Effects of melting temperature and composition on ... - OSA Publishing

Effects of melting temperature and composition on spectroscopic properties of Er3+-doped bismuth glasses Yushi Chu,1 Jing Ren,1, 3 Jianzhong Zhang,1, 4 Lu Liu,1 Pengfei Wang,1 Jun Yang,1 Gangding Peng,2 and Libo Yuan 1 1

2

Key Lab of In-fiber Integrated Optics, Ministry Education of China, Harbin Engineering University, Harbin 150001, China Photonics & Optical Communications, School of Electrical Engineering & Telecommunications, University of New South Wales, Sydney 2052, NSW, Australia 3 [email protected] 4 [email protected]

Abstract: The spectroscopic properties (SPs) of bismuth glasses are strongly influenced by the preparation conditions. Here, we studied the effects of melting temperature and concentration ratio of Bi2O3 to SiO2 on the SPs of Er3+ doped Bi2O3-B2O3-SiO2 bismuth glasses. We have shown that the low melting temperature is preferable to achieve large and broad gain of Er3+ ions. When melted at high temperatures, however, the glasses get darkened which is against the application of these materials for optical fiber amplifiers. On the other hand, we have also shown that a compromise has to be made between the laser gain and threshold regarding the optimum concentration ratio of Bi2O3 to SiO2. Theoretical radiative transition rate and quantum efficiency of the 1.55 μm emission of Er3+ have been obtained by Judd-Ofelt (JO) fitting. The variation of the JO parameters characterizing the local chemical environments around Er3+ can be explained based on the concept of optical basicity. ©2015 Optical Society of America OCIS codes: (160.5690) Rare-earth-doped materials; (160.3380) Laser materials.

References and links S. Tanabe, N. Sugimoto, S. Ito, and T. Hanada, “Broad-band 1.5 µm emission of Er3+ ions in bismuth-based oxide glasses for potential WDM amplifier,” J. Lumin. 87–89, 670–672 (2000). 2. A. Jha, S. Shen, and M. Naftaly, “Structural origin of spectral broadening of 1.5-μm emission in Er3+-doped tellurite glasses,” Phys. Rev. B 62(10), 6215–6227 (2000). 3. A. Jha, B. Richards, G. Jose, T. T. Fernandez, P. Joshi, X. Jiang, and J. Lousteau, “Rare-earth ion doped TeO2 and GeO2 glasses as laser materials,” Prog. Mater. Sci. 57(8), 1426–1491 (2012). 4. S. Ohara and N. Sugimoto, “Bi2O3-based erbium-doped fiber laser with a tunable range over 130 nm,” Opt. Lett. 33(11), 1201–1203 (2008). 5. Z. Yang, Q. Zhang, and Z. Jiang, “Photo-induced refractive index change of bismuth-based silicate glass,” J. Phys. D Appl. Phys. 38(9), 1461–1463 (2005). 6. W. Shen, J. Ren, S. Baccaro, A. Cemmi, and G. Chen, “Broadband infrared luminescence in γ-ray irradiated bismuth borosilicate glasses,” Opt. Lett. 38(4), 516–518 (2013). 7. C. Jauregui, J. Limpert, and A. Tünnermann, “High-power fibre lasers,” Nat. Photonics 7(11), 861–867 (2013). 8. O. Sanz, E. Haro-Poniatowski, J. Gonzalo, and J. M. Fernandez Navarro, “Influence of the melting conditions of heavy metal oxide glasses containing bismuth oxide on their optical absorption,” J. Non-Cryst. Solids 352(8), 761–768 (2006). 9. Y. Zhang, Y. Yang, J. Zheng, W. Hua, and G. Chen, “Effects of Oxidizing Additives on Optical Properties of Bi2O3-B2O3-SiO2 Glasses,” J. Am. Ceram. Soc. 91(10), 3410–3412 (2008). 10. G. Yang, D. Chen, J. Ren, Y. Xu, H. Zeng, Y. Yang, and G. Chen, “Effects of melting temperature on the broadband infrared luminescence of Bi-doped and Bi/Dy co-doped chalcohalide glasses,” J. Am. Ceram. Soc. 90(11), 3670–3672 (2007). 11. J. Ren, B. Li, G. Yang, W. Xu, Z. Zhang, M. Secu, V. Bercu, H. Zeng, and G. Chen, “Broadband near-infrared emission of chromium-doped sulfide glass-ceramics containing Ga2S3 nanocrystals,” Opt. Lett. 37(24), 5043– 5045 (2012). 1.

