- Email: [email protected]
Up-conversion processes in NaLaF4:Er3+
Optical Materials 31 (2009) 1517–1524
Contents lists available at ScienceDirect
Optical Materials journal homepage: www.elsevier.com/locate/optmat
Up-conversion processes in NaLaF4:Er3+ A. Sarakovskis *, J. Grube, A. Mishnev, M. Springis Institute of Solid State Physics, University of Latvia, Kengaraga Str. 8, LV-1063 Riga, Latvia
a r t i c l e
i n f o
Article history: Received 15 May 2008 Received in revised form 16 February 2009 Accepted 17 February 2009 Available online 21 March 2009 PACS: 78.20.e 78.47.+p 63.20.e 42.70.a
a b s t r a c t Structural and spectroscopic investigation of NaLaF4:Er3+ material at different doping concentrations is presented. X-ray diffraction patterns, up-conversion luminescence spectra and decay curves for 2 H9/2 ? 4I15/2, 4S3/2 ? 4I15/2 and 4F9/2 ? 4I15/2 optical transitions in the material are shown and possible excitation routes are discussed. Raman spectrum for the undoped material is presented and the effective phonon energy of the material is estimated. Based on the obtained results application of rare-earth doped NaLaF4 in the field of up-conversion phosphors is evaluated. Ó 2009 Elsevier B.V. All rights reserved.
Keywords: Up-conversion phosphors Time-resolved luminescence Raman spectroscopy Rare-earth ions
1. Introduction For many years rare-earth (RE) doped materials have been broadly used in a variety of applications. Owing to their abundant structure of energy levels many fields of science and technology like solid-state lasers, colour displays, optical telecommunication utilize the unique properties of these materials [1–3]. For some decades much attention has been paid for studies of up-conversion (UC) processes in RE doped materials, which involve the absorption of several photons (usually IR) and subsequent emission of one single photon (VIS or UV) [4]. UC effect could be used in different areas like solid-state lasers [5], temperature sensors [6], white light simulation [7], improvement of the efficiency of solar cells [8] and others. The efficiency of the UC process strongly depends on the rate of radiationless transitions within RE ion and therefore on the phonon energy of the medium: the higher the phonon energy (e.g. the higher rate of radiationless transitions), the lower the efficiency of UC. Most fluoride crystals (LaF3 [9], PbF2 [10], CaF2 [11], SrF2, BaF2 [12]) are attractive hosts for RE ions either due to their low phonon energy or multisite structure. Particular interest has been paid to RE doped NaYF4, which has been widely studied
* Corresponding author. Tel.: +371 67 187 471; fax: +371 67 132 778. E-mail address: anatoly@cfi.lu.lv (A. Sarakovskis). 0925-3467/$ - see front matter Ó 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.optmat.2009.02.015
[13–17]. In [13] NaYF4:Er3+, Yb3+, Tm3+ is claimed to be one of the most efficient UC phosphors. The high efficiency is explained partly because of the small effective phonon energy of the medium (360 cm1 [13]) and partly due to the multisite nature of NaYF4 crystalline lattice, meaning that RE ions can occupy various nonequivalent states in the crystalline lattice [18]. The chemical composition and the structure of NaYF4 crystal are similar to those of NaLaF4 [19], predicting efficient UC processes in RE doped NaLaF4. The main goal of this report is to gain first insight into the optical properties of NaLaF4:Er3+ and to estimate possible applicability of this material in the field of UC phosphors. To do this, the UC processes in NaLaF4 at different Er3+doping levels were studied. The luminescence decay profiles for the main optical transitions of Er3+ in NaLaF4 were measured by means of time-resolved spectroscopy and the effective phonon energy of NaLaF4 was determined by Raman spectroscopy. The obtained results are compared to those previously reported for NaYF4. To the best of our knowledge such survey of Er3+ doped NaLaF4 has not yet been reported. 2. Experimental Erbium doped polycrystalline NaLaF4 were synthesized from 65NaF–35LaF3–xErF3 (x = 0.05–10) or 65NaF–10LaF3–25ErF3 melt (in mol%). For each batch about 5 g of the fluoride components were fully mixed and melted in a corundum crucible at 900 °C in
1518
A. Sarakovskis et al. / Optical Materials 31 (2009) 1517–1524
air. After 1 h the furnace was switched off and the batches were slowly cooled there for about 2 h. The obtained polycrystalline materials were polished to form square coupons (5 5 2 mm) and further their structure and luminescence properties were studied. The same technique was used to synthesize NaErF4 crystalline material (50NaF–50ErF3). The structure of the samples was checked by X-ray diffraction analysis (XRD) (XPert Pro MPD, Cu anode, operating at 40 kV 30 mA). The stationary UC luminescence was excited by a temperature controlled laser diode (kem = 980 nm, variable radiant power up to 300 mW) from Thorlabs. The time-resolved luminescence was excited by a tunable pulsed solid-state laser NT342/3UV (pulse duration 4 ns) from Ekspla. Kinetics were measured by Andor SR-303i-B spectrometer coupled to a photomultiplier tube (time resolution better than 20 ns) and digital oscilloscope Tektronix TDS 684A. The same spectrometer coupled to a CCD camera (Andor DU-401-BV) was used for the studies of the stationary luminescence. Raman measurements were performed using SPEX-RAMALOG Laser Raman Spectrometer system (resolution 0.15 cm1) equipped with laser MLL-671 nm–300 mW from CNI laser. All the measurements were conducted at room temperature. 3. Results and discussion The structure of the obtained specimens was studied by XRD analysis. XRD pattern typical of samples at low Er3+ concentrations is shown in Fig. 1. The pattern is governed by the peaks related to NaLaF4 crystalline phase (JCPDS No. 50-0155). Some peaks of weaker intensity related to residual NaF are also present in the pattern. Indeed according to the phase diagram [20,21] if NaF–LaF3 melt is cooled from 900 °C down to 734 °C about 24 mol% of NaLaF4 is crystallized with the respect to the initial melt composition. Below the temperature of T = 734 °C, which corresponds to the eutectic temperature of the melt, additional amount of 56 mol% of NaLaF4 is formed and about 20 mol% of NaF remains in the melt. So the final composition of the synthesized polycrystalline material is expected to be 80 mol% NaLaF4 (24 mol% crystallites and 56 mol% fine-grained crystalline eutectic) and 20 mol% NaF. The small addition of ErF3
