Spectroscopic properties of Eu-doped Y-stabilized ZrO2 microtubes

Journal of Luminescence ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Contents lists available at ScienceDirect

Journal of Luminescence journal homepage: www.elsevier.com/locate/jlumin

Spectroscopic properties of Eu-doped Y-stabilized ZrO2 microtubes K. Utt, M. Part, T. Tätte, V. Kiisk, M.G Brik, A.A. Chaykin †, I. Sildos n Institute of Physics, University of Tartu, Riia Str. 142, 51014 Tartu, Estonia

art ic l e i nf o

Keywords: YSZ Microtubes Europium sol–gel Photoluminescence Time-resolved spectroscopy

a b s t r a c t Hereby we report a new microstructured luminescent material – Eu-doped yttria-stabilized zirconia (YSZ) microtube – prepared by a special sol–gel route. Transparent, crack-free and brightly luminescent microtubes were obtained after thermal treatment at temperatures as high as 1100 1C. The implications of time-resolved and site-selective studies of Eu3 þ luminescence are discussed. The decay kinetics of Eu3 þ luminescence is modeled following the Judd–Ofelt theory and matched to the experimental data. & 2013 Elsevier B.V. All rights reserved.

1. Introduction Functional optical materials are often required in specific geometry. For example, a microtubular geometry could have an advantage in applications such as miniature fuel cells [1], gas sensors [2], microfluidics [3], etc. The sol–gel method appears to be particularly versatile in the preparation of metal oxides with desired microscale geometrical shapes. While nanopowders and thin films are the most common products of sol–gel processing, fibers and especially tubular structures are very rare [4]. Since the sol–gel method is flexible also in the doping of oxide matrices, functionalization of the resulting structures for optical or electrical applications is of particular interest. Yttria-stabilized tetragonal or cubic ZrO2 (YSZ) is an important refractory, optical and electroceramic material which combines outstanding optical properties with excellent hardness and durability in harsh environment. In particular, the deleterious tetragonalto-monoclinic phase transition of pure ZrO2 is hindered in YSZ permitting device operation up to very high temperatures. The high ionic conductivity of YSZ is the basis for its use in ion conducting membrane applications such as gas sensors and fuel cells. Trivalent rare earth (RE3 þ ) activators are natural dopants for YSZ contributing to the phase stabilization similarly to yttrium [6] and their solubility in the stabilized phase is considerably higher. RE doping of YSZ has already allowed multifunctional use of the material as thermal barrier coating/thermographic phosphor [5]. In the microtubular geometry the range of potential applications also include miniature

n

Corresponding author. Tel.: þ 372 737 4613; fax: þ372 738 3033. E-mail addresses: [email protected] (V. Kiisk), ilmo@fi.tartu.ee, [email protected] (I. Sildos). † Deceased

solid oxide fuel cells, spray nozzles for liquid metals, miniature plasma systems etc. Hereby we focus on the optical spectroscopy of sol–gel-derived Eu3 þ -doped YSZ microtubes. Using a combination of site-selective and time-resolved spectroscopy, existence of three Eu3 þ centers in tetragonal YSZ with characteristic emission patterns and lifetimes is demonstrated for the first time. 2. Samples and experimental To prepare YSZ microtubes, first a sol with shear thinning viscoelastic properties was obtained from Zr(OBu)4 solution in butanol as a result of hydrolysis (initiated by drop-wise addition of 0.7 mol of water per 1 mol of alkoxide as 5% solution in butanol) and final evacuation of the excess solvent. 8 mol% yttrium and 1 mol% europium were incorporated by dissolving respective nitrate in the sol. By using a glass rod, the sol was pulled into jets which were first kept in air with 80–90% relative humidity at 20–22 1C for 1 min and then in air with 3–5% relative humidity at 20 1C for 15 min. The microtubes self-formed during the period due to solidification of the outer layer and formation of a hollow core. The tubes had lengths 10–15 mm, typical outer diameters 30–50 mm and wall thicknesses 10–15 mm (Fig. 1a). The tubes were annealed up to 1100 1C in air to crystallize and densify the material, remove organic residues and optimize the RE fluorescence. For some measurements (XRD, spectrofluorometry), powders were prepared from the same sol under identical conditions and were assumed to have identical structural and spectroscopic properties. Photoluminescence (PL) emission spectra were collected from single free-standing microtubes excited with a focused beam of a diode-pumped solid state laser (355 or 266 nm), laser diode (405 nm) or tunable pulsed optical parametric oscillator (pulse duration 3 ns,

