Vacuum Referred Binding Energy (VRBE)-Guided Design of Orange

Aug 23, 2017 - Orange persistent phosphors of Ca3Si2O7 doped with Eu2+ were ... For a more comprehensive list of citations to this article, users are ...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/IC

Vacuum Referred Binding Energy (VRBE)-Guided Design of Orange Persistent Ca3Si2O7:Eu2+ Phosphors Jumpei Ueda,* Ryomei Maki, and Setsuhisa Tanabe Graduate School of Human and Environmental Studies, Kyoto University, Kyoto 606-8501, Japan

Downloaded via TUFTS UNIV on June 28, 2018 at 21:59:33 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S Supporting Information *

ABSTRACT: Orange persistent phosphors of Ca3Si2O7 (CSO) doped with Eu2+ were strategically developed by codoping Sm3+ or Tm3+. First, a vacuum referred binding energy, VRBE, diagram of Ca3Si2O7 (CSO) was constructed from the measured spectroscopic data. By the zigzag curve of the divalent lanthanide ions in the VRBE diagram, Sm3+ and Tm3+ ions were predicted to be a suitable electron trap for the persistent luminescence. The initial persistent luminance of CSO:Eu2+-Sm3+ and CSO:Eu2+-Tm3+ was found to be 290 times and 9300 times stronger, respectively, compared with CSO:Eu2+. By optimizing Eu2+ and Tm3+ concentrations, the persistent luminescence duration on 0.32 mcd/m2 reached approximately 50 min in CSO:Eu2+-Tm3+. From the VRBE diagram and the persistent luminescence properties, we discuss the persistent mechanism including the charging process, detrapping process, and electron trapping centers.

1. INTRODUCTION Persistent phosphors, which emit light after ceasing excitation light, have been widely used in many applications; luminous paints in the dark for hazard signboards and road markers and luminescent markers for in vivo imaging. Almost all bright and long persistent phosphors have been developed in Eu2+-doped compounds and show green and blue persistent luminescence; SrAl2O4:Eu2+-Dy3+ (λem = 510 nm),1 CaAl2O4:Eu2+-Nd3+ (λem = 440 nm),1 Sr4Al14O25:Eu2+-Dy3+ (λem = 490 nm),2,3 and Sr2MgSi2O7:Eu2+-Dy3+ (λem = 470 nm).4 Even among them, only SrAl2O4:Eu2+-Dy3+ and Sr4Al14O25:Eu2+-Dy3+ are practically used because of their high persistent luminance, long persistent luminescence duration, and good chemical durability. On the other hand, only a few orange or red persistent phosphors have been reported mainly in Eu2+ and Mn2+ -doped materials5,6 in spite of the high demand for the wide color variation in the persistent luminescent applications. Here, it should be noted that recently Eu3+ persistent phosphors have also gathered attention because of the good red emission color and the scientific interest of hole-detrapping persistent luminescence mechanism.7,8 Focusing on Eu2+-doped orange and red persistent phosphors, we can find some phosphors such as Ca2Si5N8:Eu2+-Tm3+ (λem = 630 nm),9 Ca2SiS4:Eu2+-Nd3+ (λem = 660 nm)10, and CaS:Eu2+-Tm3+ (λem = 650 nm).11 However, they have some disadvantages, for instance, nitrides are generally synthesized under high N2 pressure and high temperature and sulfides have low water durability. Oxide host materials show better chemical durability and can be prepared under atmospheric pressure air at relatively lower temperatures, but almost all Eu2+-doped oxides show blue to green luminescence due to the weaker nephelauxetic effect and crystal field splitting.12 © 2017 American Chemical Society

As one of the potential candidates, we focused on the Ca3Si2O7:Eu2+ orange phosphor discovered by Toda et al.13 The Ca3Si2O7 crystal is found in minerals and called as rankinite, in which there are three distorted Ca2+ sites with strong crystal field splitting for Eu2+.14−16 As a result, the Ca3Si2O7:Eu2+ phosphor shows rare orange luminescence among Eu2+-doped oxide phosphors by UV or blue excitation but does not show persistent luminescence due to no appropriate electron traps. Previously, Jin et al. reported the orange persistent luminescence in Ca3Si2O7:Eu2+ codoped with Dy3+, Er3+, or Tm3+ in 2014.17 However, the electron traps from the energy level of divalent lanthanide ions and the charging mechanism were not discussed well. For considering the trap depth of lanthanide codopant and discussing the charging and detrapping processes, a vacuum referred binding energy (VRBE) diagram including the Ln3+/ Ln2+ ground state (GS), conduction band (CB), and valence band (VB), which was proposed by Dorenbos, becomes a useful tool.18−23 The Ln3+ codopant in persistent phosphors can capture one electron during charging, and the Ln3+ ion itself changes into Ln2+ state. The electron trap depth generally corresponds to the energy gap between the bottom of CB and Ln2+ GS. Because the VRBE of Ln2+ GS follows a zigzag curve as a function of the number of 4f electrons, the electron trap depth can be controlled by a type of lanthanide ion. Bos et al. reported the systematic energy shift of the electron trap depth by changing a codopant ion in YPO4:Ce3+ phosphors and the good agreement between the estimated trap depth from the thermoluminescence (TL) and that predicted from the energy Received: May 19, 2017 Published: August 23, 2017 10353