should not change the phase composition of the material, because at low concentrations Er3+ ions would rather incorporate into NaLaF4 crystalline phase than form its own phase thus doped NaLaF4:Er3+ is expected to appear. At higher Er3+ concentration a competition between La3+ and Er3+ in the melt might take place to form either NaLaF4:Er3+ or NaErF4. For better understanding of the possible variations in the phase composition at different ErF3 content, a relevant part of the XRD pattern was selected and studied in detail (Fig. 2). The positions of the peaks corresponding to NaLaF4 crystalline phase are slightly shifted to higher degrees, when Er3+ content in the initial melt is raised. Further increase of ErF3 concentration leads not only to the shift of the corresponding peaks to higher degrees but also to the appearance of a shoulder at the right side of the peaks. Moreover at even higher relative concentration of ErF3 in the melt (NaF(65%)LaF3(25%)ErF3(10%)) the peaks corresponding to NaLaF4 crystalline phase are broadened and shifted to even higher degrees. As for the samples where LaF3 is absent (NaF(65%)ErF3(35%)) the observed peaks are identified as NaErF4 (JCPDS No. 27-0689). The shift of the peaks observed at low concentrations of ErF3 could be explained by the fact that the radius of Er3+ (88 pm) ion is smaller than that of La3+ (106 pm) ion, which means that the incorporation of activator ions in NaLaF4 lattice would lead to the decrease of the lattice parameter of the latter, and the corresponding positions of XRD peaks would shift to higher degrees. This fact signifies that at low concentrations Er3+ ions mostly incorporate into NaLaF4 crystalline phase and NaLaF4:Er3+ is formed. The further shift and broadening of the peaks observed at higher concentrations of ErF3 could be explained by the formation of NaLaF4 crystallites with even higher concentration of Er3+ adjacent to much distorted NaErF4 crystalline phase. It is possible because the crystalline structures of NaLaF4 and NaErF4 are identical and the lattice parameters are close (NaLaF4–(6.178, 6.178, 3.828), NaErF4–(5.959, 5.959, 3.514)). No changes in the position or shape of the peaks corresponding to NaF crystalline phase were noticed when Er3+ content in the melt was varied. This is an important fact because it might suggest that NaF remaining in the material after melting is unlikely to be doped and its influence on the luminescent properties of the material could be negligible. To sum up
Fig. 1. XRD pattern of 1% Er3+ doped NaLaF4 polycrystalline material. The most intense peaks of NaLaF4 are indexed.
A. Sarakovskis et al. / Optical Materials 31 (2009) 1517–1524
1519
Fig. 2. XRD pattern of NaLaF4:Er3+ material for samples at different LaF3 and ErF3 contents.
the information gained from the XRD analysis of the samples: at low concentrations of ErF3 in the initial melt NaLaF4:Er3+ is formed, some residual NaF is also observed. At higher concentrations of ErF3 in the initial melt not only the relative concentration of Er3+ in NaLaF4 continues to increase but also strongly distorted Na(LaEr)F4 compound tending towards NaErF4 crystalline phase is formed. The Raman spectrum measured for the undoped NaLaF4 polycrystalline material is shown in Fig. 3. The measurements for the doped specimens showed presence of the strong luminescence signal even at the lowest doping levels
of Er3+ making the detection of the Raman signal impossible, therefore the undoped NaLaF4 was chosen for Raman measurements. Lorentzian fit of the spectrum revealed five phonon bands centred at 232, 260, 306, 331 and 365 cm1. These bands can be assigned to NaLaF4 crystalline phase since NaF is not active in the Raman spectroscopy and cannot contribute to the spectrum. The positions of the bands are in a good accordance with previously reported polarized Raman spectrum for NaLaF4 single crystal [19]. Taking into account the positions, FWHMs and the intensities of the corresponding Lorentzian sub-bands the effective phonon energy for NaLaF4 could be estimated:
Fig. 3. Raman spectrum of undoped NaLaF4 sample.
1520
A. Sarakovskis et al. / Optical Materials 31 (2009) 1517–1524
P
Eph
Ei wi Ai ¼ P 290 cm1 ; wi Ai i
i
where Ei, wi and Ai are the position, FWHM and relative intensity of each Lorentzian sub-band, respectively. The obtained value of the effective phonon energy, which to the utmost extent defines the rate of radiationless transitions in a material, is even lower than that previously reported for NaYF4 (360 cm1) [13] making NaLaF4 a very promising candidate in the field of UC phosphors. The UC luminescence spectrum for the sample with Er3+ concentration 2 mol% excited at k = 980 nm is shown in Fig. 4. The luminescence bands corresponding to the characteristic transitions in Er3+ ion
are present in the UC luminescence spectrum: violet (2H9/2 ? I15/2), green (2H11/2 ? 4I15/2 and 4S3/2 ? 4I15/2) and red (4F9/2 ? 4 I15/2) [13]. It is known that in UC processes the emission intensity IUC has a power law dependence on the excitation power Pexc:IUC (Pexc)n, where n is a number of photons required to excite the emitting state [22]. In Fig. 5 the dependence of the UC luminescence intensities measured for the main radiative transitions of Er3+ at 410, 540 and 665 nm versus excitation power are shown. The plots in the double logarithmic chart show linear behaviour of the UC luminescence intensity versus excitation power having parameters of the slopes close to 2 for the 540 nm and 665 nm bands. This is an indication that the 4S3/2 and 4F9/2 levels are 4
Fig. 4. Stationary UC luminescence spectrum of 2% Er3+ doped NaLaF4 excited at 980 nm measured at room temperature.
Fig. 5. Power dependence of the transitions 2H9/2 ? 4I15/2 (squares), 4S3/2 ? 4I15/2 (circles) and F9/2 ? 4I15/2 (triangles) in NaLaF4:Er3+ measured at room temperature.
