Photoluminescence and radioluminescence of pure and Ce3+ activated Na3Gd(PO4)2
Hongbin Liang,Zifeng Tian,Huihong Lin,Mubiao Xie,Guobin Zhang,Pieter Dorenbos,Qiang Su
DOI: https://doi.org/10.1016/j.optmat.2010.11.020
IF: 3.754
2011-01-01
Optical Materials
Abstract:The spectroscopic properties of Na 3 Gd(PO 4 ) 2 and Na 3 Gd(PO 4 ) 2 :Ce 3+ phosphors in the VUV–UV spectral range were investigated. Five excitation bands of Ce 3+ ions at Gd 3+ sites are observed at wavelengths of 205, 246, 260, 292, and 321 nm. Doublet Ce 3+ 5d → 4f emission bands are observed at 341 and 365 nm with a decay constant τ 1/ e around 26 ns. The X-ray excited luminescence of Na 3 Gd 0.99 Ce 0.01 (PO 4 ) 2 at room temperature shows a photon yield of ∼17,000 photons/MeV of absorbed X-ray energy. Keywords Luminescence Ce 3+ Na 3 Gd(PO 4 ) 2 Phosphor 1 Introduction Ce 3+ is of the lanthanide ions the one with the simplest electronic configuration. Its ground state configuration is 4f 1 and the first excited state configuration is 5d 1 . Because 5d orbitals are outer orbitals, the coordination environments have large influence on the 5d-energy levels. Therefore, the energy of 4f–5d transitions of Ce 3+ can be found in a wide spectral range, and by changing the host lattice the emission wavelength is tunable from the ultraviolet to the red. For example, Ce 3+ shows emission wavelength maxima at about (291–355) nm in La 2 CaB 10 O 19 [1] (350–540) nm in Ba 2 Ca(BO 3 ) 2 [2] (456–567) nm in Sr 3 GaO 4 F [3] , and ∼670 nm in Y 2 O 2 S [4] . This characteristic makes Ce 3+ activated phosphors a class of important luminescent materials for display and lighting. In addition, the 4f–5d transitions of Ce 3+ show high transition rates leading to fast luminescence decays because they are parity-allowed. Because of this, Ce 3+ -doped scintillators have gained widespread applications [5] . In previous work [6] , we found a high light yield for Ce 3+ -activated MGd(PO 3 ) 4 (M = Li, Na, K, Cs) phosphors under X-ray excitation. The luminescence of lanthanide ions doped in Na 3 Gd(PO 4 ) 2 has received much attention in recent years. Schwarz et al. [7] and Kloss et al. [8] presented data on the vacuum UV absorption edge of Na 3 Gd(PO 4 ) 2 , and reported on the luminescence of Na 3 Gd 1− x Eu x (PO 4 ) 2 in 1997. More recently, Legendziewicz et al. [9] studied the quantum cutting phenomenon with Gd 3+ and Gd 3+ → Eu 3+ energy transfer. In this paper, we report on the luminescence of pure and Ce 3+ activated Na 3 Gd(PO 4 ) 2 under VUV–UV and X-ray excitation. Though there are different Gd 3+ lattice sites in Na 3 Gd(PO 4 ) 2 , we found that Ce 3+ ions preferably enter one site in the host lattice. The radioluminescence shows that the phosphor Na 3 Gd 0.99 Ce 0.01 (PO 4 ) 2 has a reasonable light yield of ∼17,000 photons emitted per MeV of absorbed X-ray energy. 2 Experimental A series of samples with nominal chemical formula Na 3 Gd (1− x ) Ce x (PO 4 ) 2 ( x = 0–0.20) were prepared by a solid-state reaction route at high temperature using the following chemical reactions: 4 NaH 2 PO 4 + ( 1 - x ) Gd 2 O 3 + 2 x CeO 2 + Na 2 CO 3 + x CO → 1123 k / 5 h 2 Na 3 Gd ( 1 - x ) Ce x ( PO 4 ) 2 + 4 H 2 O + ( 1 + x ) CO 2 The reactants include analytical grade pure NaH 2 PO 4 , Na 2 CO 3 and 99.95% pure rare earth oxides CeO 2 and Gd 2 O 3 . According to above reactions, appropriate amounts of starting materials were thoroughly mixed and ground, and subsequently calcined at 1123 K during 5 h in CO reducing atmosphere. The XRD results confirm that the proper mole ratio of reactants is NaH 2 PO 4 :[(1− x )Gd 2 O 3 + 2 x CeO 2 ]:Na 2 CO 3 = 4.2:1:1.1. That is, NaH 2 PO 4 must be with an excess 20 mol.