Effect of relative alkali content on absorption linewidth in erbium-doped tellurite glasses

Effect of relative alkali content on absorption linewidth in erbium-doped tellurite glasses

Journal of Non-Crystalline Solids 255 (1999) 97±102 www.elsevier.com/locate/jnoncrysol E€ect of relative alkali content on absorption linewidth in e...

132KB Sizes 1 Downloads 10 Views

Journal of Non-Crystalline Solids 255 (1999) 97±102

www.elsevier.com/locate/jnoncrysol

E€ect of relative alkali content on absorption linewidth in erbium-doped tellurite glasses Lydia Le Neindre, Shibin Jiang *, Bor-Chyuan Hwang, Tao Luo, Jason Watson, Nasser Peyghambarian Optical Sciences Center, University of Arizona, Tucson, AZ 85721, USA

Abstract Er3‡ doped tellurite glasses with various Na2 O/Li2 O, Na2 O/K2 O and K2 O/Li2 O ratios (0±4.5) were prepared at 850°C using a platinum crucible. The e€ect of relative alkali content on absorption linewidth in erbium-doped tellurite glasses was investigated. The highest e€ective linewidth was obtained from glass samples with mixed alkali molar ratio of Na2 O/Li2 O ˆ 1, Na2 O/K2 O ˆ 1 and K2 O/Li2 O ˆ 1. Glass transition temperature, crystallization temperature, density, refractive index, Judd±Ofelt parameters and emission cross-section of Er3‡ doped tellurite glasses are also reported. Ó 1999 Elsevier Science B.V. All rights reserved.

1. Introduction Tellurite glass possesses a refractive index larger than those of silicate glasses which increases the local ®eld correction at an activator site and leads to larger radiative transition probabilities [1]. Thus rare earth ions in tellurite glasses typically have larger stimulated emission cross-sections compared to other binary oxide glasses [2]. Cooley proposed to use tellurite glasses as hosts for the Nd3‡ laser ion [3]. However, Nd3‡ doped tellurite glasses were never used for practical lasers due to their thermal expansion coecients, temperature coecients of the refractive index, and physical strength compared to commercial silicate and phosphate neodymium laser glasses. Er3‡ doped tellurite glasses have been attractive research subjects for upconversion emission due to their low

* Corresponding author. Tel.: +1-520 621 8241; fax: +1-520 621 4442; e-mail: [email protected]

phonon energy [4,5]. Recently, Ohishi et al., demonstrated 75 nm ¯attened gain bandwidth near 1.54 lm using an Er3‡ doped tellurite glass ®ber [6]. This demonstration attracted attention since ¯attened broad bandwidth of erbium doped ®ber ampli®ers is key issue in increasing the transmission capacity of wavelength-divisionmultiplexing transmission systems. We investigated the e€ect of glass composition on the e€ective absorption linewidth for the 4 I15=2 ± 4 I13=2 transition of Er3‡ ions in tellurite glasses. This paper reports the e€ect of relative alkali content on the absorption linewidth of erbium ions in tellurite glasses with di€erent alkali ions (Li, Na and K).

2. Experiments Three series of glass samples were prepared in the present study, which are (A): 75TeO2 ±20ZnO±xNa2 O±(4.5 ÿ x)Li2 O± 0.5NaF; (B): 75TeO2 ±20ZnO±xNa2 O±

0022-3093/99/$ - see front matter Ó 1999 Elsevier Science B.V. All rights reserved. PII: S 0 0 2 2 - 3 0 9 3 ( 9 9 ) 0 0 4 2 8 - 7