0022-2313/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jlumin.2013.10.056

Please cite this article as: K. Utt, et al., J. Lumin. (2013), http://dx.doi.org/10.1016/j.jlumin.2013.10.056i

K. Utt et al. / Journal of Luminescence ∎ (∎∎∎∎) ∎∎∎–∎∎∎

2

Fig. 1. (a) Optical micrographs of a YSZ:Eu3 þ microtube annealed at 700 1C. Left: illumination with white light. Right: fluorescence of Eu3 þ excited by 405 nm laser (focused at a point slightly above the field of view) and observed through 550 nm high-pass filter. (b) Raman spectrum of YSZ:Eu3 þ microtubes depending on annealing temperature (indicated).

Fig. 2. PL spectra of YSZ:Eu3 þ microtubes annealed at 1100 1C depending on the excitation wavelength (indicated).

pulse energy  20 mJ, repetition rate 20 Hz). The spectra were recorded by using a spectrograph (Andor SR-303i, spectral resolution 1 nm) equipped with an image-intensified CCD (Andor DH-501). PL spectra were corrected to instrumental response. PL decay kinetics was recorded with a Hamamatsu photomultiplier tube (H8259-01) operating in photon counting mode.

3. Results and discussion The crystal structure of the microtubes was evaluated from Raman spectra recorded with a Renishaw inVia micro-Raman spectrometer (Fig. 1b). The Raman pattern is easily assigned to tetragonal YSZ; there are no peaks due to monoclinic phase. The cubic phase should be identified by a relatively broad band centered at 617 cm  1 which unfortunately has an overlap with the 611 and 641 cm  1 peaks due to the tetragonal phase [7]. Moreover, it is argued that Raman-scattering is relatively less sensitive to the cubic phase [8]. An increase of annealing temperature (from 600–1100 1C) preserves the Raman pattern but leads to the removal of the broad background and a slight narrowing of Raman peaks. Rietveld refinement of XRD pattern (not shown) also confirmed that the tetragonal phase is prevalent. Under focused blue or UV laser beam, the Eu-doped microtubes emitted bright red luminescence easily seen by naked eye. Optical quality of the microtubes was characterized by inspecting the propagation of light in the microtube under a microscope (Fig. 1a). To inject light into a microtube, a bright luminescent spot was created inside the microtube by focusing a UV laser beam on its side area. In general, the microtubes remained transparent even after annealing at 1100 1C. Practically no attenuation of light was detected over distances of a few mm. The fluorescence image reveals that in addition to the weak uniform glow caused by scattering from inhomogeneities on the nano scale, some brighter spots are observed which indicate extended defects inside the microtubes not easily seen with white light illumination. The defects probably result from deficiencies in the preparation. A survey of PL spectra at several excitation wavelengths is shown in Fig. 2. The observed crystal-field-split spectral lines are markedly broadened, even after thermal treatment at 1100 1C. In YSZ the contribution of inhomogeneous broadening is expected to be quite strong since each oxygen vacancy produces a large number of differently distorted cationic sites due to lattice

Fig. 3. Time-resolved PL spectra of YSZ:Eu3 þ microtubes annealed at 1100 1C excited at 464 nm. The detection time window with respect to the laser pulse is indicated.