DOI: 10.1021/acs.inorgchem.7b01214 Inorg. Chem. 2017, 56, 10353−10360

Article

Inorganic Chemistry diagram.24−26 Also some persistent phosphors such as SrSi2AlO2N3:Eu2+-Ln3+ and Sr3AlxSi1−xO5:Ce3+-Ln3+ show the similar systematic trap depth shift by the type of Ln3+ based on the VRBE diagram.27,28 In this study, on the basis of the VRBE diagram we constructed for the Ca3Si2O7 host, two different lanthanides ions, Sm3+ and Tm3+, were selected as electron traps among 14 lanthanides and the orange persistent oxide phosphors of Ca3Si2O7:Eu2+-Sm3+ and Ca3Si2O7:Eu2+-Tm3+ were strategically developed. The luminescence properties were carefully investigated by the measurement of photoluminescence (PL) and PL excitation (PLE) spectra, temperature dependence of lifetime, persistent luminescence (PersL) and PersL excitation (PersLE) spectra, PersL decay curves, and TL glow curves. From the VRBE diagram and the luminescence properties, we discuss the PersL mechanism including the charging process, detrapping process, and electron trapping centers.

Science, Okazaki, was used. Emission spectra were detected using a combined spectroscopy system of a grating monochromator (Princeton Instruments, Acton SP 2300i) and a CCD detector (Roper Scientific, LN/CCD-100EB-GI). The excitation spectra were measured with a photomultiplier tube (Hamamatsu, R928) attached at another output port of the grating monochromator. The PersLE spectrum was measured by a spectrofluorophotometer (Shimadzu, RF-5000). The sample was charged by monochromatic light for 1 min, and then the PersL spectrum (450−800 nm) was detected at 1 min after ceasing the charging light. This procedure was repeated with decreasing charging wavelength by 10 nm throughout the wavelength range (580 to 200 nm) by a homemade program. For the measurement of PersL decay curves, the samples in a dark box were charged for 5 min by UV light (250 nm ∼400 nm) obtained by the Xe lamp and a UV cold mirror. The PersL luminescence intensity after UV charging was monitored by the PMT. The luminance at 30 s after ceasing the excitation light was obtained by using a calibrated multichannel spectrometer (B&W TEK, Glacier X) and the overall persistent luminescence intensity in the persistent luminescence decay curves was converted into the absolute luminance. For the TL measurements, the sample was mounted on a cryostat (JANIS, VPF-800) to control the temperature from 80 to 600 K. After illumination of the UV light at 80 K for 10 min, the sample was kept in the dark at the same temperature for another 10 min. Then, the sample was heated at the rate of 10 K/min to 600 K and the TL intensity was monitored by the PMT. The TL spectra were also detected every 5 K from 80 to 600 K using the multichannel spectrometer. The temperature dependence of luminescence decay curves was performed in a cryostat (Iwatani, D105) using a dye laser of coumarin (475 nm) pumped by a N2 laser (USHO Optical Systems, KEC-200), a silicon photodiode (Thorlabs Inc., DET110), and an oscilloscope (Teledyne-LeCroy, HDO4104). The luminescence decay curves of Eu2+ luminescence were obtained from 25 to 450 K.

2. EXPERIMENTAL SECTION 2.1. Materials and Synthesis Procedures. Ca3Si2O7 (CSO) phosphors doped with Eu2+(1%), Eu3+(1%), Ce3+(0.5%), Eu2+(0.5%)Sm3+(0.1%), Eu2+(0.5, 0.1%)-Tm3+(0.1%, 0.05%) were prepared by solid-state reaction. The detail compositions and sample notations of the prepared phosphors are shown in Table 1. For the charge

Table 1. Sample Composition and Notation composition

notation

(Ca0.99Eu2+0.01)3Si2O7 (Ca0.99Eu3+0.01)3Si2O7 (Ca0.995Ce3+0.005)3Si2O7 (Ca0.994Eu2+0.005Sm3+0.001)3Si2O7 (Ca0.994Eu2+0.005Tm3+0.001)3Si2O7 (Ca0.998Eu2+0.001Tm3+0.0005)3Si2O7

CSO:Eu2+ CSO:Eu3+ CSO:Ce3+ CSO:Eu2+-Sm3+ CSO:Eu2+-Tm3+ CSO:Eu2+(0.1%)-Tm3+(0.05%)

3. RESULTS AND DISCUSSION 3.1. XRD and Crystal Structure of Ca3Si2O7. Figure 1a shows the XRD pattern of CSO:Eu2+ and the simulated pattern of Ca3Si2O7. Because there is no reference card of the Ca3Si2O7 XRD pattern in the wide angle range, the XRD pattern of CSO:Eu2+ was compared with the simulated pattern. On the basis of the matching between the measured and the simulated XRD pattern, all the obtained samples were identified as a single phase of Ca3Si2O7, which has rankinite monoclinic structure with P21/a space group.14,15 As shown in Figure 1b, the Ca3Si2O7 crystal structure has three crystallographically different Ca sites with 7-fold coordination. On the basis of the similar ionic radius to Ca2+, Eu2+ ions occupy these irregular 7fold Ca sites. 3.2. PL and PLE of CSO:Eu2+, CSO:Eu3+, and CSO:Ce3+. To construct the VRBE diagram for Ca3 Si 2 O 7 from spectroscopic data, the PL and PLE spectra for CSO:Eu2+, CSO:Eu3+, and CSO:Ce3+ were measured as shown in Figure 2. The PL spectrum of CSO:Eu2+ shows a broad band centered at 620 nm due to the Eu2+: 4f65d1 → 4f7 transition. From the PLE spectrum, three main excitation bands due to the Eu2+: 4f65d1 ← 4f7 transitions were observed at 345, 385, and 485 nm and the host exciton peak was observed at 187 nm. For the PL and PLE of CSO:Eu3+, a typical sharp PL peak due to the Eu3+:5D0−7F2 transitions was observed at around 610 nm. In the PLE of CSO:Eu3+, there are the multiple f−f transitions of Eu3+ in the range of 300−600 nm and one broad band due to the charge transfer (CT) band from O2− to Eu3+ in the range of 200−300 nm. For the PL and PLE spectrum of CSO:Ce3+, a broad luminescence at around 430 nm due to the Ce3+: 4f05d1 → 4f 1 5d 0 transition was observed. When monitoring