98

L. Le Neindre et al. / Journal of Non-Crystalline Solids 255 (1999) 97±102

(4.5 ÿ x)K2 O±0.5NaF; (C): 75TeO2 ±20ZnO± 1, xK2 O±(4.5 ÿ x)Li2 O±0.5NaF, where x ˆ 0, 2.25, 3.5, 4.5. One weight percent Er2 O3 (99.99%) was added into all glass samples. 0.5 mol% NaF was used to remove OHÿ in glass samples. Batches of 15 g were prepared from commercial powders of TeO2 (99%), anhydrous Li2 CO3 , K2 CO3 and Na2 CO3 , and ZnO (99.9%). The powders were melted in a platinum crucible at 850°C for 1 h in an electrically heated furnace in air. When the melting was completed, the liquids were cast into a mold at 300°C and samples with a thickness of 4 mm were prepared. All samples were annealed 2 h at a temperature ranging from 290°C to 300°C, depending upon the glass composition. The glasses were then cooled to room temperature at a rate of 20°C/h. The glass transition temperature (Tg ), crystallization onset temperature (Tx ) and melting temperature (Tm ) were determined by di€erential thermal analysis (DTA) at a heating rate of 10°C/ min, using aluminum oxide ceramic pans. Densities were measured following the ArchimedesÕ principle using water as the medium. The use of water does not a€ect either the density measurement or subsequent optical measurement. Refractive indices of the samples were measured on optically polished

glass plates at 1550 nm using a prism coupler refractive index measurement system. Absorption spectra were recorded with a spectrophotometer (CARY) from 300 to 3300 nm using samples with a thickness of approximately 2 mm. A background spectrum was ®rst recorded and then subtracted from the subsequent measurements. Emission spectra measurements at 1.5 lm were made using a 980 nm laser diode as a pumping source. 3. Results 3.1. Physical and thermal properties Table 1 lists densities and refractive indices at 1.55 lm of all three series of samples. Densities of the samples increase with the addition of Li‡ ion for the (A) series of glasses, and decrease with the addition of K‡ ion for the (B) and (C) series of glasses. Densities of the (A) series of glasses were larger than those of the (B) or (C) series of glasses, which we suggest is due to more ecient packing of Li‡ ions in the network structure [7]. Refractive indices of all samples increase with increasing density.

Table 1 Density, refractive index and peak absorption wavelength of tellurite glasses x (mol%)

Density (g/cm3 ) (‹0.01)

Refractive index (‹0.0002)

Peak wavelength (lm) (‹0.001)

75TeO2 ÿ20ZnOÿ(4.5 ÿ x)Na2 O±xLi2 O±0.5NaF±1 wt% Er2 O3 (A) 0 5.41 1.9973 1 5.44 2.0002 2.25 5.45 2.0034 3.5 5.46 2.0085 4.5 5.48 2.0117

1.531 1.531 1.532 1.532 1.532

75TeO2 ÿ20ZnOÿ(4.5 ÿ x)Na2 OÿxK2 Oÿ0.5NaFÿ1 wt% Er2 O3 (B) 0 5.41 1.9973 1 5.36 1.9908 2.25 5.30 1.9871 3.5 5.20 1.9811 4.5 5.17 1.9778

1.531 1.531 1.532 1.531 1.532

75TeO2 ÿ20ZnOÿ(4.5 ÿ x)Li2 O±xK2 O±0.5NaF±1 wt% Er2 O3 (C) 0 5.48 2.0117 1 5.43 2.0032 2.25 5.33 1.9952 3.5 5.25 1.9850 4.5 5.17 1.9778

1.532 1.531 1.531 1.531 1.532

L. Le Neindre et al. / Journal of Non-Crystalline Solids 255 (1999) 97±102

Fig. 1. DTA trace of a tellurite glass (75TeO2 ±20ZnO± 2.25Na2 O±2.25Li2 O±0.5NaF).

A typical DTA trace is illustrated in Fig. 1. Glass transition, exothermic crystallization and endothermic glass melting are observed. The glass transition temperature (Tg ), the crystallization onset temperature (Tx ) are marked in Fig. 1. The quantity, Tx ÿ Tg , has been frequently used as a measure of glass stability. To achieve a working range of temperature during our samples ®ber drawing, it is desirable to have (Tx ÿ Tg ) as large as possible [8]. The studied glasses have a (Tx ÿ Tg ) exceeding 100°C, indicating these samples are stable against devitri®cation in agreement with previous reports [9].