relaxation towards oxygen vacancies although the number of differently coordinated sites remains small [9]. Jang and Meltzer [10] have used spectral hole burning to show that both the inhomogeneous and homogeneous linewidths of Eu3 þ are dramatically increased in YSZ relative to ordinary crystals. Yet, the correlations between different peaks (with changing excitation wavelength) suggest at least three distinguishable Eu3 þ sites in the YSZ microtubes. A distinct emission pattern is obtained by site-selective excitation at 466 nm (in resonant with the 7F0-5D2 transition of Eu3 þ ). Corresponding Eu center is labeled as Eu I. Only a slight variation of excitation wavelength (464 nm) leads to completely different spectrum. However, this spectrum turns out to be a mixture of two Eu centers as revealed by time-resolved spectroscopy (Fig. 3). The longer-living Eu3 þ site is labeled as Eu II and the shorterliving one as Eu III. The Eu II center is also separately observed under excitation with 210 nm, which presumably leads to host-mediated excitation. A common feature of all emission spectra is the presence of a weak line due to the 5D0-7F0 transition. It is known that this emission is generally obtained only from noncentrosymmetric sites [11,12]. At least for Eu II and Eu III centers, also a threefold splitting of the 5D0-7F1 band is clearly observed. Although for Eu I emission we observe only single line at 5D0-7F1 transition, some works on tetragonal YSZ:Eu have reported similar emission

Please cite this article as: K. Utt, et al., J. Lumin. (2013), http://dx.doi.org/10.1016/j.jlumin.2013.10.056i

K. Utt et al. / Journal of Luminescence ∎ (∎∎∎∎) ∎∎∎–∎∎∎

pattern where one or even two weak additional components in the 5 D0-7F1 band were resolved [14,15]. Yet, the small intensity of the 5 D0-7F0 band and only a weak splitting of the 5D0-7F1 band suggest that a high symmetry is approximately preserved. The original cation site symmetry of a tetragonal structure (D2d) would allow twofold splitting of both 5D0-7F1 and 5D0-7F2 bands [12]. Although there are indeed two strong 5D0-7F2 components in the Eu I spectrum, also two weaker components seem to be present. This can be explained by a slightly lower symmetry (C2v) which would also account for the presence of the 5D0-7F0 emission [15]. For the Eu II center we distinguish experimentally only two components in the 5D0-7F2 band. A symmetry, which would allow threefold splitting of 5D0-7F1 and twofold splitting of 5 D0-7F2, cannot be found. If we assume that three components exist in the 5D0-7F2 band, then the spectrum is in agreement with D2 symmetry, but this would not allow the 5D0-7F0 emission [12]. Again, the only possibility seems to be C2v. In the original study of spectral-structural relationship of Eu3 þ :YSZ by Dexpert-Ghys et al. [9] the Eu II emission pattern was obtained from tetragonal YSZ (total lanthanide concentration 10 at%) and was assigned to 8-fold coordinated Eu3 þ site with the oxygen vacancy in the second coordination sphere. Unchanged first coordination around Eu3 þ is also suggested by later EXAFS studies [13]. Several authors have reported either Eu I or Eu II type of spectrum from ZrO2:Eu containing tetragonal phase [14–19]. The similar spectra and identical symmetries imply that the two sites may differ only by a slight variation of the surrounding. Aside from the unresolved splitting of the 5D0-7F2 transition, the Eu III emission has some similarity with the spectrum obtained from cubic YSZ containing 425 at% of lanthanide and assigned to 6-fold coordinated Eu3 þ sites of C2v symmetry, similar to RE3 þ sites in cubic RE sesquioxides [9]. We cannot rule out the presence of a small amount of cubic phase in the microtubes. However, Eu III spectrum bears even stronger resemblance to characteristic Eu3 þ spectrum in glasses [20]. The spectrum contains wider emission peaks and has a notably higher ratio of 5D0-7F2 (electric dipole) and 5D0-7F1 (magnetic dipole) intensities which all suggest that the Eu3 þ sites are located in less ordered and lowsymmetry (possibly C1) surrounding. The luminescence kinetics after pulsed excitation at 466 nm (Eu I) reveals nearly single exponential decay with lifetime 2.2 ms (Fig. 4). Ref. [15] reports 2.33 ms average lifetime for this center. In contrast, at least two exponential components are required to fit the decay excited at 464 nm. Based on the time-resolved spectra, these components (with lifetimes 1.0 and 2.7 ms) must correspond to the Eu III and Eu II sites, respectively.

Fig. 4. PL decay of YSZ:Eu3 þ microtubes (annealed at 1100 1C) at different excitation and emission wavelengths.