neutrality, cation vacancies can be generated in Ln3+-doped samples. CaCO3 (3N), SiO2 (3N), Eu2O3 (4N), Sm2O3 (3N), and Tm2O3 (3N) were used as starting materials. All oxides were weighed according to stoichiometric ratio and mixed thoroughly in an alumina mortar. The homogenized powder was preheated for decarbonation of CaCO3 at 1173 K for 4 h. The fired powder was pulverized again and dry-pressed at 20 MPa into 13 mm × 2 mm thick pellet. The Eu3+ singly doped sample was sintered at 1573 K for 6 h under O2 while other samples including Eu2+ or Ce3+ were sintered under reducing atmosphere of 5% H2/95% N2. The crystal phases were identified by an XRD measurement system (Rigaku, Ultima IV). The XRD pattern of Ca3Si2O7 was simulated using the Rietan-FP program29 from the crystal structure reported by Kusachi.14 2.2. Optical Measurements. The PL spectra were obtained by two measurement systems. One is measured by a spectrofluorophotometer (Shimadzu, RF-5300) and another is measured by a multichannel spectrometer (Ocean Optics, QE65 Pro) with monochromatic light excitation obtained by a Xe lamp (Asahi Spectra, MAX-302) and a band-pass filter (400 nm). The latter system is for the purpose of the precise comparison with PersL spectrum. Thus, the PersL spectrum was also measured by the multichannel spectrometer after the 400 nm charging for 1 min. The PLE spectra between 300 and 550 nm were measured by the spectrofluorophotometer. The PLE spectra in the vacuum ultraviolet (VUV) region (160−300 nm) were measured using a D 2 lamp (Hamamatsu, L9519), a VUV monochromator (Ritu Oyo Kougaku Co., Ltd., custom-made) and a detection system using a photomultiplier tube (PMT) (Hamamatsu, R1104). The both PLE spectra were represented as one united spectrum (160−550 nm) by normalizing the spectra using the intensity at 300 nm. For high-resolution VUV spectroscopy in CSO:Ce3+ at 8K, the beamline BL3B of the UVSOR facility at the Institute for Molecular 10354

DOI: 10.1021/acs.inorgchem.7b01214 Inorg. Chem. 2017, 56, 10353−10360

Article

Inorganic Chemistry

3.3. Selection of Electron Trap from VRBE Diagram. To predict the best codopant for persistent luminescence at room temperature in the CSO:Eu2+ phosphor, the VRBE diagram for Ca3Si2O7 was constructed from the obtained spectroscopic data as shown in Figure 3 according to the

Figure 3. Host referred binding energy (HRBE) and vacuum referred binding energy (VRBE) diagram for Ca3Si2O7 host.

method by Dorenbos.22,23 For the VRBE diagram, the horizontal axis represents lanthanides as a function of atomic number, while the vertical axis is the binding energy based on the vacuum level. Red and light blue zigzag curves connect the Ln2+ GSs and Ln3+ GSs, respectively. Before constructing the VRBE diagram, host referred binding energy (HRBE) diagram, in which the vertical axis is the binding energy based on the top of VB, was constructed first. To estimate the band gap at low temperature, the host exciton energy (6.63 eV) at room temperature obtained from the PLE of CSO:Eu2+ in Figure 2 was corrected by adding 0.15 eV due to the energy shift by the temperature,30 and then by multiplying 1.08 due to the electron−hole binding energy.22 Finally, the bandgap energy for Ca3Si2O7 was estimated to be 7.32 eV as shown in the orange arrow in Figure 3 and the HRBE of the bottom of CB was determined respect to the top of VB. It should be noted that the bottom of CB is found to be mainly composed of the 4s orbital of Ca atoms and the top of VB is of the 2p orbital of O atoms from the DFT (density functional theory) calculation (see the Supporting Information). The CT energy of Eu3+ was estimated to be 5.11 eV from the PLE of CSO:Eu3+ in Figure 2. Because the CT energy is equivalent to the energy gap between the top of the VB and the Eu2+ GS as shown in the black arrow of Figure 3, the HRBE of Eu2+ GS was determined. To estimate the HRBE of trivalent lanthanide ions, the electron repulsion energy for U(6, A), which corresponds to the energy difference between the ground state of divalent and trivalent Eu ion in host A, was estimated from the centroid shift of Ce3+: 5d level in Ca3Si2O7 using eq 1 proposed by Dorenbos.23,31

Figure 1. (a) XRD pattern of Ca3Si2O7:Eu2+ sample and simulated pattern of Ca3Si2O7. (b) Crystal structure of Ca3Si2O7 represented by blue (Si1, Si2), red (Ca1), yellow (Ca2), and green (Ca3) spheres and polyhedron.