99

bands seen in this spectrum correspond to various absorption transitions of erbium in this tellurite glass. The 1.53 lm transition 4 I15=2 ±4 I13=2 is important for ®ber ampli®ers operating in the 1.5 lm telecommunication window. The peak absorption wavelengths for the 4 I15=2 ±4 I13=2 transition of Er3‡ ions in all three series of samples are listed in Table 1. Table 1 also shows that changes in the peak wavelength are less than 1 nm. The e€ective absorption linewidth for the 4 I15=2 ± 4 I13=2 transition was obtained for each sample, which was calculated by integrating the absorption coecient over wavelength and dividing by the peak intensity. The e€ective absorption linewidth as a function of alkali ions concentration in all three series of samples is illustrated in Fig. 3. The largest e€ective absorption linewidth was obtained for glass samples with an alkali ion ratio of R2 O/ M2 O ˆ 1 in the three series (R and M stand for alkali ions). The e€ective absorption linewidth of samples with mixed alkali ions is 10% greater than in single alkali glasses. We also calculated the e€ective absorption linewidth for composition (A) series samples at two other transitions. Fig. 4 shows the e€ect of the alkali ion concentration on the e€ective absorption linewidths of the 975 nm 4 I15=2 ±4 I11=2 and the 653 nm 4 I15=2 ±4 F9=2 transitions. Fig. 4 indicates that in each case the largest e€ective linewidth is obtained from mixed alkali glasses with a ratio of Na2 O/

3.2. E€ect on e€ective absorption linewidth Fig. 2 illustrates the absorption spectrum of Er3‡ ion in 75TeO2 ±20ZnO±2.25Na2 O±2.25Li2 O± 0.5NaF (1 wt% Er2 O3 ) glass. The absorption

Fig. 2. Absorption spectrum of a tellurite glass.

Fig. 3. E€ective absorption linewidths as a function of alkali ions concentration in all three series of glasses }: 75TeO2 ±20ZnO±xNa2 O±(4.5 ÿ x)Li2 O±0.5NaF; n: 75TeO2 ±20ZnO± M: 75TeO2 ±20ZnO±xK2 O± xNa2 O±(4.5 ÿ x)K2 O±0.5NaF; (4.5 ÿ x)Li2 O±0.5NaF. Lines are drawn as guides for the eye.

100

L. Le Neindre et al. / Journal of Non-Crystalline Solids 255 (1999) 97±102 Table 2 Judd±Ofelt parameters in erbium doped glasses Glass

X2 (10ÿ20 cm2 )

X4 (10ÿ20 cm2 )

X6 (10ÿ20 cm2 )

Tellurite Silica Fluoride

4.0 3.8 2.5

1.2 0.6 1.5

0.8 0.3 1.0

Fig. 4. E€ect of the alkali ion concentration on the e€ective absorption linewidths of 975 nm 4 I15=2 ±4 I11=2 transition (N) and 653 nm 4 I15=2 ±4 F9=2 transition (h) in 75TeO2 ±20ZnO±xNa2 O± (4.5 ÿ x)Li2 O±0.5NaF. Lines are drawn as guides for the eye.

Li2 O ˆ 1. The e€ective absorption linewidths of samples with mixed alkali ions is 10% and 9% greater for the 4 I15=2 ±4 F9=2 and 4 I15=2 ±4 I11=2 transitions, respectively, compared to the Li‡ only glass. 3.3. The Judd±Ofelt parameters

Fig. 5. Absorption and emission cross-sections of Er3‡ ions in 75TeO2 ±20ZnO±25Na2 O±2.25Li2 O±0.5NaF.

cross-section is around 0.79 ´ 10ÿ20 cm2 at 1.53 lm for most of the samples. The emission cross-section is calculated from McCumber theory [12]. According to McCumber theory, the absorption and emission cross-sections are related by

The glass sample with the composition, 75TeO2 ±20ZnO±2.25Na2 O±2.25Li2 O±0.5NaF, was selected for further experiments due to their thermal stability and spectral properties. The Judd±Ofelt parameters were calculated for this glass sample using absorption intensity at three di€erent wavelengths: 521, 653 and 975 nm [10]. We obtained the following parameters: cm2 , X4 ˆ 1.2 ´ 10ÿ20 cm2 , X2 ˆ 4.0 ´ 10ÿ20 ÿ20 2 cm . As Jorgensen et al., described, X6ˆ 0.8 ´ 10 the X2 parameter is a€ected by covalent chemical bonding, and the X4 and X6 parameters are related to the rigidity of the medium in which the ions are situated [11]. Table 2 shows the Judd±Ofelt parameters in tellurite, silica and ¯uoride glass hosts. X2 for tellurite glass is larger than for other glasses, indicating a greater degree of covalency in the tellurite glass.

