Journal of Non-Crystalline Solids 322 (2003) 319–323 www.elsevier.com/locate/jnoncrysol A comparison between different methods of calculating the radiative lifetime of the 4I13=2 level of Er3þ in various glasses B.J. Chen a, G.C. Righini a,* , M. Bettinelli b, A. Speghini b a b Optoelectronics and Photonics Department, Nello Carrara, Institute of Applied Physics (IFAC-CNR), via Panciatichi 64, I-50127 Firenze, Italy Dipartimento Scientifico e Tecnologico and INSTM, Universita di Verona, Strada Le Grazie 15, I-37134 Verona, Italy Abstract A comparison between the radiative lifetimes of the 4 I13=2 level of Er3þ derived from the conventional Judd–Ofelt (J–O) method and from a simple formula, that corresponds to Einstein relation for the emission probability of a twolevel system, has been carried out in some typical glasses, such as germanate, phosphate, silicate, and tellurite. The uncertainties of the radiative lifetimes are evaluated using standard error analysis; it comes out that uncertainties are much larger (some tens %) when using the J–O method than with the simple formula (error less than 1%). The same simple method can be used to calculate the radiative lifetimes of other rare earth ions, such as Ho3þ , Tm3þ and Yb3þ . 2003 Elsevier B.V. All rights reserved. PACS: 78.20.)e; 78.20.Bh; 78.30.Ly; 78.40.)q 1. Introduction Recent activity on rare earth (RE) doped glasses has been aimed at the development of integrated optical lasers and amplifiers which exploit the 4 I13=2 fi 4 I15=2 transition of Er3þ centred at 1.54 lm [1]. The search for better and better host glass materials is still under way, and Judd–Ofelt (J–O) theory [2,3] represents an important tool associating experimental and theoretical data in the description of the f–f optical transitions of RE ions. With J–O theory, in particular, the radiative transition probabilities of RE ions can be calcu- lated from their absorption spectra; as such it is often used, even if some problems have been discussed, related to its application to practical modelling processes and to the correctness of J–O representation for a particular RE ion [4,5]. In the present paper, we use a simple method to calculate the radiative lifetime of the 4 I13=2 level of Er3þ in different glass hosts, and we compare its results with those obtained from the J–O procedure. A similar method has been already suggested for other RE ions [6], but, to the best of our knowledge, not for the Er3þ ion. 2. Experimental * Corresponding author. Tel.: +39-055 423 5239; fax: +39055 423 5350. E-mail address: [email protected] (G.C. Righini). Small samples, of typical size 20 mm · 10 mm · 1 mm, of trivalent erbium doped phosphate, 0022-3093/03/$ - see front matter 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0022-3093(03)00221-7 320 B.J. Chen et al. / Journal of Non-Crystalline Solids 322 (2003) 319–323 Table 1 Chemical composition and refractive index for the Er3þ -doped glasses under investigation Glass Chemical composition (mol%) Refractive index B–Ge–Na silicate 67.35SiO2 –8.79GeO2 –6.10B2 O3 – 15.98Na2 O–0.52Er2 O3 –1.26Yb2 O3 49PbO–50GeO2 –1Er2 O3 99Ba(PO3 )2 –1Er(PO3 )3 99Cd(PO3 )2 –1Er(PO3 )3 99Ca(PO3 )2 –1Er(PO3 )3 97.5Sr(PO3 )2 –2.5Er(PO3 )3 19ZnO–80TeO2 –1Er2 O3 1.53 ± 0.01 Lead germanate Barium phosphate Cadmium phosphate Calcium phosphate Strontium phosphate Zinc tellurite silicate, lead-germanate and zinc-tellurite glasses were prepared by conventional melt-quenching techniques, by using the procedures described elsewhere [7–9]. The chemical compositions and the refractive indexes of these glasses are reported in Table 1. The refractive index of the silicate glass was measured using m-line spectroscopy [10], employing an SF-6 glass coupling prism. For the other glasses, the refractive indexes were taken from the literature [8,11,12]. Electronic absorption spectra were measured at room temperature in the visible region using a ouble beam spectrophotometer (spectral bandwidth of 0.5 nm) (Kontron Uvikon 941 Plus) and in the near infrared region using a Fourier transform infrared (FTIR) spectrometer (spectral bandwidth of 2 cm1 ) (Nicolet Magna 760). 