3

Fig. 5. PL excitation spectra of YSZ:Eu3 þ powder (annealed at 1100 1C) depending on the detection wavelength. The inset depicts a properly weighted difference of the spectra to obtain more narrow and symmetric CT bands characteristic of Eu I and Eu II sites.

Rather strong PL is also obtained when intermediate excitation wavelengths (e.g. 355 nm and 266 nm) are used (Fig. 2). While 355 nm radiation could still induce 4f–4f transitions of Eu3 þ , the 266 nm radiation most probably induces charge transfer (CT) excitation typically located around 5 eV for oxides [21]. Due to the overlapping of the emission of several Eu3 þ centers it was not possible to record their excitation spectra separately. Yet, rather different CT bands could be obtained by using different detection wavelengths (Fig. 5). Further distinction of individual CT bands was achieved by taking a weighted difference of the spectra (Fig. 5 inset). The CT band centered at 245 nm was previously obtained for Eu I type of emission [16,19]. The CT band located at 270 nm was reported in Refs. [17,18] for Eu II type of emission. This data suggests that the Eu I center has lower coordination because the latter implies shorter Eu–O bond length which, in turn, implies higher ionic bond character and correspondingly increased CT energy [21]. The experimental Eu3 þ emission spectra can be used for determination of the Judd-Ofelt intensity parameters [22,23] which allow for a theoretical estimate of the radiative lifetime. It is well known that the intensity of the 5D0-7F1 magnetic dipole transition is nearly host-independent and its radiative transition probability AR ð5 D0 -7 F 1 Þ in different hosts varies between 30 and 60 s  1 [24–26]. For 5D0-7FJ (J ¼2, 4) we can express the transition probabilities as AR ð5 D0 -7 F J Þ ¼

! D E2 64π 4 e2 v3 nðn2 þ 2Þ2 ΩJ  5 D0 ‖U ðJÞ ‖7 F J  3h 9

ð1Þ

where e is elementary charge, n is energy of transition (in cm  1), h is Planck's constant and n is refractive index. The matrix elements of the doubly reduced tensor operators between the corresponding states are denoted by h…i. Only the terms with non-zero reduced matrix elements are included. The values of AR ð5 D0 -7 F J Þ, and hence the Ω2 and Ω4 parameters, can be estimated easily by knowing AR ð5 D0 -7 F 1 Þ as reference and by using the fact that AR is proportional to the area under the corresponding spectral band. The value of AR ð5 D0 -7 F 1 Þ for the YSZ host (with refractive index n) can be estimated as Amd ¼ ðn=n′Þ3 Amd ′, where Amd ′ ¼ 57:34 s  1 is the 5D0-7F1 radiative transition probability for the 50(NaPO3)6 þ 10TeO2 þ20AlF3 þ19LiF þEu2O3 glass with refractive index n′¼1.591 [27]. Bulk YSZ has relatively high refractive index which depends on the content of yttria. It is reported that cubic single crystals of YSZ containing 10 mol% of Y2O3 have refractive index 2.19 at 550 nm [28]. To account for some nanoporosity in sol–gel-derived structures, we assume refractive index of  2.1 in our case. Based on this value of n, the estimated Judd–Ofelt intensity parameters are

Please cite this article as: K. Utt, et al., J. Lumin. (2013), http://dx.doi.org/10.1016/j.jlumin.2013.10.056i

K. Utt et al. / Journal of Luminescence ∎ (∎∎∎∎) ∎∎∎–∎∎∎

4

(in units of 10-20 cm2): Ω2 ¼2.55, Ω4 ¼ 1.08 for the Eu I center, Ω2 ¼0.93, Ω4 ¼0.60 for the Eu II center, and Ω2 ¼ 6.59, Ω4 ¼ 2.47 for the Eu III center. Corresponding theoretical 5D0 state radiative lifetimes are 2.3, 3.9 and 1.2 ms. For the Eu I and Eu III centers the agreement with experimental values is very good. The calculated lifetime for the Eu II center is notably greater than the corresponding experimental result. However, the experimental lifetime value for this center is less accurate (Fig. 4). It is also possible that a specific non-radiative mechanism applies to this particular impurity center.