where ra and re are the absorption and stimulated emission cross-section, respectively, m is the photon frequency, e is the net free energy required to excite one Er3‡ ion from the 4 I15=2 to 4 I13=2 state at temperature T, h is the Planck constant, and k is the Boltzmann constant. e was determined using the procedure provided in [13]. Fig. 5 illustrates the calculated absorption and emission cross-section for the 4 I15=2 ±4 I13=2 transition of Er3‡ ions in the glass with a composition of 75TeO2 ±20ZnO± 2.25Na2 O±2.25Li2 O±0.5NaF.

3.4. Absorption and emission cross-section

4.1. Mixed alkali e€ect

The absorption cross-section was determined from the absorption spectrum. The maximum

The mixed alkali e€ect is a well-known phenomenon in various oxide glasses [14±16]. When

re …m† ˆ ra exp ‰…e ÿ hm†=kT Š;

…1†

4. Discussion

L. Le Neindre et al. / Journal of Non-Crystalline Solids 255 (1999) 97±102

one alkali replaces another one in a glass, many properties such as electrical conductivity, dielectric loss, alkali di€usion and viscosity deviate from the principle of additivity [17,18]. Research on the mixed alkali e€ect was performed in silicate glasses. Recently, Komatsu et al., observed a mixed alkali e€ect on the thermal stability, viscosity, electrical and dielectric properties in (20 ÿ x)Li2 Oáx Na2 Oá80TeO2 glasses [19]. We observed that the e€ective absorption linewidth increased until the ratio Na2 O/Li2 O ˆ 1, Na2 O/ K2 O ˆ 1 and K2 O/Li2 O ˆ 1. The e€ective absorption linewidth in mixed alkali glasses is 10% greater than in single alkali glasses. We assume this change in linewidth might be a mixed alkali e€ect. This observation is important because it indicates that a greater linewidth could be obtained by selecting glasses with mixed alkali ions. Since the e€ective absorption linewidth was calculated by integrating the absorption coecient over wavelength and dividing by the peak intensity, it is not a€ected by errors of chemical batch weight and doping concentration, or intensity measurement errors. Therefore, the change of the e€ective absorption linewidth is due to the changes in local structure around Er3‡ ions. Generally, the linewidth broadening of Er3‡ ions in glasses is inhomogeneous broadening which is caused by differences in the ligand ®eld from Er3‡ site to site [20]. This variation means that the variation in ligand ®elds around Er3‡ ions in glasses with mixed alkali ions is greater than that in glasses with single alkali ions. This greater variation is reasonable and agrees with other publications. Balaya et al., proposed that the phenomenon of the mixed alkali e€ect in tellurite glasses is closely related to the distortion in the TeO4 units by the substituting alkali ions [21]. 4.2. Absorption and emission linewidth Broad emission linewidths and emission crosssections are important in the realization of broad bandwidth ampli®cation. The emission cross-section (re ) of Er3‡ ions in tellurite glasses is high since the emission cross-section increases with the refractive index of the host as a function of 2 …n2 ‡ 2† =n [12]. Table 1 shows that the refractive

101

Fig. 6. Emission cross-sections of Er3‡ ions in tellurite (75TeO2 ±20ZnO±2.25Na2 O±2.25Li2 O±0.5NaF) and phosphate glasses (Ref. [23]).

index of tellurite samples is approximately 2.0, which is larger than that of typical Er3‡ doped phosphate and silicate glasses [22]. Emission linewidth is related to absorption linewith as expressed in Eq. (1). Absorption linewidth can be determined more easily and reliably than emission linewidth. So we focused our studies on the characterization of absorption linewidth. For the Er3‡ ion 4 I15=2 ±4 I13=2 transition, the greater the absorption linewidth, the greater the emission linewidth. Er3‡ ions in tellurite glasses have an broad emission band near 1.53 lm. Fig. 6 shows the emission cross-sections as a function of wavelength for the Er3‡ ion 4 I13=2 ±4 I15=2 transition in phosphate and tellurite glasses [23]. The emission band of Er3‡ ions in tellurite glass has a larger width than that in phosphate glass. Tellurite glass also has a larger emission cross-section in the 1450±1650 nm wavelength region. We note that the cross-section at around 1600 nm in tellurite glass is twice as large as in phosphate glass. Our theoretical calculation indicates that a 80 nm ¯attened gain bandwidth, from 1530 to 1610 nm, could be obtained in Er3‡ doped tellurite glasses. 5. Conclusion The e€ect of relative alkali content on the effective absorption linewidth of Er3‡ ions in tellurite glasses was observed in our samples, indicating