% and Na 2 CO 3 with an excess 10 mol.%. Otherwise, a pure phase can not be obtained. In addition, the concentration of lanthanide ions was not measured. Therefore, there may be some differences between experimental and theoretical values for the doping concentrations. The structure of the final products was examined by powder X-ray diffraction (XRD) using Cu K α radiation on Rigaku D/max 2200 vpc X-ray Diffractometer, and compared with that registered in the International Center for Diffraction Data (ICDD)-Powder Diffraction Files (PDF). The steady-state UV excitation and corresponding emission spectra and the luminescence decay curves were measured on an Edinburgh FLS 920 combined fluorescence lifetime and steady state spectrometer, which was equipped with a CTI-Cryogenics temperature controlling system. A Xe900 lamp (450 W) was used as excitation source for the steady-state spectra, while an nF900 nanosecond lamp (150 W, pulse width 1 ns, pulse repetition rate 40–100 kHz) or a μF900 microsecond lamp (60 W, pulse width 1.5–3.0 μs, maximal pulse repetition rate 100 Hz) were used as excitation sources for the luminescence decay curves of Ce 3+ and Gd 3+ , respectively [1] . The VUV excitation and emission spectra were measured at the VUV spectroscopy experimental station of the National Synchrotron Radiation Laboratory (NSRL, Hefei, China) under normal operating conditions, and the excitation spectra were corrected for the photon flux at different wavelengths using the excitation spectrum of sodium salicylate (o-C 6 H 4 OHCOONa) as a standard. The details of these optical spectroscopy measurements can be found elsewhere [2] . X-ray excited emission spectra were recorded with an X-ray tube with Cu anode operating at 55 kV and 40 mA at Delft University of Technology, The Netherlands [6] . A 3 mm aluminum filter was used to filter the soft X-ray component from the beam. The emission of the sample was dispersed by means of an ARC (Acton Research Corporation) model VM-504 monochromator (blazed at 300 nm, 1200 groves/mm) and detected by a Hamamatsu R934-04 PMT (photomultiplier tube). The slit width was about 1 mm. All spectra were corrected for the wavelength dependence of the photomultiplier quantum efficiency as well as for the monochromator transmission. With this method, the wavelength integrated emission intensity provides a good relative estimate for the total light yield. The total light yield was compared with that of a pure BaF 2 sample. Pure BaF 2 emits about 9000 ph/MeV photons when 1 MeV of X-ray or γ -ray energy is absorbed in the material. Because the photon collection efficiency on the entrance slit of the monochromator depends on the sample dimensions we estimate the error in this comparative method quite large, about 30%. So the numbers for the light yield should be treated as an indicative number. 3 Results and discussion 3.1 Structural characterization The compound Na 3 Gd(PO 4 ) 2 , sodium gadolinium orthophosphate, crystallizes in three structure modifications. The alpha-phase is tetragonal with cell parameters a = 1.2990(3) nm, c = 1.0624(3) nm and space group P 4 ¯ 2 1 c . The beta-phase is orthorhombic with cell parameters a = 1.5864(7) nm, b = 1.3867(4) nm, c = 1.8375(8) nm and space group Pbcm or Pca 2 1 . The gamma-phase is monoclinic with cell parameters a = 2.7590(2) nm, b = 1.3897(4) nm, c = 0.6466(1) nm, β = 91.39(3) and