Figure 2. PL and PLE spectra of CSO:Eu3+(RT), CSO:Eu2+(RT), and CSO:Ce3+(8K).

luminescence wavelength was changed in the PLE measurement, the PLE spectral shape was also changed as shown in Figure 2. In the two PLE spectra of Ce3+, totally 10 excitation bands were observed (see the Supporting Information), thus we concluded that there are two different Ce3+ sites, in which each Ce3+: 5d level splits to five states by crystal field. From the two sets of five excitation bands, the centroid shift (εc) for Ce3+: 5d energy level23 was estimated to be 1.61 eV for both sites.

U (6, A) = 5.44 + 2.834 exp(−Ec /2.2)

(1)

Using the centroid shift (εc) estimated from the PLE of CSO:Ce3+, U(6, Ca3Si2O7) was estimated to be 6.81 eV. The HRBE of Eu2+ GS was converted using these energy values into 10355

DOI: 10.1021/acs.inorgchem.7b01214 Inorg. Chem. 2017, 56, 10353−10360

Article

Inorganic Chemistry the VRBE of Eu2+ GS, E4f(7, 2+, A), according to eq 2 proposed by Dorenbos.22,23 E4f (7, 2 + , A) = −24.92 +

18.05 − U (6, A) 0.777 − 0.0353U (6, A) (2)

The VRBE of Eu GS is determined to be −3.97 eV and finally all HRBE of VB, CB, Ln3+ GS, and Ln2+ GS were converted into the VRBE. From this VRBE diagram, only the GS of Sm2+, Tm2+, Yb2+, and Eu2+ was found to be located below the bottom of CB, which indicates that they act as electron traps. Especially, Sm3+ and Tm3+ were selected as optimum codopants for the persistent phosphor of CSO:Eu2+ at ambient temperature. This is because the energy gaps between the bottom of CB and Tm2+ or Sm3+ are estimated to be 0.49 and 0.96 eV, respectively, and these trap depths are close to the value ∼0.65 eV in the most famous SrAl2O4:Eu2+Dy3+ phosphor.1 3.4. PersL and TL of CSO:Eu2+, CSO:Eu2+-Sm3+, and CSO:Eu2+-Tm3+. The CSO:Eu2+ did not show detectable persistent luminescence by human eyes, but the samples codoped with Sm3+ and Tm3+ showed orange persistent luminescence after UV charging. Figure 4a shows the comparison of the PL and PersL spectra for CSO:Eu2+-Tm3+. The PersL band due to the Eu2+: 4f65d1 → 4f7 was observed at around 620 nm. Its spectral shape completely overlaps with that of PL spectrum, which indicates that the Eu2+ center contributing to the persistent luminescence is the same as that for the photoluminescence. Figure 4b shows the PersL decay curves for CSO:Eu2+, CSO:Eu2+-Sm3+ and CSO:Eu2+Tm3+. The initial PersL intensities of CSO:Eu2+-Sm3+ and CSO:Eu2+-Tm3+ at 10 s after ceasing the excitation are improved by 290 and 9300 times than that for CSO:Eu2+. This result strongly indicates that the codopant ions act as suitable electron trapping centers. The persistent luminance time on the 0.32 mcd/m2 for CSO:Eu2+-Sm3+ and CSO:Eu2+Tm3+ are 1202 and 1191 s, respectively. The PersL decay rate of the Tm3+ codoped sample is much faster than that of the Sm3+ codoped sample because the slope of persistent luminescence decay curve of CSO:Eu2+-Tm3+ is much steeper than that of CSO:Eu2+-Sm3+. This result can be explained by trap depth difference of two lanthanide ions as discussed later. In addition, the persistent luminescence performance of CSO:Eu2+-Tm3+ was improved by optimizing the Eu2+ and Tm3+ concentrations. The persistent luminescence decay time on 0.32 mcd/m2 in CSO:Eu2+(0.1%)-Tm3+(0.05%) is 2926 s (∼50 min). In order to check the electron trap depth by Sm3+ and Tm3+, TL glow curves were measured as shown in Figure 4c. Because the intensity of TL glow curves are affected by the thermal quenching, the TL intensity was calibrated by the Eu2+ luminescence intensity at the corresponding temperature.32,33 In addition, the TL glow curve of CSO:Eu2+ is intensified by 150 times because of the weak signal. The rise part above 500 K in the TL glow curve of CSO:Eu2+ is due to the blackbody radiation, and this part is not related to the nature of the phosphors. From the TL glow curves, two TL glow peaks were observed both in the CSO:Eu2+-Tm3+ and CSO:Eu2+-Sm3+ while in CSO:Eu2+ only a very weak TL glow peak at 260 K was observed. For the CSO:Eu2+-Tm3+, the TL peak temperatures are 209 and 278 K while for the CSO:Eu2+-Sm3+ the TL peak temperatures are 369 and 452 K. Because the CSO:Eu2+-Tm3+ and CSO:Eu2+-Sm3+ show different additional TL peaks 2+

Figure 4. Comparison of PL and PersL spectra of CSO:Eu2+-Tm3+ by and after 400 nm excitation. (b) Persistent decay curve of CSO:Eu2+, CSO:Eu2+-Sm3+, and CSO:Eu2+-Tm3+ after UV illumination at room temperature. (c) TL glow curves of CSO:Eu2+, CSO:Eu2+-Tm3+, and CSO:Eu2+-Sm3+ after UV charging.