A. Sarakovskis et al. / Optical Materials 31 (2009) 1517–1524
populated by two photon processes, while the slope of 2.98 for the band at 410 nm signifies the involvement of three-photon absorption process in the creation of the violet UC band. The possible excitation routes of the UC luminescence for Er3+ ion are shown in Fig. 6. The appearance of the green luminescence band at 540 nm might be a result of an excited state absorption process (ESA) i.e. sequential two photon absorption process: 4I15/2 + hm ? 4I11/2 + hm ? 4F7/2. Afterwards the relaxation to 2H11/2 and 4S3/2 populates the emitting levels for the green luminescence. Another route to populate the 4S3/2 level is by energy transfer (ET) process: (4I11/2,
1521
4
I11/2) ? (4I15/2, 4F7/2) (ET1). The involvement of the both ESA and ET processes is evident in the time-resolved luminescence decay plot for the green luminescence band excited at 980 nm (Fig. 7a). It can be seen that the luminescence decay curve has a complex nature: it has an intensity offset at t = 0 and also the rise part of about 5 ls signifying the involvement of the both processes in the creation of the luminescence. The appearance of the red luminescence at 660 nm might be caused by an ET process: (4I13/2, 4I11/2) ? (4I15/2, 4F9/2) (ET2). Clear evidence of pure ET process is a time-resolved luminescence decay plot for the red band excited at 980 nm (Fig. 7b) – it can be seen
Fig. 6. Schematic energy level diagram and the main radiative transitions in Er3+ ion. Full and dashed arrows are radiative and non-radiative transitions, respectively.
Fig. 7. Decay profiles for (a) 4S3/2 ? 4I15/2, (b) 4F9/2 ? 4I15/2, (c) 2H9/2 ? 4I15/2 optical transitions in NaLaF4:Er3+(0.5 mol%) excited at 980 nm measured at room temperature.
1522
A. Sarakovskis et al. / Optical Materials 31 (2009) 1517–1524
that the luminescence signal does not have any intensity offset at t = 0. The appearance of the violet band could be explained by an ET process (4S3/2, 4I13/2) ? (2H9/2, 4I15/2) (ET3). Again the ET process is supported by the luminescence decay kinetics excited at 980 nm (Fig. 7c) – no intensity offset is present at t = 0. In Fig. 8 the UC luminescence spectra for the samples at different doping levels are shown. It can be seen that the increase of Er3+ doping level leads to the decrease of the green-to-red bands intensity ratio. It could be explained by an interionic cross-relaxation process, which becomes dominant at higher Er3+ concentrations, when the average distance between the Er3+ ions diminishes. On one hand the cross-relaxa-
tion process depopulates 4S3/2 level (4S3/2, 4I15/2) ? (4I9/2, 4I13/2) (in Fig. 6 CR) thus decreasing the intensity of the green band. On the other hand it provides an extra route for 4F9/2 level population: (4I13/2, 4I11/2) ? (4I15/2, 4F9/2) (ET2) increasing the population of 4F9/2 level and the intensity of the red band. Another important parameter, which is essential for the characterization of the optical properties of the material, is the decay profile of the luminescence under direct excitation of the emitting level. In this work we have performed time-resolved luminescence measurements for the green (4S3/2 ? 4I15/2) and red (4F9/2 ? 4I15/2) luminescence bands at direct excitation not higher than 300 cm1 above the respective emitting level. The luminescence kinetics
Fig. 8. Stationary UC luminescence spectra of Er3+ doped NaLaF4 samples at different doping levels excited at 980 nm measured at room temperature.
Fig. 9. Decay profiles for the 4S3/2 ? 4I15/2 optical transition in NaLaF4 at different Er3+ concentrations: direct excitation at room temperature. Inset – double exponential fit of the decay for the sample with 4 mol% Er3+.
1523
A. Sarakovskis et al. / Optical Materials 31 (2009) 1517–1524
Fig. 10. Decay profiles for the 4F9/2 ? 4I15/2 optical transition in NaLaF4 at different Er3+ concentrations: direct excitation at room temperature. Inset – double exponential fit of the decay for the sample with 4 mol% Er3+.
Table 1 Lifetimes (in ls) of the main optical transitions in Er3+ doped NaLaF4 samples at different doping levels and NaErF4, measured under direct excitation at room temperature. The lifetime of NaYF4:Er3+ 2 mol% is given for the comparison [13].