2.03 ± 0.01 [8] 1.5849 ± 0.0002 1.6039 ± 0.0002 1.5479 ± 0.0002 1.5585 ± 0.0002 2.08 ± 0.01 [12] [11] [11] [11] [11] 3. Calculations For the B–Ge–Na silicate glass the experimental oscillator strengths of the f $ f transitions were obtained from the absorption spectrum (not shown) and used in the frame of the J–O parameterisation scheme [13,14]. The three Xk parameters (k ¼ 2, 4 and 6) were calculated by fitting with a least-squares method the electric dipole (ED) contributions of the experimental oscillator strengths of the observed transitions to the calculated ones. The matrix elements given by Carnall et al. [15] were employed in the calculation. Magnetic dipole (MD) contributions were subtracted from the observed oscillator strengths using the procedure described elsewhere [16]. The values of the intensity parameters for the B–Ge–Na silicate glass are reported in Table 2. Table 2 4 I13=2 fi 4 I15=2 transition barycenter vb , peak absorption cross-section r (peak), J–O parameters Xk and calculated lifetimes of the 4 I13=2 level for the Er3þ ion in the glasses under investigation (sJO : J–O procedure; sR : from Eqs. (2) and (3); d: relative difference between sJO and sR ) Glass vb (cm1 ) r (peak) (1020 cm2 ) X2 (pm2 ) X4 (pm2 ) X6 (pm2 ) sJO (ms) sR (ms) B–Ge–Na silicate Lead germanate Barium phosphate Cadmium phosphate Calcium phosphate Strontium phosphate Zinc tellurite 6643 ± 25 0.564 ± 0.002 4.05 ± 0.13 0.70 ± 0.15 0.403 ± 0.071 13.4 ± 5.6 14.10 ± 0.09 5.2 6563 ± 17 0.799 ± 0.002 4.76 ± 0.11 1.20 ± 0.12 0.807 ± 0.041 3.88 ± 0.70 3.68 ± 0.02 5.1 6560 ± 20 0.604 ± 0.002 5.02 ± 0.15 1.27 ± 0.18 0.923 ± 0.082 8.3 ± 2.1 7.22 ± 0.05 13.0 6580 ± 20 0.617 ± 0.002 5.34 ± 0.17 1.22 ± 0.21 0.851 ± 0.093 8.3 ± 2.5 8.46 ± 0.06 1.9 6601 ± 17 0.577 ± 0.002 5.48 ± 0.13 1.30 ± 0.15 0.765 ± 0.080 9.8 ± 2.4 8.23 ± 0.06 16.0 6587 ± 19 0.608 ± 0.002 5.033 ± 0.065 1.292 ± 0.089 0.926 ± 0.032 8.5 ± 1.1 8.23 ± 0.05 3.2 6602 ± 18 0.848 ± 0.002 6.30 ± 0.19 1.55 ± 0.26 1.29 ± 0.10 2.55 ± 0.71 2.59 ± 0.02 1.6 d (%) B.J. Chen et al. / Journal of Non-Crystalline Solids 322 (2003) 319–323 sJO 1 ¼ ; AED þ AMD ð1Þ which takes into account both the ED and MD contributions. On the other hand, given a general two-level system, denoted as 1 and 2 the ground and the excited state respectively, the radiative lifetime, sR , of the excited state, 2 is given by [19] sR ¼ 1 ; A21 ð2Þ where A21 is the spontaneous emission probability between states 2 and 1 (Einstein A coefficient). The A21 coefficient, in turn, is related to the wave number dependent absorption cross-section of the 2 ‹ 1 transition, rðvÞ, by the well-known equation [19] Z g1 2 2 ð3Þ A21 ¼ 8pcn vb rðvÞ dv; g2 where g1 and g2 are the degeneracies of state 1 and 2, c is the velocity of light, n is the refractive index of the medium, vb is the barycenter (in wave numbers) of the absorption band. The absorption and emission transitions between 4 I13=2 and 4 I15=2 of Er3þ ion are allowed by both ED and MD mechanisms, and therefore A21 contains both the ED and MD contributions. From the absorbance spectra of the Er3þ -doped glasses under investigation the absorption crosssections as a function of wave number were obtained from the well-known relation [13] rðvÞ ¼ 2:303 AbsðvÞ ; cl ð4Þ 0.4 B-Ge-Na Silicate 0.0 Absorption Cross Section (10-20 cm2) For the zinc-tellurite, lead-germanate and phospate glasses under investigation, the J–O parameters were taken from the literature [8,17,18] and are reported in Table 2 for comparison. The J– O intensity parameters were then used to calculate the ED contribution, AED , of the spontaneous emission probability of the 4 I13=2 fi 4 I15=2 emission transition of the Er3þ ion. The MD