space group P 2 1 / m [10] . Lately, another monoclinic phase Na 3 Gd(PO 4 ) 2 with the cell parameters a = 2.755(25) nm, b = 0.5312(4) nm, c = 1.3935(11) nm, β = 91.30(1) and space group C 2/ c was reported [11] . As examples, the XRD patterns of samples Na 3 Gd(PO 4 ) 2 and Na 3 Gd 0.99 Ce 0.01 (PO 4 ) 2 are displayed in Fig. 1 a and b. The two patterns are in good agreement with Fig. 1 c which pertains to Na 3 Gd(PO 4 ) 2 with orthorhombic structure ( β phase) [7–9] . No impurity phase was detected, demonstrating that Ce 3+ ions were dissolved in the host lattice and formed complete solid solutions. To our knowledge, the atomic positions of Na 3 Gd(PO 4 ) 2 is not reported in literature. However, Salmon et al. [12] reported structural data of single crystal Na 3 Nd(PO 4 ) 2 and they found that its X-ray-diffraction pattern is similar to those of other sodium and rare-earth double orthophosphates Na 3 RE(PO 4 ) 2 (RE = Y, La–Er). There are three types of Nd 3+ ions in Na 3 Nd(PO 4 ) 2 , respectively with 6-, 7-, and 8-fold coordination. In Na 3 RE(PO 4 ) 2 , although RE 3+ ions occupy different sites in this series of compounds [8,9,12,13] , we expect that the larger Ce 3+ ions prefers to occupy the large 8-fold coordinated site with C 1 symmetry in Na 3 Gd(PO 4 ) 2 . Since the bond lengths from Ce 3+ to O 2− are important to calculate the centroid shift of 5d orbital, we could use the bond length of Na 3 Nd(PO 4 ) 2 to estimate that of Na 3 Gd(PO 4 ) 2 . 3.2 The excitation and emission spectra of Na 3 Gd(PO 4 ) 2 Fig. 2 presents the VUV excitation spectrum (curve a, λ em = 313 nm) and the emission spectrum (curve b, λ e x = 172 nm) of Na 3 Gd(PO 4 ) 2 at RT. The dominant line-like absorptions around 274 nm and a series of weak sharp peaks near 196, 203, 247 and 254 nm are observed in (curve a), which correspond to the 8 S 7/2 → 6 G J , 6 D J , 6 I J transitions of Gd 3+ as marked in Fig. 2 . Additionally, broad bands below 207 nm with a maximum at 169 nm are clearly seen. The broad bands in this spectral range are attributed to the host-related absorptions [7–9] . Under 172 nm excitation, we recorded the emission spectrum of Na 3 Gd(PO 4 ) 2 as shown in Fig. 2 b, in which a peak at about 313 nm due to 6 P J → 8 S 7/2 transitions of Gd 3+ appears, indicating effective energy transfer from the host to Gd 3+ . The decay constant ( τ 1/ e ) of this transition is about 4.5 ms as shown in Fig. 2 c. 3.3 The luminescence of Na 3 Gd(PO 4 ) 2 :Ce 3+ under VUV–UV excitation The VUV excitation spectrum (a) of sample Na 3 Gd (1− x ) Ce x (PO 4 ) 2 ( x = 0.01) measured at 375 nm emission at 25 K is displayed in Fig. 3 I. Six broad bands labeled A, B, C, D, E and HA (∼164 nm) are observed in this spectrum. To estimate the wavelength maxima of these bands, curve a in the 185–350 nm spectral range is fitted by a sum of five Gaussian functions that peak at ∼321 nm (A), ∼292 nm (B), ∼260 nm (C), ∼246 nm (D), and ∼205 nm (E). These five broad bands are attributed to the transitions from the 4f ground state to crystal field split 5d states of Ce 3+ ion in the host lattice. The band HA peaking at ∼164 nm (7.55 eV) is the host-related absorption attributed to exciton creation. It is the same band as the 169 band in pure Na 3 Gd(PO 4 ) 2 but slightly shifted to shorter wavelengths which is quite normal when temperature decreases from RT to 25 K. A series of weak peaks at 194, 201, and 273 nm (see black arrows in the spectrum), respectively from 8 S 7/2 → 6 G J , 6 I J transitions of Gd 3+ , appear in the