Excellence “Mesosystems: Theory and Applications”, TK114). The authors are grateful to Dr. Hugo Mändar for XRD analysis. References [1] [2] [3] [4] [5] [6] [7] [8]

4. Conclusions The special sol–gel route carried out under the conditions of high humidity followed by a proper thermal treatment provided highly transparent and luminescent YSZ:Eu microtubes containing mostly tetragonal phase. Site-selective and time-resolved spectral analysis allowed a clear distinction of three different Eu3 þ emission centers, of which at least two sites are situated in the crystalline phase (with a probable site symmetry C2v, but different coordination). It was possible to achieve a good agreement between the theoretically modeled and experimentally observed decay rates of different Eu3 þ sites assuming quite high refractive index of 2.1 for the microtube material.

Acknowledgments This work was supported by Estonian Science Foundation Grants nos. 8699, 8377, 9283, 7603 and SF0180058s07 and by the European Union through the European Regional Development Fund (Centre of

[9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28]

K. Kendall, Int. J. Appl. Ceram. Technol. 7 (2010) 1. J.X. Wang, et al., Appl. Phys. A 88 (2007) 611. S. Shopova, et al., Opt. Express 15 (2007) 12735. K. Saal, T. Tätte, M. Järvekülg, V. Reedo, A. Lõhmus, I. Kink, Intern. J. Mater. Product Tech. 40 (2011) 2. M.D. Chambers, D.R. Clarke, Ann. Rev. Mater. Res. 39 (2009) 325. C. Wang, M. Zinkevich, F. Aldinger, Pure Appl. Chem. 79 (2007) 1731. D. Gazzoli, G. Mattei, M. Valigi, J. Raman Spectr. 38 (2007) 824. E. Fernández López, V. Sánchez Escribano, M. Panizza, M.M. Carnasciali, G. Busca, J. Mater. Chem. 11 (2001) 1891. J. Dexpert-Ghys, M. Faucher, P. Caro, J. Solid State Chem. 54 (1984) 179. K.W. Jang, R.S. Meltzer, Phys. Rev. B 52 (1995) 6431. X.Y. Chen, G.K. Liu, J. Solid State Chem. 178 (2005) 419. P.A. Tanner, Chem. Soc. Rev. 42 (2013) 5090. P. Ghosh, K.R. Priolkar, A. Patra, J. Phys. Chem. C 111 (2006) 571. A. Speghini, M. Bettinelli, P. Riello, S. Bucella, A. Benedetti, J. Mater. Res. 20 (2005) 2780. J. Liao, D. Zhou, B. Yang, R. Liu, Q. Zhang, Opt. Mater. 35 (2012) 274. L. Chen, Y. Liu, Y. Li, J. Alloys Compd. 381 (2004) 266. M.R.N. Soares, C. Nico, M. Peres, N. Ferreira, A.J.S. Fernandes, T. Monteiro, F. M. Costa, J. Appl. Phys. 109 (2011) 013516. L. Li, et al., J. Nanosci. Nanotech. 11 (2011) 350. P. Ghosh, A. Patra, Langmuir 22 (2006) 6321. D.K. Rice, L.G. DeShazer, Phys. Rev. 186 (1969) 387. P. Dorenbos, J. Lumin. 111 (2005) 89. B.R. Judd, Phys. Rev. 127 (1962) 750. G.S. Ofelt, J. Chem. Phys. 37 (1962) 511. R. Balda, J. Fernandez, J.L. Adam, M.A. Arriandiaga, Phys. Rev. B 54 (1996) 12076. M. Dejneka, F. Snitzer, R.E. Riman, J. Lumin. 65 (1995) 227. R. van Deun, K. Binnemans, C. Görller-Walrand, J.L. Adam, J. Phys. Condens. Matter 10 (1998) 7231. D. Maheswari, J.S. Kumar, L.R Moorthy, K. Jang, M. Jayasimhadri, Physica B 403 (2008) 1690. D.L. Wood, K. Nassau, T.Y. Kometani, Appl. Opt. 29 (1990) 2485.

Please cite this article as: K. Utt, et al., J. Lumin. (2013), http://dx.doi.org/10.1016/j.jlumin.2013.10.056i