compared with CSO:Eu2+, Tm3+ and Sm3+ are found to act as electron traps in Ca3Si2O7. From the TL peak temperature, the trap depth ε was estimated by the following equation: ⎛ ε ⎞ βε = s exp⎜ − ⎟ 2 kTm ⎝ kTm ⎠

(3) −1

where β is the heating rate (K/s ), k is the Boltzmann constant (eV/K), s is the frequency factor (s−1) and Tm is the absolute temperature at the TL glow curve peak maximum (K). Here, we assumed s to be a typical value of 1.0 × 1013 s−1. The trap depths of CSO:Eu2+-Tm3+ are 0.60 and 0.81 eV, while that of CSO:Eu2+-Sm3+ are 1.08 and 1.34 eV. The energy gaps between two TL peaks for CSO:Eu2+-Tm3+ and CSO:Eu2+Sm3+ are 0.21 and 0.26 eV, respectively, which are very similar to each other. In addition, the shape of TL glow curve of CSO:Eu2+-Sm3+ is almost the same as that of CSO:Eu2+-Tm3+, the difference is only the TL peak temperature. The Ca3Si2O7 crystal structure possesses three crystallographically different 10356

DOI: 10.1021/acs.inorgchem.7b01214 Inorg. Chem. 2017, 56, 10353−10360

Article

Inorganic Chemistry Ca sites.14 On the basis of these facts, we suggest that two main TL glow peaks in CSO:Eu2+-Tm3+ and CSO:Eu2+-Sm3+ come from the substitution of Sm3+ or Tm3+ into two different Ca sites. As mentioned above, the energy gap between the bottom of CB and Ln2+ (Ln = Tm, Sm) based on the VRBE diagram are 0.49 and 0.96 eV, respectively. On the other hand, when we focus on the TL glow peak at the lower temperature, the trap depths of Tm3+ and Sm3+ are estimated to be 0.60 and 1.08 eV, respectively. When we focus on the TL glow peak at the higher temperature, the trap depths of Tm3+ and Sm3+ are estimated to be 0.81 and 1.34 eV, respectively. The trap depths based on TL glow peaks at lower and higher temperatures appear to be systematically 0.11 and 0.35 eV larger, respectively, in both CSO:Eu2+-Tm3+ and CSO:Eu2+-Sm3+ than that predicted from the VRBE diagram we have constructed. This deviation is probably because there is some systematic error from the TL glow peak, band gap or CT energy, or the effect by the different site occupancy of Eu3+ and Ln3+ codopants (Ln=Tm and Sm). However, it is clear that Sm3+ forms much deeper electron trap than Tm3+ from the VRBE and TL analysis. Because the detrapping rate, p, is simply expressed by p = s exp(−ε/kT), PersL decay rate CSO:Eu2+-Sm3+ with deeper trap is much slower than CSO:Eu2+-Tm3+ as shown in Figure 4. Figure 5 shows wavelength-temperature contour plots of TL intensity for (a) CSO:Eu2+-Sm3+ and (b) CSO:Eu2+-Tm3+. The

luminescence through the recombination process. Thus, Eu2+ is found to act as a stable hole trap even at high temperature of 500 K. This assumption can also be understood from the VRBE diagram. The VRBE of Eu2+ and the top of VB are −3.97 and −9.08 eV, respectively, so that Eu2+ can form very deep hole trap with trap depth of 5.11 eV and becomes a good PersL center for the electron detrapping PersL mechanism. This is one of the reasons why there are a number of Eu2+-doped persistent phosphors. One may notice that the TL peak wavelength at higher temperatures is blue-shifted compared with that at lower temperatures in Figure 5. This is because the excited state depends on the Boltzmann distribution and the electrons in the excited state are populated into the higher energy level.34 3.5. PersLE Spectra for Charging Process. In addition to the detrapping process of persistent luminescence, the charging process is discussed in this section. The energy gap between the lowest Eu2+ 5d excited state and the bottom of the CB composed by the 4s orbital of Ca is important for the charging process. This energy gap can be estimated by investigating the temperature dependence of the nonradiative decay rate if the quenching process is caused only by thermal ionization. Actually, it is reported that the quenching process of Eu2+ is caused by thermal ionization in many Eu2+-doped phosphors instead of the thermally activated crossover relaxation.35−37 Figure 6 shows the temperature dependence of luminescence

Figure 6. Temperature dependence of luminescence lifetime for CSO:Eu2+ and fitting curve.

lifetime for CSO:Eu2+. The lifetime of Eu2+: 4f65d1 → 4f7 is 3.50 μs at lower temperature, starts to decrease from 200 K, and reaches almost zero at 450 K. The quenching temperature (T50%), which is defined as the temperature at which the luminescence lifetime becomes half of that at low temperature, was estimated to be 384 K. The temperature dependence of luminescence lifetime was fitted by the following equation: 1 τ (T ) = Eq Γυ + Γ0 exp − kT

Figure 5. Wavelength-temperature contour plots of thermoluminescence for (a) CSO:Eu2+-Sm3+ and (b) CSO:Eu2+-Tm3+.