#248120 © 2016 OSA

Received 18 Aug 2015; revised 13 Nov 2015; accepted 15 Nov 2015; published 23 Dec 2015 1 Jan 2016 | Vol. 6, No. 1 | DOI:10.1364/OME.6.000279 | OPTICAL MATERIALS EXPRESS 279

12. A. D. Guzman-Chavez, Y. O. Barmenkov, and A. V. Kir’yanov, “Spectral dependence of the excited-state absorption of erbium in silica fiber within the 1.48–1.59 m range,” Appl. Phys. Lett. 92(19), 191111 (2008). 13. A. V. Kir’yanov, Y. O. Barmenkov, and A. D. Guzman-Chavez, “Er3+ excited-state absorption in an Erbiumdoped silica fiber at the wavelengths 1490–1580 and 978 nm,” Laser Phys. 18(11), 1251–1256 (2008). 14. A. V. Kir’yanov, Y. O. Barmenkov, G. E. Romero, and L. E. Zarate, “Concentration effects in commercial erbium-doped silica fibers fabricated through the MCVD and DND technologies,” IEEE J. Quantum Electron. 49(6), 511–521 (2013). 15. S. Dai, T. Xu, Q. Nie, X. Shen, and X. Wang, “Investigation of concentration quenching in Er3+:Bi2O3–B2O3– SiO2 glasses,” Phys. Lett. A 359(4), 330–333 (2006). 16. X. Wang, Q. Nie, T. Xu, S. Dai, X. Shen, and L. Liu, “Optical spectroscopy and energy transfer of Er3+ /Ce3+ in B2O3-doped bismuth-silicate glasses,” J. Opt. Soc. Am. B 24(4), 972–978 (2007). 17. J. Yang, S. Dai, N. Dai, S. Xu, L. Wen, L. Hu, and Z. Jiang, “Effect of Bi2O3 on the spectroscopic properties of erbium-doped bismuth silicate glasses,” J. Opt. Soc. Am. B 20(5), 810–815 (2003). 18. J. Yang, S. Dai, Y. Zhou, L. Wen, L. Hu, and Z. Jiang, “Spectroscopic properties and thermal stability of erbiumdoped bismuth-based glass for optical amplifier,” J. Appl. Phys. 93(2), 977–983 (2003). 19. X. Feng, S. Tanabe, and T. Hanada, “Hydroxyl groups in erbium-doped germanotellurite glasses,” J. Non-Cryst. Solids 281(1–3), 48–54 (2001). 20. H. Sun, J. Zhou, and J. Qiu, “Recent advances in bismuth activated photonic materials,” Prog. Mater. Sci. 64, 1– 72 (2014). 21. V. L. Stolyarova, A. L. Shilov, S. I. Lopatin, and S. M. Shugurov, “High-temperature mass spectrometric study and modeling of thermodynamic properties of binary glass-forming systems containing Bi2O3.,” Rapid Commun. Mass Spectrom. 28(7), 801–810 (2014). 22. Y. Xia, T. Tang, C. Chen, M. Jin, and M. Chen, “Preparation of α-Bi2O3 from bismuth powders through lowtemperature oxidation,” Trans. Nonferrous Met. Soc. China 22(9), 2289–2294 (2012). 23. V. Dimitrova and T. Komatsu, “Optical basicity and chemical bonding of Bi2O3 containing glasses,” J. NonCryst. Solids 382, 18–23 (2013). 24. X. Zhu, C. Mai, and M. Li, “Effects of B2O3 content variation on the Bi ions in Bi2O3-B2O3-SiO2 glass structure,” J. Non-Cryst. Solids 388, 55–61 (2014). 25. S. Tanabe, T. Ohyagi, N. Soga, and T. Hanada, “Compositional dependence of Judd-Ofelt parameters of Er3+ ions in alkali-metal borate glasses,” Phys. Rev. B Condens. Matter 46(6), 3305–3310 (1992). 26. J. Ren, T. Wagner, M. Bartos, M. Frumar, J. Oswald, M. Kincl, B. Frumarova, and G. Chen, “Intense nearinfrared and midinfrared luminescence from the Dy3+-doped GeSe2-Ga2Se3-MI (M=K, Cs, Ag) chalcohalide glasses at 1.32, 1.73, and 2.67 μm,” J. Appl. Phys. 109(3), 033105 (2011). 27. S. Tanabe, T. Ohyagi, S. Todoroki, T. Hanada, and N. Soga, “Relation between the Ω6 intensity parameter of Er3+ ions and the 151Eu isomer shift in oxide glasses,” J. Appl. Phys. 73(12), 8451–8454 (1993). 28. M. Wachtler, A. Speghini, K. Gatterer, H. P. Fritzer, D. Ajo, and M. Bettinelli, “Optical properties of rare-earth ions in lead germanate glasses,” J. Am. Ceram. Soc. 81(8), 2045–2052 (1998).