4
4
4
4
S3/2 ; I15/2 F9/2 ; I15/2
Fast Slow Fast Slow
0.05%
0.1%
0.5%
1%
2%
4%
10%
NaErF4
NaYF4 2% [13]
50 510 110 530
40 490 100 520
35 470 90 510
30 410 85 480
25 350 70 450
20 240 60 390
7 80 30 230
0.275
360
5
430
related to the transitions from the different Er3+ energy levels to the ground state are shown in Figs. 9 and 10. It can be seen in the semilogarithmic chart that the luminescence decay curves for the green band (Fig. 9) contain more than one component. Exponential fit (Fig. 9 inset) shows that the experimental curve can be approximated by two exponents: ‘‘fast” and ‘‘slow”. At highest doping levels the slow component almost disappears and only the fast one remains. Similar tendency can be observed for the red luminescence band (Fig. 10), which also seems to be built of two components (Fig. 10 inset), although the concentration impact on the lifetime value of the slow component is less pronounced compared to the green band. The values of the lifetimes for the both luminescence bands at different doping levels are summarized in Table 1. The presence of the fast and slow components in the luminescence decay plots at direct excitation could be explained by the formation of several crystalline phases in the polycrystalline material as it was seen in the XRD patterns (Figs. 1 and 2). Since the decay profiles of the green luminescence at low Er3+ concentrations are mostly governed by a single slow exponent and XRD analysis reveals mainly the presence of NaLaF4:Er3+ in the samples, this slow exponential component could be attributed to NaLaF4:Er3+. It is also supported by the fact that the obtained lifetime value of the slow exponent is similar to that of NaYF4 (Table 1), which has similar structure and close effective phonon energy as NaLaF4 does. At higher doping levels the value of the slow component for the green luminescence band in the exponential fit decreases. This might be due to the cross-relaxation processes between Er3+ ions in NaLaF4,
in which the emitting 4S3/2 level is non-radiatively depopulated (cross-relaxation (4S3/2, 4I15/2) ? (4I9/2, 4I13/2)). The appearance of the fast component in the exponential decay of the green luminescence band and the growth of its relative intensity along with the increase of Er3+ concentration in NaLaF4 correlates with the XRD observations (Fig. 2) where the increase of ErF3 amount in the initial melt led to the shift of the corresponding XRD peaks towards those attributable to NaErF4 crystalline phase. This fact suggests that the fast exponential component in the decay curve of the green luminescence band could be attributed to the distorted NaErF4 crystalline phase. Nevertheless additional experiments are needed to elucidate the problem of the coexisting of the fast and slow components in the decay profiles. 4. Conclusions Erbium doped polycrystalline NaLaF4 material has been synthesized. The XRD patterns measured for NaLaF4 material at different doping levels of ErF3 showed considerable lattice distortion of NaLaF4 when Er3+ content was increased. The UC luminescence spectra measured for the samples at different doping levels of Er3+ revealed the characteristic luminescence bands in the violet (2H9/2 ? 4I15/2), green (2H11/2 ? 4I15/2 and 4S3/2 ? 4I15/2) and red (4F9/2 ? 4I15/2) spectral regions. The luminescence decay curves originating from 4S3/2 and 4F9/2 levels of Er3+ ions in NaLaF4 crystalline structure have been measured the lifetimes of the corresponding levels were determined. The effective phonon energy of NaLaF4 derived from Raman measurements for the undoped material was
1524
A. Sarakovskis et al. / Optical Materials 31 (2009) 1517–1524
estimated Eph 290 cm1. It is slightly lower than the one reported for NaYF4 suggesting that the potential efficiency of UC processes in NaLaF4 could be comparable to that in NaYF4. The obtained results imply that NaLaF4 crystalline phase properly synthesized and doped is a promising media in the field of UC phosphors. The work related to the synthesis of NaLaF4 crystallites free from other crystalline phases is under progress and the results will be published shortly. Acknowledgments This work was financially supported by National Research Programme of Latvia in Materials Science. The financial support of ESF is greatly acknowledged. The authors would like to thank Dr. A. Veispals for the help in the sample preparation, Dr. G. Chikvaidze for Raman measurements and Prof. U. Rogulis for fruitful discussions. References [1] J. Mendez-Ramos, V.K. Tikhomirov, V.D. Rodriguez, D. Furniss, Journal of Alloys and Compounds 440 (1–2) (2007) 328. [2] T. Schweizer, D.W. Hewak, D.N. Payne, T. Jensen, G. Huber, Electronics Letters 32 (7) (1996) 666. [3] D.N. Payne, L. Reekie, in: Proceedings Fourteenth European Conference on Optical Communication, Brighton 1988, vol. 1, published by Institution of Electrical Engineers, London, 1988, p. 49.
[4] F. Auzel, Chemical Review 2004 104 (2004) 139. [5] F. Tong, W.P. Risk, R.M. Macfarlane, W. Lenth, Electronics Letters 25 (20) (1989) 1389. [6] B. Dong, T. Yang, M.K. Lei, Sensors and Actuators B 123 (2007) 667. [7] J.E.C. Silva, G.F. de Sa, P.A. Santa-Cruz, Journal of Alloys and Compounds 323– 324 (2001) 336. [8] A. Shalav, B.S. Richards, M.A. Green, Solar Energy Materials and Solar Cells 91 (2007) 829. [9] A.C. Yanes, J. Del-Castillo, J. Mendez-Ramos, V.D. Rodriguez, M.E. Torres, J. Arbiol, Optical Materials 29 (8) (2007) 999. [10] M. Mortier, G. Patriarche, Optical Materials 28 (2006) 1401. [11] D. Chen, Y. Wang, E. Ma, Y. Yu, F. Liu, Optical Materials 29 (12) (2007) 1693. [12] Yu.V. Orlovskii, T.T. Basiev, K.K. Pukhov, I.N. Vorobiev, A.G. Papashvili, F. Pelle, V.V. Osiko, Journal of Luminescence 94–95 (2001) 791. [13] J.F. Suyver, J. Grimm, M.K. van Veen, D. Biner, K.W. Krämer, H.U. Güdel, Journal of Luminescence 117 (1) (2006) 1. [14] F. Liu, Y. Wang, D. Chen, Y. Yu, Materials Science and Engineering: B 136 (2–3) (2007) 106. [15] Y. Wei, F. Lu, X. Zhang, D. Chen, Journal of Alloys and Compounds 455 (1–2) (2008) 376. [16] K.W. Krämer, D. Biner, G. Frei, H.U. Güdel, M.P. Hehlen, S.R. Lüthi, Chemistry of Materials 16 (2004) 1244. [17] A. Aebischer, M. Hostettler, J. Hauser, K. Krämer, T. Weber, H.U. Güdel, H.-B. Bürgi, Angewandte Chemie International Edition 45 (2006) 2802. [18] A. Grzechnik, P. Bouvier, M. Mezouar, M.D. Mathews, A.K. Tyagi, J. Köhler, Journal of Solid State Chemistry 165 (1) (2002) 159. [19] M.M. Lage, F.M. Matinaga, J.-Y. Gesland, R.L. Moreira, Journal of Applied Physics 99 (5) (2006) 053510. [20] R.E. Thoma, H. Insley, G.M. Hebert, Inorganic Chemistry 5 (7) (1966) 1222. [21] P.P. Fedorov, I.I. Buchinskaya, O.S. Bondareva, A.A. Bystrova, L.L. Vistin, D.A. Ershov, S.P. Ivanov, V.A. Stasyuk, B.P. Sobolev, Russian Journal of Inorganic Chemistry 45 (6) (2000) 949. [22] M. Polnau, D.R. Gamelin, S.R. Lüthi, H.U. Güdel, M.P. Hehlen, Physical Review B 61 (5) (2000) 3337.