( )

broad TL spectra in both samples at all TL glow peak temperature in the range from 150 to 500 K are originated from the Eu2+: 4f65d1 → 4f7 transition and there is no TL luminescence center except Eu2+. This result shows that the trapped electron by Ln3+(Ln = Tm and Sm) is released at given temperatures and transferred back to the hole-trapped Eu2+, and then the phosphor exhibits the Eu2+: 4f65d1 → 4f7

(4)

where Γυ is radiative rate, Γ0 is attempt rate of nonradiative process, Eq is the activation energy of nonradiative process. As a result, Eq for CSO:Eu2+ is estimated to be 0.33 eV. The lifetime at 300 K is approximately 20% smaller compared with that at low temperatures. This implies that some portion of the excited electrons in the lower 5d energy levels could be used as the 10357

DOI: 10.1021/acs.inorgchem.7b01214 Inorg. Chem. 2017, 56, 10353−10360

Article

Inorganic Chemistry

the PersLE spectrum at 473 K for CSO:Eu2+-Sm3+, a stronger PersLE band was observed in the range of 350−550 nm compared with the PersLE spectrum at RT for CSO:Eu2+Tm3+. Obviously, the charging efficiency at the 5dlower levels in the range of 350−550 nm is improved at higher temperatures for Eu2+-doped Ca3Si2O7 persistent phosphors. These results strongly indicate that Eu2+-doped Ca3Si2O7 undergoes the thermal ionization process from the 5dlower to the bottom of CB and this thermal ionization process causes the thermal luminescence quenching and also the electron charging by the excitation to the 5dlower level. Although the charging efficiency by blue illumination is lower than that by UV illumination for CSO:Eu2+-Tm3+ at room temperature, the CSO:Eu2+-Tm3+ can be charged by blue illumination as shown in Figure 7a,b. In order to discuss the charging process from energy diagrams, the energy scheme combining VRBE and configuration coordinate (CC) diagrams is constructed as shown in Figure 8. For the construction of CC diagram of Ca3Si2O7:Eu2+,

ionization to the CB at RT. The effective charging wavelength was investigated from the PersLE spectra for revealing the charging process as well as the quenching process in detail. Figure 7a represents the charging wavelength−luminescence wavelength contour plot for persistent luminescence in

Figure 8. Schematic diagram combing VRBE and CC diagrams for Ca3Si2O7.

the obtained transition energies from the PLE in CSO:Eu2+ was used. In the VRBE, the 5dlower level is located at 0.32 eV above the bottom of CB, which means that the strong autoionization process from the 5dlower level to CB is caused. However, the CSO:Eu2+ shows efficient luminescence even at room temperature. These results imply that the VRBE includes some energetic error. As discussed in the TL-analysis, two trap depths of Ln3+ (Ln = Tm and Sm) in two different Ca sites are also 0.11 or 0.35 eV larger than that predicted from the VRBE diagram. On the basis of the results of trap depth from TL analysis and the strong luminescence even at RT, the actual bottom of CB is expected to be located at approximately 0.32− 0.35 eV above that determined first in the VRBE diagram (Figure 3). When the bottom of CB is increased by 0.35 eV as shown in Figure 8, the 5dlower appears below the bottom of CB, but still the energy gap between the 5dlower and the bottom of CB is small. This is probably because the 5d excited level takes much lower energy after lattice relaxation. The VRBE diagram does not take into account the lattice relaxation while the CC diagram represents this relaxation process. After lattice relaxation in the excited state in the CC diagram, the bottom of 5dlower potential curve is located below the bottom of CB with much larger energy gap compared with the VRBE diagram as shown in Figure 8. In this case, the activation energy of 0.37 eV for the nonradiative process can be understood by the

Figure 7. (a) Charging wavelength and PersL wavelength contour plot for persistent luminescence of CSO:Eu2+-Tm3+. Comparison of PLE spectra of CSO:Eu2+ and PersLE spectrum (b) at ambient temperature for CSO:Eu2+-Tm3+ and (c) at 473 K for CSO:Eu2+-Sm3+.

CSO:Eu2+-Tm3+ at RT. The wavelength of the PersL peak is always at around 620 nm, and it is independent of charging wavelength. However, the PersLE spectrum of CSO:Eu2+-Tm3+ is largely different from the PLE spectrum as shown in Figure 7b. With increasing charging wavelength, the charging efficiency in the persistent luminescence drops significantly although there is the strong excitation band up to 550 nm in the PLE spectrum. The CSO:Eu2+-Tm3+ is efficiently charged by 250− 350 nm illumination at room temperatures but not by 350−550 nm. The excitation band in the range of 250−350 nm and 350−550 nm corresponds to the higher 5d (5dhigher) and the lower 5d (5dlower) energy levels of Eu2+, respectively. From the PersLE spectrum at RT, it is expected that there is a thermal energy barrier from the 5dlower levels to the bottom of CB. Figure 7c displays the comparison of PersLE spectrum of CSO:Eu2+-Sm3+ at 473 K and PLE spectrum of CSO:Eu2+. In 10358

DOI: 10.1021/acs.inorgchem.7b01214 Inorg. Chem. 2017, 56, 10353−10360

Article

Inorganic Chemistry

was performed at the UVSOR facility under the Joint Studies Program (Grant 28-521) of the Institute for Molecular Science.

thermal ionization process from the bottom of 5dlower potential curve to the bottom of CB, not by the thermally activated crossover process due to the large energy barrier to the cross point with 4f GS. On the other hand, the 5dhigher levels are completely located within the CB even if the lattice relaxation is taken into account, which indicates that the charging process from the 5dhigher levels to the CB is caused by strong autoionization. Thus, the charging efficiency through the autoionization from the 5dhigher level in the range of 250−350 nm is much higher than that through the thermal ionization from the 5dlower level in the range of 350−550 nm.