excitation spectrum evidencing that energy transfer from Gd 3+ to Ce 3+ takes place. The emission spectra of Ce 3+ in Na 3 Gd(PO 4 ) 2 were recorded under excitation at wavelengths of 165 (curve b) and 210 nm (curve c) at 25 K, as shown in Fig. 3 II. Two characteristic doublet emission bands of Ce 3+ are observed. To evaluate the emission peak positions for Ce 3+ , emission spectrum (c) was fitted with a sum of two Gaussian profiles yielding band maxima at ∼341 nm and ∼365 nm with an energy separation of 1928 cm −1 . According to above spectral information, we conclude that although Ce 3+ might occupy three different Gd 3+ sites, only one site is responsible for the main emission doublet. Spectra (a) and (b) in Fig. 4 show excitation spectra of Na 3 Gd (1− x ) Ce x (PO 4 ) 2 ( x = 0.04) at RT normalized to the maximum excitation intensities at 304 nm. We observe two nearly overlapping excitation spectra for the two emissions bands at 335 and 370 nm corresponding to the 5d → 2 F J ( J = 5/2, 7/2) emissions of Ce 3+ , see Fig. 3 IIb. The emission spectra normalized to their maximum emission intensities at 366 nm under 246 nm, 260 nm, 274 nm (the 6 I J – 8 S 7/2 transition of Gd 3+ ), 292 nm, 300 nm, and 321 nm are shown in curves c–g. Again, these spectra nearly overlap each another. Due to the higher Ce 3+ concentration and because of room temperature the excitation and emission spectra appear broader than those in Fig. 3 . Most noteworthy is the absence of the 321 nm band (band A in Fig. 3 Ia in the results of Fig. 4 ). By taking the 321 nm band or the 304 nm band as the lowest energy crystal field 5d component we would obtain a total crystal-field splitting (CFS) with the 205 nm band of 17,600 cm −1 or 15,900 cm −1 , respectively. Note that 17,000–18,000 cm −1 is a typical value for the crystal-field splitting at an 8-fold oxygen coordinated Gd 3+ site [14] . Two factors, the shape of the anion coordination polyhedron around Ce 3+ and the size of that polyhedron, have large influence on the crystal-field splitting. With fixed shape, the CFS decreases with the increase of the polyhedral size, i.e., the average bond length to the coordinating oxygen ligands. With fixed size, the CFS depends on the polyhedral shape. When we consider Na 3 Ln(PO 4 ) 2 (Ln = Gd, Nd) compounds share same structure and think that the Gd–O bond lengths are nearly same with those of Nd–O. In Na 3 Gd(PO 4 ) 2 we deal with a 8-fold irregular coordination with average bond length between Gd 3+ and O 2− of ∼248 pm. Here, it is interesting to compare the CFS of Ce 3+ 5d state in Na 3 Gd(PO 4 ) 2 with that in NaGd(PO 3 ) 4 where we found [15] a CFS value of ∼17.35 × 10 3 cm −1 for Ce 3+ on 8-fold Gd 3+ sites with C 1 symmetry and ∼240 pm average bond length. At this stage of our research both possible values of 17,600 cm −1 and 15,900 cm −1 for the crystal-field splitting are consistent with expectation. The centroid shift of the Ce 3+ 5d configuration is defined as the shift of the average energy of the five 4f–5d excitation bands of Ce in a compound from the value of 51,230 cm −1 pertaining to free (gaseous) Ce 3+ . Its value can be estimated from knowledge on the average Ce 3+ to oxygen bond length and a weighted average χ av of the electronegativity χ of the cations in that compound [16] . The eight phosphate compounds compiled in Ref. [16] show values for the centroid shift ranging from roughly 7000 cm −1 to 10,000 cm −1 with χ av ranging from 1.7 to 2.1. We found for the centroid shift in NaGd(PO 3 ) 4 a value of 7600 cm −1 [15] and from Ref. [17] a value of 8200 cm −1 is obtained for NaGdP 2 O 7 . In the series of compounds NaGd(PO 3 ) 4 , NaGdP 2 O 7 , and Na 3 Gd(PO 4 ) 2 the average electronegativity as defined in Ref. [16] decreases from 2.01 to 1.80 to 1.76. A centroid shift of 8000–9000 cm −1 is to be expected for Na 3 Gd(PO 4 ) 2 . If we take the bands A, B, C, D, and E in Fig. 3 Ia as the five crystal field split excitation bands of Ce on the 8-fold coordinated Gd-site, a centroid shift of 13,000 cm −1 is obtained. This value is much larger than the expected one suggesting that the bands A to E not all belong to one type of Ce center. We may speculate that the 205 nm (E-band) belongs to a doublet band and that the 321 nm band is from Ce 3+ on a different site. In that case the crystal field split states might be at 2 × 205, 246, 260, ∼300 nm giving a centroid shift of 9200 cm −1 which would agree much better with expectation. The above shows that assigning the excitation bands observed for NaGd(PO 4 ) 3 with three possible sites for Ce 3+ is not trivial. And although the emission seems to be dominated by one type of Ce 3+ site, the excitation spectra is not necessarily due to excitation bands belonging to one type of Ce 3+ site. Energy may be transferred from Ce 3+ on one site to Ce 3+ on another site, and also transfer can by via the Gd levels. More detailed experiments at low temperature and low Ce 3+ concentrations is required to further resolve the complex spectroscopy of Ce 3+ centers in NaGd(PO 4 ) 3 . In Fig. 3 II spectrum (b), the 6 P → 8 S emission of Gd 3+ at about 312 nm is seen on excitation at 165 nm. Apparently, like in the pure compound (see Fig. 2 ) there is energy transfer from the host to Gd 3+ . Upon 210 nm excitation, however, Gd 3+ emission is undetectable, showing that energy transfer from Ce 3+ to Gd 3+ is inefficient. Observation of the Gd 6 P → 8 S emission implies that energy transfer from the 6 P state to Ce 3+ is inefficient. Indeed in the excitation spectrum of Fig. 3 I the Gd 3+ 8 S → 6 P excitation line at 312 nm is missing whereas the higher energy Gd 3+ excitations are clearly observed. Energy transfer from the 6 P level of Gd 3+ to the 5d-state of the lower energy band A at 321 nm should be possible. The observation that the 6 P level of Gd 3+ does not sensitize the main Ce 3+ emission in Fig. 3 IIb indicates that excitation band A is not related to that main Ce 3+ emission. 3.4 The decay of Ce 3+ luminescence in Na 3 Gd(PO 4 ) 2 The decay curves of Ce 3+ in Na 3 Gd 0.99 Ce 0.01 (PO 4 ) 2 were measured at RT as shown in Fig. 5 . The four recorded curves nearly overlap each other and can all be well fitted by a single exponential equation, I t = I 0 exp(− t / τ ), where I and I 0 correspond to the luminescence intensity at time t and the initial intensity immediately after the exciting pulse, τ is the decay time constant. After fitting, we found τ ∼ 25.4 ns at 340 nm emission and ∼26.5 ns at 365 nm emission, respectively. If there is only one emitting type of Ce center, both decay times should be equal. The small difference may indicate emission from more than one Ce center. 