Contents lists available at ScienceDirect
Optical Materials journal homepage: www.elsevier.com/locate/optmat
Up-conversion processes in NaLaF4:Er3+ A. Sarakovskis *, J. Grube, A. Mishnev, M. Springis Institute of Solid State Physics, University of Latvia, Kengaraga Str. 8, LV-1063 Riga, Latvia
a r t i c l e
i n f o
Article history: Received 15 May 2008 Received in revised form 16 February 2009 Accepted 17 February 2009 Available online 21 March 2009 PACS: 78.20.e 78.47.+p 63.20.e 42.70.a
a b s t r a c t Structural and spectroscopic investigation of NaLaF4:Er3+ material at different doping concentrations is presented. X-ray diffraction patterns, up-conversion luminescence spectra and decay curves for 2 H9/2 ? 4I15/2, 4S3/2 ? 4I15/2 and 4F9/2 ? 4I15/2 optical transitions in the material are shown and possible excitation routes are discussed. Raman spectrum for the undoped material is presented and the effective phonon energy of the material is estimated. Based on the obtained results application of rare-earth doped NaLaF4 in the field of up-conversion phosphors is evaluated. Ó 2009 Elsevier B.V. All rights reserved.
Keywords: Up-conversion phosphors Time-resolved luminescence Raman spectroscopy Rare-earth ions
1. Introduction For many years rare-earth (RE) doped materials have been broadly used in a variety of applications. Owing to their abundant structure of energy levels many fields of science and technology like solid-state lasers, colour displays, optical telecommunication utilize the unique properties of these materials [1–3]. For some decades much attention has been paid for studies of up-conversion (UC) processes in RE doped materials, which involve the absorption of several photons (usually IR) and subsequent emission of one single photon (VIS or UV) [4]. UC effect could be used in different areas like solid-state lasers [5], temperature sensors [6], white light simulation [7], improvement of the efficiency of solar cells [8] and others. The efficiency of the UC process strongly depends on the rate of radiationless transitions within RE ion and therefore on the phonon energy of the medium: the higher the phonon energy (e.g. the higher rate of radiationless transitions), the lower the efficiency of UC. Most fluoride crystals (LaF3 [9], PbF2 [10], CaF2 [11], SrF2, BaF2 [12]) are attractive hosts for RE ions either due to their low phonon energy or multisite structure. Particular interest has been paid to RE doped NaYF4, which has been widely studied
* Corresponding author. Tel.: +371 67 187 471; fax: +371 67 132 778. E-mail address: anatoly@cfi.lu.lv (A. Sarakovskis). 0925-3467/$ - see front matter Ó 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.optmat.2009.02.015
[13–17]. In [13] NaYF4:Er3+, Yb3+, Tm3+ is claimed to be one of the most efficient UC phosphors. The high efficiency is explained partly because of the small effective phonon energy of the medium (360 cm1 [13]) and partly due to the multisite nature of NaYF4 crystalline lattice, meaning that RE ions can occupy various nonequivalent states in the crystalline lattice [18]. The chemical composition and the structure of NaYF4 crystal are similar to those of NaLaF4 [19], predicting efficient UC processes in RE doped NaLaF4. The main goal of this report is to gain first insight into the optical properties of NaLaF4:Er3+ and to estimate possible applicability of this material in the field of UC phosphors. To do this, the UC processes in NaLaF4 at different Er3+doping levels were studied. The luminescence decay profiles for the main optical transitions of Er3+ in NaLaF4 were measured by means of time-resolved spectroscopy and the effective phonon energy of NaLaF4 was determined by Raman spectroscopy. The obtained results are compared to those previously reported for NaYF4. To the best of our knowledge such survey of Er3+ doped NaLaF4 has not yet been reported. 2. Experimental Erbium doped polycrystalline NaLaF4 were synthesized from 65NaF–35LaF3–xErF3 (x = 0.05–10) or 65NaF–10LaF3–25ErF3 melt (in mol%). For each batch about 5 g of the fluoride components were fully mixed and melted in a corundum crucible at 900 °C in
1518
A. Sarakovskis et al. / Optical Materials 31 (2009) 1517–1524
air. After 1 h the furnace was switched off and the batches were slowly cooled there for about 2 h. The obtained polycrystalline materials were polished to form square coupons (5 5 2 mm) and further their structure and luminescence properties were studied. The same technique was used to synthesize NaErF4 crystalline material (50NaF–50ErF3). The structure of the samples was checked by X-ray diffraction analysis (XRD) (XPert Pro MPD, Cu anode, operating at 40 kV 30 mA). The stationary UC luminescence was excited by a temperature controlled laser diode (kem = 980 nm, variable radiant power up to 300 mW) from Thorlabs. The time-resolved luminescence was excited by a tunable pulsed solid-state laser NT342/3UV (pulse duration 4 ns) from Ekspla. Kinetics were measured by Andor SR-303i-B spectrometer coupled to a photomultiplier tube (time resolution better than 20 ns) and digital oscilloscope Tektronix TDS 684A. The same spectrometer coupled to a CCD camera (Andor DU-401-BV) was used for the studies of the stationary luminescence. Raman measurements were performed using SPEX-RAMALOG Laser Raman Spectrometer system (resolution 0.15 cm1) equipped with laser MLL-671 nm–300 mW from CNI laser. All the measurements were conducted at room temperature. 3. Results and discussion The structure of the obtained specimens was studied by XRD analysis. XRD pattern typical of samples at low Er3+ concentrations is shown in Fig. 1. The pattern is governed by the peaks related to NaLaF4 crystalline phase (JCPDS No. 50-0155). Some peaks of weaker intensity related to residual NaF are also present in the pattern. Indeed according to the phase diagram [20,21] if NaF–LaF3 melt is cooled from 900 °C down to 734 °C about 24 mol% of NaLaF4 is crystallized with the respect to the initial melt composition. Below the temperature of T = 734 °C, which corresponds to the eutectic temperature of the melt, additional amount of 56 mol% of NaLaF4 is formed and about 20 mol% of NaF remains in the melt. So the final composition of the synthesized polycrystalline material is expected to be 80 mol% NaLaF4 (24 mol% crystallites and 56 mol% fine-grained crystalline eutectic) and 20 mol% NaF. The small addition of ErF3