■

(1) Matsuzawa, T.; Aoki, Y.; Takeuchi, N.; Murayama, Y. A new long phosphorescent phosphor with high brightness, SrAl2O4:Eu2+,Dy3+. J. Electrochem. Soc. 1996, 143, 2670. (2) Lin, Y.; Tang, Z.; Zhang, Z. Preparation of long-afterglow Sr4Al14O25-based luminescent material and its optical properties. Mater. Lett. 2001, 51, 14. (3) Lin, Y.; Tang, Z.; Zhang, Z.; Nan, C. W. Anomalous luminescence in Sr4Al14O25:Eu, Dy phosphors. Appl. Phys. Lett. 2002, 81, 996. (4) Lin, Y.; Tang, Z.; Zhang, Z.; Wang, X.; Zhang, J. Preparation of a new long afterglow blue-emitting Sr2MgSi2O7-based photoluminescent phosphor. J. Mater. Sci. Lett. 2001, 20, 1505. (5) Van den Eeckhout, K.; Smet, P. F.; Poelman, D. Persistent Luminescence in Eu2+-Doped Compounds: A Review. Materials 2010, 3, 2536. (6) Van den Eeckhout, K.; Poelman, D.; Smet, P. Persistent Luminescence in Non-Eu2+-Doped Compounds: A Review. Materials 2013, 6, 2789. (7) Luo, H.; Bos, A. J. J.; Dorenbos, P. Controlled Electron−Hole Trapping and Detrapping Process in GdAlO3 by Valence Band Engineering. J. Phys. Chem. C 2016, 120, 5916. (8) Luo, H.; Bos, A. J. J.; Dorenbos, P. Charge Carrier Trapping Processes in RE2O2S (RE = La, Gd, Y, and Lu). J. Phys. Chem. C 2017, 121, 8760. (9) Miyamoto, Y.; Kato, H.; Honna, Y.; Yamamoto, H.; Ohmi, K. An Orange-Emitting, Long-Persistent Phosphor, Ca2Si5N8: Eu2+, Tm3+. J. Electrochem. Soc. 2009, 156, J235. (10) Smet, P. F.; Avci, N.; Poelman, D. Red Persistent Luminescence in Ca2SiS4: Eu, Nd. J. Electrochem. Soc. 2009, 156, H243. (11) Jia, D.; Zhu, J.; Wu, B. Trapping Centers in CaS: Bi3+ and CaS: Eu2+, Tm3+. J. Electrochem. Soc. 2000, 147, 386. (12) Dorenbos, P. Energy of the first 4f7 ·f65d transition of Eu2+ in inorganic compounds. J. Lumin. 2003, 104, 239. (13) Toda, K.; Kawakami, Y.; Kousaka, S. I.; Ito, Y.; Komeno, A.; Uematsu, K.; Sato, M. New silicate phosphors for a white LED. IEICE Transactions on Electronics 2006, E89-C, 1406. (14) Kusachi, I.; Henmi, C.; Kawahara, A.; Henmi, K. The structure of rankinite. Mineral. J. 1975, 8, 38. (15) Saburi, S.; Kusachi, I.; Henmi, C.; Kawahara, A.; Henmi, K.; Kawada, I. Refinement of the structure of rankinite. Mineral. J. 1976, 8, 240. (16) Kim, S. W.; Toda, K.; Hasegawa, T.; Uematsu, K.; Sato, M. In Phosphors, Up Conversion Nano Particles, Quantum Dots and Their Applications: Vol. 1; Liu, R.-S., Ed.; Springer Berlin Heidelberg: Berlin, Heidelberg, 2017; p 219. (17) Jin, Y.; Hu, Y.; Chen, L.; Wang, X.; Mu, Z.; Ju, G.; Wang, T. A novel orange emitting long afterglow phosphor Ca3Si2O7:Eu2+ and the enhancement by R3+ ions (R = Tm, Dy and Er). Mater. Lett. 2014, 126, 75. (18) Dorenbos, P. Mechanism of persistent luminescence in Eu2+ and Dy3+ codoped aluminate and silicate compounds. J. Electrochem. Soc. 2005, 152, 107. (19) Dorenbos, P. Lanthanide charge transfer energies and related luminescence, charge carrier trapping, and redox phenomena. J. Alloys Compd. 2009, 488, 568. (20) Dorenbos, P. Modeling the chemical shift of lanthanide 4f electron binding energies. Phys. Rev. B: Condens. Matter Mater. Phys. 2012, 85, 165107 DOI: 10.1103/PhysRevB.85.165107. (21) Dorenbos, P. A Review on How Lanthanide Impurity Levels Change with Chemistry and Structure of Inorganic Compounds. ECS J. Solid State Sci. Technol. 2013, 2, R3001. (22) Dorenbos, P. Determining binding energies of valence-band electrons in insulators and semiconductors via lanthanide spectroscopy. Phys. Rev. B: Condens. Matter Mater. Phys. 2013, 87, 035118.