contribution, AMD , of the spontaneous emission probability for the 4 I13=2 fi 4 I15=2 emission transition was calculated following the procedure described in [16], and finally the radiative lifetime, sJO , for the 4 I13=2 level was calculated using the relation [13] 321 0.4 Barium phosphate 0.0 0.4 Strontium phosphate 0.0 0.4 Calcium phosphate 0.0 0.4 Cadmium phosphate 0.0 0.4 Lead germanate 0.0 0.4 Zinc tellurite 0.0 6200 6400 6600 6800 7000 7200 Wavenumber (cm-1) Fig. 1. Absorption cross-sections for the 4 I13=2 ‹ 4 I15=2 transition of the Er3þ ion in different oxide glasses. where AbsðvÞ is the wave number dependent absorbance, l and c are the thickness and the ion concentration (in ions/cm3 ), respectively. In Fig. 1 the absorption cross-sections in the range 6200– 7200 cm1 for the glasses under investigation are shown. The observed band centered at about 6600 cm1 is due to the 4 I13=2 ‹ 4 I15=2 absorption transition of the Er3þ ion. We see how large is the dependence of the absorption cross-section on the glass host. These absorption cross-sections are typical of the glasses investigated and in agreement with recent results reported in the literature [20,21]. 4. Results The sJO s of the 4 I13=2 level of the Er3þ ion for the glasses under investigation, calculated using relation (1), are reported in Table 2. The uncertainties on the calculated radiative lifetimes, sJO , were 322 B.J. Chen et al. / Journal of Non-Crystalline Solids 322 (2003) 319–323 obtained using standard error analysis [22] and are also reported in Table 2. The uncertainties are in some cases surprisingly high (even 20–30% of sJO ) but they can be considered as an upper limit of the experimental errors, because they are calculated using the simplifying assumption that the uncertainties of the Xk parameters are uncorrelated. The radiative lifetimes, sR , for the same glasses were instead calculated by using Eqs. (2) and (3); the sR are reported in Table 2 as well, together with the transition barycenters for the different samples. The uncertainties on sR were also calculated by standard error analysis using the standard deviations of the integrated absorption crosssection. 5. Discussion Table 2 shows that the relative differences of the radiative lifetimes sJO and sR are equal or less than 5% for most of the oxide glasses under investigation, and are much larger only for barium and calcium phosphate glasses (13% and 16%, respectively). Let us now analyse the uncertainties on sJO and sR . Usually, the r.m.s. error for the J–O procedure is 8% [7–9,17,18] and that is the reason why some authors have considered the J–O theory as a semiquantitative theory. However it can give a useful contribution in estimating transition probabilities with an accuracy generally not worse than 10%. Due to the theoretical complexity, an accurate theory for the calculation of f–f transition intensities of trivalent lanthanide ions is still absent. From Eqs. (1)–(3), considering the propagation of the errors of different experimental quantities in an equation [22], calculations show that the precision on the sR is mainly dependent on the standard deviation of the integrated absorption cross-section and of the barycenter of the 4 I13=2 ‹ 4 I15=2 absorption band. In fact, the measurement of the refractive index usually is quite accurate: as an example, according to the data reported in Table 1, in the worst case the error on n is ±0.7%. On the other hand, the precision on the sJO is due to a complicated function involving the sum of the standard deviations on the absorption cross-sections and on the barycenters of all the bands included in the J–O procedure, and the relative uncertainties on the J–O parameters could be as high as 10% (see Table 2). As a consequence, the data in Table 2 indicate that the uncertainties on the sR are at least an order of magnitude smaller than those on sJO . 