should not change the phase composition of the material, because at low concentrations Er3+ ions would rather incorporate into NaLaF4 crystalline phase than form its own phase thus doped NaLaF4:Er3+ is expected to appear. At higher Er3+ concentration a competition between La3+ and Er3+ in the melt might take place to form either NaLaF4:Er3+ or NaErF4. For better understanding of the possible variations in the phase composition at different ErF3 content, a relevant part of the XRD pattern was selected and studied in detail (Fig. 2). The positions of the peaks corresponding to NaLaF4 crystalline phase are slightly shifted to higher degrees, when Er3+ content in the initial melt is raised. Further increase of ErF3 concentration leads not only to the shift of the corresponding peaks to higher degrees but also to the appearance of a shoulder at the right side of the peaks. Moreover at even higher relative concentration of ErF3 in the melt (NaF(65%)LaF3(25%)ErF3(10%)) the peaks corresponding to NaLaF4 crystalline phase are broadened and shifted to even higher degrees. As for the samples where LaF3 is absent (NaF(65%)ErF3(35%)) the observed peaks are identified as NaErF4 (JCPDS No. 27-0689). The shift of the peaks observed at low concentrations of ErF3 could be explained by the fact that the radius of Er3+ (88 pm) ion is smaller than that of La3+ (106 pm) ion, which means that the incorporation of activator ions in NaLaF4 lattice would lead to the decrease of the lattice parameter of the latter, and the corresponding positions of XRD peaks would shift to higher degrees. This fact signifies that at low concentrations Er3+ ions mostly incorporate into NaLaF4 crystalline phase and NaLaF4:Er3+ is formed. The further shift and broadening of the peaks observed at higher concentrations of ErF3 could be explained by the formation of NaLaF4 crystallites with even higher concentration of Er3+ adjacent to much distorted NaErF4 crystalline phase. It is possible because the crystalline structures of NaLaF4 and NaErF4 are identical and the lattice parameters are close (NaLaF4–(6.178, 6.178, 3.828), NaErF4–(5.959, 5.959, 3.514)). No changes in the position or shape of the peaks corresponding to NaF crystalline phase were noticed when Er3+ content in the melt was varied. This is an important fact because it might suggest that NaF remaining in the material after melting is unlikely to be doped and its influence on the luminescent properties of the material could be negligible. To sum up
Fig. 1. XRD pattern of 1% Er3+ doped NaLaF4 polycrystalline material. The most intense peaks of NaLaF4 are indexed.
A. Sarakovskis et al. / Optical Materials 31 (2009) 1517–1524
1519
Fig. 2. XRD pattern of NaLaF4:Er3+ material for samples at different LaF3 and ErF3 contents.
the information gained from the XRD analysis of the samples: at low concentrations of ErF3 in the initial melt NaLaF4:Er3+ is formed, some residual NaF is also observed. At higher concentrations of ErF3 in the initial melt not only the relative concentration of Er3+ in NaLaF4 continues to increase but also strongly distorted Na(LaEr)F4 compound tending towards NaErF4 crystalline phase is formed. The Raman spectrum measured for the undoped NaLaF4 polycrystalline material is shown in Fig. 3. The measurements for the doped specimens showed presence of the strong luminescence signal even at the lowest doping levels
of Er3+ making the detection of the Raman signal impossible, therefore the undoped NaLaF4 was chosen for Raman measurements. Lorentzian fit of the spectrum revealed five phonon bands centred at 232, 260, 306, 331 and 365 cm1. These bands can be assigned to NaLaF4 crystalline phase since NaF is not active in the Raman spectroscopy and cannot contribute to the spectrum. The positions of the bands are in a good accordance with previously reported polarized Raman spectrum for NaLaF4 single crystal [19]. Taking into account the positions, FWHMs and the intensities of the corresponding Lorentzian sub-bands the effective phonon energy for NaLaF4 could be estimated:
Fig. 3. Raman spectrum of undoped NaLaF4 sample.
1520
A. Sarakovskis et al. / Optical Materials 31 (2009) 1517–1524
P
Eph
Ei wi Ai ¼ P 290 cm1 ; wi Ai i
i
where Ei, wi and Ai are the position, FWHM and relative intensity of each Lorentzian sub-band, respectively. The obtained value of the effective phonon energy, which to the utmost extent defines the rate of radiationless transitions in a material, is even lower than that previously reported for NaYF4 (360 cm1) [13] making NaLaF4 a very promising candidate in the field of UC phosphors. The UC luminescence spectrum for the sample with Er3+ concentration 2 mol% excited at k = 980 nm is shown in Fig. 4. The luminescence bands corresponding to the characteristic transitions in Er3+ ion
are present in the UC luminescence spectrum: violet (2H9/2 ? I15/2), green (2H11/2 ? 4I15/2 and 4S3/2 ? 4I15/2) and red (4F9/2 ? 4 I15/2) [13]. It is known that in UC processes the emission intensity IUC has a power law dependence on the excitation power Pexc:IUC (Pexc)n, where n is a number of photons required to excite the emitting state [22]. In Fig. 5 the dependence of the UC luminescence intensities measured for the main radiative transitions of Er3+ at 410, 540 and 665 nm versus excitation power are shown. The plots in the double logarithmic chart show linear behaviour of the UC luminescence intensity versus excitation power having parameters of the slopes close to 2 for the 540 nm and 665 nm bands. This is an indication that the 4S3/2 and 4F9/2 levels are 4
Fig. 4. Stationary UC luminescence spectrum of 2% Er3+ doped NaLaF4 excited at 980 nm measured at room temperature.
Fig. 5. Power dependence of the transitions 2H9/2 ? 4I15/2 (squares), 4S3/2 ? 4I15/2 (circles) and F9/2 ? 4I15/2 (triangles) in NaLaF4:Er3+ measured at room temperature.