4. CONCLUSION Orange persistent phosphors were successfully developed in Ca3Si2O7:Eu2+ by codoping Tm3+ and Sm3+. By using the transition energies of host exciton, CT of Eu3+ and centroid shift of Ce3+: 5d, which are obtained experimentally, the VRBE diagram was constructed. For the best lanthanide electron trap in Ca3Si2O7:Eu2+, Tm3+ and Sm3+ were carefully selected from the constructed VRBE diagram. From the PersL properties for Ca3Si2O7:Eu2+, Ca3Si2O7:Eu2+-Tm3+, and Ca3Si2O7:Eu2+-Sm3+, the persistent luminescence was surely induced by codoping with Tm3+ and with Sm3+. In the CSO:Eu2+(0.1%)-Tm3+(0.5%) in which Eu and Tm concentration was optimized, the persistent luminescence duration on 0.32 mcd/m2 reached approximately 50 min. In CSO:Eu2+-Tm3+, two main trap depths of Tm3+ in two different Ca sites are estimated to be 0.60 and 0.81 eV, while in CSO:Eu2+-Sm3+ two main trap depth are estimated to be 1.08 and 1.34 eV. The tendency of TL glow peaks by Tm3+ and Sm3+ corresponds to that estimated from the constructed VRBE diagram. From the PersLE spectrum, the charging efficiency through the autoionization from the 5dhigher level in the range of 250−350 nm is much higher than that through the thermal ionization from the 5dlower level in the range of 350−550 nm.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.7b01214. High-resolution PLE spectrum of CSO:Ce3+; total density of state (TDOS) and partial density of state (PDOS) of Ca3Si2O7 (PDF)

■

REFERENCES

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Jumpei Ueda: 0000-0002-7013-9708 Author Contributions

The manuscript was written through contributions of all authors. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by JST matching planners program (Grant MP27215667896) and by JSPS KAKENHI (Grant Numbers 16K05934 and 16H06441). The VUV spectroscopy 10359

DOI: 10.1021/acs.inorgchem.7b01214 Inorg. Chem. 2017, 56, 10353−10360

Article

Inorganic Chemistry (23) Dorenbos, P. Ce3+ 5d-centroid shift and vacuum referred 4felectron binding energies of all lanthanide impurities in 150 different compounds. J. Lumin. 2013, 135, 93. (24) Bos, A. J. J.; Dorenbos, P.; Bessière, A.; Lecointre, A.; Bedu, M.; Bettinelli, M.; Piccinelli, F. Study of TL glow curves of YPO4 double doped with lanthanide ions. Radiat. Meas. 2011, 46, 1410. (25) Dorenbos, P.; Bos, A. J. J. Lanthanide level location and related thermoluminescence phenomena. Radiat. Meas. 2008, 43, 139. (26) Bos, A. J. J.; Dorenbos, P.; Bessière, A.; Viana, B. Lanthanide energy levels in YPO4. Radiat. Meas. 2008, 43, 222. (27) Zhuang, Y.; Lv, Y.; Li, Y.; Zhou, T.; Xu, J.; Ueda, J.; Tanabe, S.; Xie, R.-J. Study on Trap Levels in SrSi2AlO2N3:Eu2+,Ln3+ Persistent Phosphors Based on Host-Referred Binding Energy Scheme and Thermoluminescence Analysis. Inorg. Chem. 2016, 55, 11890. (28) Luo, H.; Bos, A. J. J.; Dobrowolska, A.; Dorenbos, P. Lowtemperature VUV photoluminescence and thermoluminescence of UV excited afterglow phosphor Sr3AlxSi1‑xO5:Ce3+,Ln3+ (Ln = Er, Nd, Sm, Dy and Tm). Phys. Chem. Chem. Phys. 2015, 17, 15419. (29) Izumi, F.; Momma, K. Three-Dimensional Visualization in Powder Diffraction. Solid State Phenom. 2007, 130, 15. (30) Dorenbos, P. Electronic Structure and Optical Properties of the Lanthanide Activated RE3(Al1−xGax)5O12 (RE = Gd, Y, Lu) Garnet Compounds. J. Lumin. 2013, 134, 310. (31) Dorenbos, P. Lanthanide 4f-electron binding energies and the nephelauxetic effect in wide band gap compounds. J. Lumin. 2013, 136, 122. (32) Bos, A. J. J. Theory of Thermoluminescence. Radiat. Meas. 2006, 41, S45. (33) Ueda, J.; Dorenbos, P.; Bos, A. J. J.; Meijerink, A.; Tanabe, S. Insight into the Thermal Quenching Mechanism for Y3Al5O12:Ce3+ through Thermoluminescence Excitation Spectroscopy. J. Phys. Chem. C 2015, 119, 25003. (34) Yen, W. M.; Shionoya, S.; Yamamoto, H. Phosphor Handbook; CRC Press/Taylor and Francis, 2007. (35) Dorenbos, P. Thermal quenching of Eu2+ 5d−4f luminescence in inorganic compounds. J. Phys.: Condens. Matter 2005, 17, 8103. (36) Ueda, J.; Nakanishi, T.; Katayama, Y.; Tanabe, S. Optical and optoelectronic analysis of persistent luminescence in Eu2+-Dy3+ codoped SrAl2O4 ceramic phosphor. Phys. Stat. Sol. C 2012, 9, 2322. (37) Ueda, J.; Shinoda, T.; Tanabe, S. Evidence of three different Eu2+ sites and their luminescence quenching processes in CaAl2O4:Eu2+. Opt. Mater. 2015, 41, 84.

10360

DOI: 10.1021/acs.inorgchem.7b01214 Inorg. Chem. 2017, 56, 10353−10360