despite of majority of Ca excess segregation from the CaTiO 3 lattice, showing that the small amount of remaining soluble CaO in the CaTiO 3 structure has a stronger effect on the luminescence than the higher concentrations in the samples treated at lower temperatures and indicating that the performance of long persistent luminescence pigments is the consequence of more than only one factor, e.g. thermal treatment and defect concentration in the lattice.