A. Sarakovskis et al. / Optical Materials 31 (2009) 1517–1524
populated by two photon processes, while the slope of 2.98 for the band at 410 nm signifies the involvement of three-photon absorption process in the creation of the violet UC band. The possible excitation routes of the UC luminescence for Er3+ ion are shown in Fig. 6. The appearance of the green luminescence band at 540 nm might be a result of an excited state absorption process (ESA) i.e. sequential two photon absorption process: 4I15/2 + hm ? 4I11/2 + hm ? 4F7/2. Afterwards the relaxation to 2H11/2 and 4S3/2 populates the emitting levels for the green luminescence. Another route to populate the 4S3/2 level is by energy transfer (ET) process: (4I11/2,
1521
4
I11/2) ? (4I15/2, 4F7/2) (ET1). The involvement of the both ESA and ET processes is evident in the time-resolved luminescence decay plot for the green luminescence band excited at 980 nm (Fig. 7a). It can be seen that the luminescence decay curve has a complex nature: it has an intensity offset at t = 0 and also the rise part of about 5 ls signifying the involvement of the both processes in the creation of the luminescence. The appearance of the red luminescence at 660 nm might be caused by an ET process: (4I13/2, 4I11/2) ? (4I15/2, 4F9/2) (ET2). Clear evidence of pure ET process is a time-resolved luminescence decay plot for the red band excited at 980 nm (Fig. 7b) – it can be seen
Fig. 6. Schematic energy level diagram and the main radiative transitions in Er3+ ion. Full and dashed arrows are radiative and non-radiative transitions, respectively.
Fig. 7. Decay profiles for (a) 4S3/2 ? 4I15/2, (b) 4F9/2 ? 4I15/2, (c) 2H9/2 ? 4I15/2 optical transitions in NaLaF4:Er3+(0.5 mol%) excited at 980 nm measured at room temperature.
1522
A. Sarakovskis et al. / Optical Materials 31 (2009) 1517–1524
that the luminescence signal does not have any intensity offset at t = 0. The appearance of the violet band could be explained by an ET process (4S3/2, 4I13/2) ? (2H9/2, 4I15/2) (ET3). Again the ET process is supported by the luminescence decay kinetics excited at 980 nm (Fig. 7c) – no intensity offset is present at t = 0. In Fig. 8 the UC luminescence spectra for the samples at different doping levels are shown. It can be seen that the increase of Er3+ doping level leads to the decrease of the green-to-red bands intensity ratio. It could be explained by an interionic cross-relaxation process, which becomes dominant at higher Er3+ concentrations, when the average distance between the Er3+ ions diminishes. On one hand the cross-relaxa-
tion process depopulates 4S3/2 level (4S3/2, 4I15/2) ? (4I9/2, 4I13/2) (in Fig. 6 CR) thus decreasing the intensity of the green band. On the other hand it provides an extra route for 4F9/2 level population: (4I13/2, 4I11/2) ? (4I15/2, 4F9/2) (ET2) increasing the population of 4F9/2 level and the intensity of the red band. Another important parameter, which is essential for the characterization of the optical properties of the material, is the decay profile of the luminescence under direct excitation of the emitting level. In this work we have performed time-resolved luminescence measurements for the green (4S3/2 ? 4I15/2) and red (4F9/2 ? 4I15/2) luminescence bands at direct excitation not higher than 300 cm1 above the respective emitting level. The luminescence kinetics
Fig. 8. Stationary UC luminescence spectra of Er3+ doped NaLaF4 samples at different doping levels excited at 980 nm measured at room temperature.
Fig. 9. Decay profiles for the 4S3/2 ? 4I15/2 optical transition in NaLaF4 at different Er3+ concentrations: direct excitation at room temperature. Inset – double exponential fit of the decay for the sample with 4 mol% Er3+.
1523
A. Sarakovskis et al. / Optical Materials 31 (2009) 1517–1524
Fig. 10. Decay profiles for the 4F9/2 ? 4I15/2 optical transition in NaLaF4 at different Er3+ concentrations: direct excitation at room temperature. Inset – double exponential fit of the decay for the sample with 4 mol% Er3+.
Table 1 Lifetimes (in ls) of the main optical transitions in Er3+ doped NaLaF4 samples at different doping levels and NaErF4, measured under direct excitation at room temperature. The lifetime of NaYF4:Er3+ 2 mol% is given for the comparison [13].