3.5 Luminescence of Na 3 Gd(PO 4 ) 2 :Ce 3+ under X-ray excitation The X-ray excited emission spectra of sample Na 3 Gd 0.99 Ce 0.01 (PO 4 ) 2 and a BaF 2 reference sample with known light yield of 8800 ph/MeV of absorbed X-ray energy were recorded at RT under the same experimental conditions and are shown in Fig. 6 . Note that due to thermal broadening, the Ce 3+ emission at room temperature does not show the doublet structure as observed at 25 K in Fig. 3 but resembles the room temperature emission of Fig. 4 . Using equation N ph = ∫ I sample ( λ ) d λ ∫ I BaF 2 ( λ ) d λ · 8800 photons / MeV , the absolute light yield of sample Na 3 Gd 0.99 Ce 0.01 (PO 4 ) 2 is roughly estimated to be 17,000 ph/MeV. Here I sample and I BaF 2 are the emission intensities as function of wavelength for Na 3 Gd 0.99 Ce 0.01 (PO 4 ) 2 and for BaF 2 that were corrected for the detection efficiency of the equipment used. The yield of our phosphate sample is about two times larger than the BaF 2 reference sample. 4 Conclusions The investigations on the VUV–UV spectroscopic properties of Na 3 Gd(PO 4 ) 2 :Ce 3+ show that the main emission is from Ce 3+ ions on one type of Gd 3+ site in the lattice. Most likely this is the largest 8-fold coordinated Gd-site. The decay time constant measured at the 5d to 2 F 5/2 emission is 25.4 ns, and at the 5d to 2 F 7/2 emission at 365 nm is 26.5 ns. Excitation bands of Ce 3+ emission at 25 K are observed at ∼205, ∼246, ∼260, ∼292, and ∼321 nm. By analysis of the total crystal-field splitting and the centroid shift of the 5d excitation bands and by comparison with the spectroscopy of Ce 3+ in NaGd(PO 3 ) 4 and in NaGdP 2 O 7 , we conclude that these five bands cannot belong to one and the same Ce 3+ center. Most likely, a small but significant fraction of Ce 3+ resides on the Gd-sites with either 7-fold or 6-fold coordination. It will require detailed spectroscopic studies at low temperature and low Ce 3+ concentration to observe more details in excitation spectra and to arrive at an unambiguous assignment of excitation bands. For the sample Na 3 Gd 0.99 Ce 0.01 (PO 4 ) 2 the emission and light yield under X-ray excitation was studied. The sample emits about 17,000 photons/MeV of absorbed X-ray energy at RT. Acknowledgments The work is financially supported by the National Basic Research Program of China (973 Program, Grant No. 2007CB935502 ), by the National Natural Science Foundation of China ( Grant Nos. 20871121 and 10979027 ), and by the Natural Science Foundation of Guangdong Province ( Grant No. 9151027501000003 ). References [1] L. Li H.B. Liang Z.F. Tian H.H. Lin Q. Su G.B. Zhang J. Phys. Chem. C 112 2008 13763 [2] H.H. Lin H.B. Liang Z.F. Tian B. Han J. Wang Q. Su G.B. Zhang J. Phys. D: Appl. Phys. 42 2009 165409 [3] W.P. Chen H.B. Liang B. Han J.P. Zhong Q. Su J. Phys. Chem. C 113 2009 17194 [4] S. Yokono T. Abe T. Hoshina J. Lumin. 24–25 1981 309 [5] K.W. Krämer P. Dorenbos H.U. Güdel C.W.E. van Eijk J. Mater. Chem. 16 2006 2773 [6] J.P. Zhong H.B. Liang Q. Su J.Y. Zhou I.V. Khodyuk P. Dorenbos Opt. Mater. 32 2009 378 [7] L. Schwarz B. Finke M. Kloss A. Rohmann U. Sasum D. Haberland J. Lumin. 72–74 1997 257 [8] M. Kloss B. Finke L. Schwarz D. Haberland J. Lumin. 72–74 1997 684 [9] J. Legendziewicz M. Guzik J. Cybińska Opt. Mater. 31 2009 567 [10] M. Kizilyalli A. Sungur D.S. Jones J. Less Common Met. 110 1985 249 [11] M. Fang W.-D. Cheng H. Zhang D. Zhao W.-L. Zhang S.-L. Yang J. Solid State Chem. 181 2008 2165 [12] R. Salmon C. Parent M. Vlasse G.L. Flem Mater. Res. Bull. 13 1978 439 [13] M. Guzik T. Aitasalo W. Szuszkiewicz J. Hölsä B. Keller J. Legendziewicz J. Alloys Compd. 380 2004 368 [14] P. Dorenbos Phys. Rev. B 64 2001 125117 [15] J.P. Zhong H.B. Liang H.H. Lin B. Han Q. Su G.B. Zhang J. Mater. Chem. 17 2007 4679 [16] P. Dorenbos Phys. Rev. B 65 2002 235110 [17] J.L. Yuan J. Wang D.B. Xiong J.T. Zhao Y.B. Fu G.B. Zhang C.S. Shi J. Lumin. 126 2007 717