102

L. Le Neindre et al. / Journal of Non-Crystalline Solids 255 (1999) 97±102

that an increase of linewidth could be obtained by selecting glasses with mixed alkali ions. These tellurite glasses are thermally stable for ®ber drawing. Theoretical modeling indicates that an 80 nm ¯attened gain bandwidth, from 1530 to 1610 nm, is achievable in Er3‡ doped tellurite glasses. All these data indicate that tellurite glasses with mixed alkali ions are promising materials for broad band erbium doped ®ber ampli®ers. Acknowledgements We acknowledge support from BMDO through Center for Optoelectronic Devices, Interconnects and Packaging (COEDIP). References [1] R.R. Jacobs, M.J. Weber, IEEE J. Quantum Electron 12 (1976) 102. [2] M.J. Weber, J.D. Myers, D.H. Blackburn, J. Appl. Phys. 52 (1981) 2944. [3] R.F. Cooley, US patent No. 3883357, 1975. [4] S. Tanabe, K. Hirao, N. Soga, J. Non-Cryst. Solids 122 (1990) 79. [5] H. Nii, K. Ozaki, M. Herren, M. Morita, J. Lumin. 76/77 (1998) 116. [6] A. Mori, Y. Ohishi, S. Sudo, Electron. Lett. 33 (1997) 863.

[7] J. Heo, D. Lam, G.H. Sigel, E.A. Mendoza, D.A. Hensley, J. Am. Ceram. Soc. 75 (1992) 227. [8] M.G. Drexhage, O.H. El Bayoumi, C.T. Moyniyan, A.J. Bruce, K.H. Chung, D.L. Gavin, J.T. Loretz, J. Am. Ceram. Soc. C 65 (1982) 168. [9] J.S. Wang, E.M. Vogel, E. Snitzer, J.L. Juckel, V.L. Lasilva, Y. Silberberg, J. Non-Cryst. Solids 178 (1994) 109. [10] E. Desurvire, Erbium-Doped Fiber Amplif. (1994) 215. [11] C.K. Jorgensen, R. Reisfeld, J. Less-Common Met. 93 (1983) 107. [12] D.E. Mc Cumber, Phys. Rev. A 134 (1964) 299. [13] W.J. Miniscalo, R.S. Quimby, Opt. Lett. 16 (1992) 258. [14] B. Schulz, M. Schulz, S. Trimper, Phys. Rev. 58 (1998) 3368. [15] M. Tomozawa, V. Mc Gahay, in: L.D. Pye, W.C. La Course, H.J. Stevens (Eds.), Physics of Non-Crystalline Solids, Taylor and Francis, London, 1992, p. 429. [16] A.K. Agarwal, D.E. Day, J. Am. Ceram. Soc. 65 (1982) 231. [17] R. Terai, J. Non-Cryst. Solids 6 (1971) 121. [18] R.J. Charles, J. Am. Ceram. Soc. 48 (1965) 432. [19] T. Komatsu, R. Ike, R. Sato, K. Matusita, Phys. Chem. Glasses 36 (1995) 216. [20] R. Powell, Physics of Solid-State Laser Materials, AIP, New York, 1998, pp. 103. [21] P. Balaya, C.S. Sunandana, in: B.V.R. Chowdari, Q.G. Liu, L.Q. Chen (Eds.), Recent Advances in Fast Ion Conducting Materials and Devices, World Scienti®c, Singapore, 1990, p. 539. [22] S. Jiang, M. Myers, N. Peyghambarian, J. Non-Cryst. Solids 239 (1998) 143. [23] S. Jiang, T. Luo, B. Hwang, G. Nunzi-Conti, M. Myers, D. Rhonehouse, S. Honkanen, N. Peyghambarian, Opt. Eng. 37 (1998) 3282.