4
4
4
4
S3/2 ; I15/2 F9/2 ; I15/2
Fast Slow Fast Slow
0.05%
0.1%
0.5%
1%
2%
4%
10%
NaErF4
NaYF4 2% [13]
50 510 110 530
40 490 100 520
35 470 90 510
30 410 85 480
25 350 70 450
20 240 60 390
7 80 30 230
0.275
360
5
430
related to the transitions from the different Er3+ energy levels to the ground state are shown in Figs. 9 and 10. It can be seen in the semilogarithmic chart that the luminescence decay curves for the green band (Fig. 9) contain more than one component. Exponential fit (Fig. 9 inset) shows that the experimental curve can be approximated by two exponents: ‘‘fast” and ‘‘slow”. At highest doping levels the slow component almost disappears and only the fast one remains. Similar tendency can be observed for the red luminescence band (Fig. 10), which also seems to be built of two components (Fig. 10 inset), although the concentration impact on the lifetime value of the slow component is less pronounced compared to the green band. The values of the lifetimes for the both luminescence bands at different doping levels are summarized in Table 1. The presence of the fast and slow components in the luminescence decay plots at direct excitation could be explained by the formation of several crystalline phases in the polycrystalline material as it was seen in the XRD patterns (Figs. 1 and 2). Since the decay profiles of the green luminescence at low Er3+ concentrations are mostly governed by a single slow exponent and XRD analysis reveals mainly the presence of NaLaF4:Er3+ in the samples, this slow exponential component could be attributed to NaLaF4:Er3+. It is also supported by the fact that the obtained lifetime value of the slow exponent is similar to that of NaYF4 (Table 1), which has similar structure and close effective phonon energy as NaLaF4 does. At higher doping levels the value of the slow component for the green luminescence band in the exponential fit decreases. This might be due to the cross-relaxation processes between Er3+ ions in NaLaF4,
in which the emitting 4S3/2 level is non-radiatively depopulated (cross-relaxation (4S3/2, 4I15/2) ? (4I9/2, 4I13/2)). The appearance of the fast component in the exponential decay of the green luminescence band and the growth of its relative intensity along with the increase of Er3+ concentration in NaLaF4 correlates with the XRD observations (Fig. 2) where the increase of ErF3 amount in the initial melt led to the shift of the corresponding XRD peaks towards those attributable to NaErF4 crystalline phase. This fact suggests that the fast exponential component in the decay curve of the green luminescence band could be attributed to the distorted NaErF4 crystalline phase. Nevertheless additional experiments are needed to elucidate the problem of the coexisting of the fast and slow components in the decay profiles. 4. Conclusions Erbium doped polycrystalline NaLaF4 material has been synthesized. The XRD patterns measured for NaLaF4 material at different doping levels of ErF3 showed considerable lattice distortion of NaLaF4 when Er3+ content was increased. The UC luminescence spectra measured for the samples at different doping levels of Er3+ revealed the characteristic luminescence bands in the violet (2H9/2 ? 4I15/2), green (2H11/2 ? 4I15/2 and 4S3/2 ? 4I15/2) and red (4F9/2 ? 4I15/2) spectral regions. The luminescence decay curves originating from 4S3/2 and 4F9/2 levels of Er3+ ions in NaLaF4 crystalline structure have been measured the lifetimes of the corresponding levels were determined. The effective phonon energy of NaLaF4 derived from Raman measurements for the undoped material was
1524
A. Sarakovskis et al. / Optical Materials 31 (2009) 1517–1524
estimated Eph 290 cm1. It is slightly lower than the one reported for NaYF4 suggesting that the potential efficiency of UC processes in NaLaF4 could be comparable to that in NaYF4. The obtained results imply that NaLaF4 crystalline phase properly synthesized and doped is a promising media in the field of UC phosphors. The work related to the synthesis of NaLaF4 crystallites free from other crystalline phases is under progress and the results will be published shortly. Acknowledgments This work was financially supported by National Research Programme of Latvia in Materials Science. The financial support of ESF is greatly acknowledged. The authors would like to thank Dr. A. Veispals for the help in the sample preparation, Dr. G. Chikvaidze for Raman measurements and Prof. U. Rogulis for fruitful discussions. References [1] J. Mendez-Ramos, V.K. Tikhomirov, V.D. Rodriguez, D. Furniss, Journal of Alloys and Compounds 440 (1–2) (2007) 328. [2] T. Schweizer, D.W. Hewak, D.N. Payne, T. Jensen, G. Huber, Electronics Letters 32 (7) (1996) 666. [3] D.N. Payne, L. Reekie, in: Proceedings Fourteenth European Conference on Optical Communication, Brighton 1988, vol. 1, published by Institution of Electrical Engineers, London, 1988, p. 49.
[4] F. Auzel, Chemical Review 2004 104 (2004) 139. [5] F. Tong, W.P. Risk, R.M. Macfarlane, W. Lenth, Electronics Letters 25 (20) (1989) 1389. [6] B. Dong, T. Yang, M.K. Lei, Sensors and Actuators B 123 (2007) 667. [7] J.E.C. Silva, G.F. de Sa, P.A. Santa-Cruz, Journal of Alloys and Compounds 323– 324 (2001) 336. [8] A. Shalav, B.S. Richards, M.A. Green, Solar Energy Materials and Solar Cells 91 (2007) 829. [9] A.C. Yanes, J. Del-Castillo, J. Mendez-Ramos, V.D. Rodriguez, M.E. Torres, J. Arbiol, Optical Materials 29 (8) (2007) 999. [10] M. Mortier, G. Patriarche, Optical Materials 28 (2006) 1401. [11] D. Chen, Y. Wang, E. Ma, Y. Yu, F. Liu, Optical Materials 29 (12) (2007) 1693. [12] Yu.V. Orlovskii, T.T. Basiev, K.K. Pukhov, I.N. Vorobiev, A.G. Papashvili, F. Pelle, V.V. Osiko, Journal of Luminescence 94–95 (2001) 791. [13] J.F. Suyver, J. Grimm, M.K. van Veen, D. Biner, K.W. Krämer, H.U. Güdel, Journal of Luminescence 117 (1) (2006) 1. [14] F. Liu, Y. Wang, D. Chen, Y. Yu, Materials Science and Engineering: B 136 (2–3) (2007) 106. [15] Y. Wei, F. Lu, X. Zhang, D. Chen, Journal of Alloys and Compounds 455 (1–2) (2008) 376. [16] K.W. Krämer, D. Biner, G. Frei, H.U. Güdel, M.P. Hehlen, S.R. Lüthi, Chemistry of Materials 16 (2004) 1244. [17] A. Aebischer, M. Hostettler, J. Hauser, K. Krämer, T. Weber, H.U. Güdel, H.-B. Bürgi, Angewandte Chemie International Edition 45 (2006) 2802. [18] A. Grzechnik, P. Bouvier, M. Mezouar, M.D. Mathews, A.K. Tyagi, J. Köhler, Journal of Solid State Chemistry 165 (1) (2002) 159. [19] M.M. Lage, F.M. Matinaga, J.-Y. Gesland, R.L. Moreira, Journal of Applied Physics 99 (5) (2006) 053510. [20] R.E. Thoma, H. Insley, G.M. Hebert, Inorganic Chemistry 5 (7) (1966) 1222. [21] P.P. Fedorov, I.I. Buchinskaya, O.S. Bondareva, A.A. Bystrova, L.L. Vistin, D.A. Ershov, S.P. Ivanov, V.A. Stasyuk, B.P. Sobolev, Russian Journal of Inorganic Chemistry 45 (6) (2000) 949. [22] M. Polnau, D.R. Gamelin, S.R. Lüthi, H.U. Güdel, M.P. Hehlen, Physical Review B 61 (5) (2000) 3337.