6. Conclusions The radiative lifetime of the 4 I13=2 level of Er3þ ion in various oxide glasses was calculated both by the standard J–O method (sJO ) and by using a simple procedure based on the Einstein relation for the spontaneous emission probability between two states (sR ). The latter procedure implies the calculation of the oscillator strength of the 4 I13=2 ‹ 4 I15=2 absorption band, and avoids the data fitting process which is necessary in the J–O method. The comparison between the results obtained with the two methods confirms that the simpler procedure is feasible and gives results in agreement with those obtained from the more laborious J–O procedure. Indeed, the values of sR could be considered also more reliable than sJO , as the assessed uncertainties on the sR are at least an order of magnitude smaller than those on sJO . Moreover, we mention that the simple procedure used in the present paper, of immediate physical meaning, can be usefully employed at least in the first stage of design and characterization of novel glasses or crystalline materials containing REs. Besides the 4 I13=2 ‹ 4 I15=2 transition in Er3þ , this method can also be used for calculating the radiative lifetimes for other transitions from the first excited state to the ground state, such as 5 I7 fi 5 I8 for Ho3þ , 3 F4 fi 3 H6 for Tm3þ or 2 F5=2 fi 2 F7=2 for Yb3þ . Acknowledgements B.J.C. thanks ICTP, Trieste, for his fellowship in the frame of TRIL Program. M.B. and A.S. thank Erica Viviani (Universita di Verona) for expert technical assistance. The technical assis- B.J. Chen et al. / Journal of Non-Crystalline Solids 322 (2003) 319–323 tance of Roberto Calzolai (IFAC-CNR) is gratefully acknowledged. References [1] G.C. Righini, S. Pelli, M. Brenci, M. Ferrari, C. Duverger, M. Montagna, R. DallÕIgna, J. Non-Cryst. Solids 284 (2001) 223. [2] B.R. Judd, Phys. Rev. 127 (1961) 750. [3] G.S. Ofelt, J. Chem. Phys. 37 (1962) 511. [4] O. Florez, L. Malta, Y. Messaddeq, M.A. Aegerter, J. NonCryst. Solids 213 & 214 (1997) 315. [5] A.A. Kaminskii, A.A. Kornienko, M.I. Chertanov, Phys. Stat. Sol. (b) 134 (1986) 717. [6] M. Bettinelli, G. Ingletto, P. Polato, G. Pozza, G. Zanella, R. Zannoni, Phys. Chem. Glasses 37 (1996) 4. [7] R. Rolli, K. Gatterer, M. Wachtler, M. Bettinelli, A. Speghini, D. Aj o, Spectrochim. Acta, Part A 57 (2001) 2009. [8] M. Wachtler, A. Speghini, K. Gatterer, H.P. Fritzer, D. Aj o, M. Bettinelli, J. Am. Ceram. Soc. 81 (1998) 2045. [9] S. Zaccaria, M. Casarin, A. Speghini, D. Aj o, M. Bettinelli, Spectrochim. Acta, Part A 55 (1999) 171. [10] R. Ulrich, R. Torge, Appl. Opt. 12 (1973) 2901. 323 [11] O.V. Mazurin, M.V. Streltsina, T.P. Shvaiko-Shvaikovskaya, Handbook of Glass Data, Part B, Elsevier Science, The Netherlands, 1985, p. 539. [12] N. Mochida, K. Takahashi, K. Nakata, S. Shibusawa, J. Ceram. Soc. Jpn. 86 (1978) 317. [13] R. Reisfeld, Struct. Bonding 22 (1975) 123. [14] R.D. Peacock, Struct. Bonding 22 (1975) 83. [15] W.T. Carnall, P.R. Fields, K. Rajnak, J. Chem. Phys. 49 (1968) 4424. [16] S. Zaccaria, M. Casarin, A. Speghini, D. Aj o, M. Bettinelli, Spectrochim. Acta, Part A 55 (1999) 171. [17] R. Rolli, K. Gatterer, M. Wachtler, M. Bettinelli, A. Speghini, D. Aj o, Spectrochim. Acta, Part A 57 (2001) 2009. [18] A. Speghini, M. Bettinelli, D. Dal Dosso, S. Zaccaria, D. Aj o, in: P. Vincenzini, G.C. Righini (Eds.), Advances in Science and Technology – Innovative Light Emitting Materials, Techna Srl, Faenza, Italy, 1999, p. 181. [19] R. Loudon, The Quantum Theory of Light, 3rd Ed., Oxford University, New York, 2000. [20] N. Jaba, A. Kanoun, H. Mejri, A. Selmi, S. Alaya, H. Maaref, J. Phys. Condens. Matter 12 (2000) 4523. [21] A. Speghini, R. Francini, A. Martinez, M. Tavernese, M. Bettinelli, Spectrochim. Acta, Part A 57 (2001) 2001. [22] P.R. Bevington, D.K. Robinson, Data Reduction and Error Analysis for the Physical Sciences, 3rd Ed., McGraw-Hill, New York, 2002.
© Copyright 2026 Paperzz