1. Introduction Fiber amplifiers with broadband and flat gain characteristics are required for dense wavelength division multiplexing (DWDM) optical network. Although Er3+-doped SiO2based fiber amplifiers (EDFA) have shown tremendous commercial success, the gain width of EDFA at 1.55 μm is limited to only few tens of nanometer [1,2]. Thus, much effort has been devoted to searching for new host materials with broader gain spectra [3]. Thanks to the wide gain at 1.55 µm and being fusion-spliceable to SiO2-based fibers, Er3+ doped bismuth fibers have attracted a lot of attention [4]. Such type of fibers are also inherently photosensitive [5,6], offering the opportunity to directly write fiber Bragg gratings in the same gain media, and reducing the splicing losses when inserted into optical networks [7]. When melted at high temperatures and for a long period of time, however, they suffer from darkening and crystallization resulted from the reduction of bismuth oxide (Bi2O3) to bismuth metal. Such problems limit to a great extent the optical applications of these glasses [8]. To avoid the undesirable reduction of Bi2O3 during melting, various oxidants such as CeO2 and mixture of Sb2O3 (or As2O3) and nitrate salts have been used, leading to much improved optical quality of the glasses [9]. Although the influence of melting temperature (MT) on spectroscopic properties (SPs) of un-doped bismuth glasses has been studied, its effect on those of Er3+ doped glasses is still open to questions [10]. In this work, we prepared Er3+ doped Bi2O3-B2O3-SiO2 (BBS) bismuth glasses with the addition of extra Ba(NO3)2 and Sb2O3 as the oxidizing agents. Apart from the advantages of bismuth glasses mentioned above, the present system also has the following merits such as the

#248120 © 2016 OSA


raw materials being relatively cheap, excellent fiber drawing ability and large refractive index (>2.0). Our study consists of two parts: first, fixing the concentration ratio of Bi2O3 to SiO2, the effect of MT on SPs of the glasses including absorption spectra, up-conversion (UC) luminescence, near infrared (NIR) emission and gain spectra was studied; second, fixing the MT, the influence of the concentration ratio of Bi2O3 to SiO2 on SPs of the glasses was examined. Judd-Ofelt (JO) analysis was carried out to obtain the radiative transition rate and quantum efficiency of the 1.55 μm emission of Er3+. The variation of the JO parameters characterizing the local chemical environments around Er3+ was explained based on the concept of optical basicity. 2. Experimental 2.1 Glass preparation Glasses with the nominal compositions of xBi2O3-10B2O3-(85-x)SiO2-1Er2O3 (x = 30, 40, 45, 50, 60 in mol.%, referred to as H1 ~H5, respectively) were prepared by conventional meltquenching method. Very pure compounds Bi2O3 (99.999%), SiO2 (99.99%), Er2O3 (99.99%), Sb2O3 (99.99%), H3BO3 (Guaranteed Reagent, GR) and Ba(NO3)2 (Analytical Reagent, AR) were used as the starting materials. The raw materials were weighed 20g, mixed in an agate mortar for at least 15 min., stored in an alumina crucible and then put in an electric furnace. To prevent evaporation of Bi2O3, the materials were pre-sintered at 450 °C for 30 min, then the samples H1 to H5 were melted at 1150 °C for 90 min, whereas five samples of the same composition to H3 (x = 45 mol.%) were prepared at different melting temperatures from 950 to 1350 °C with an interval of 100 °C (referred to as T1 ~T5, respectively). The thickness of the samples is approximately 2 ± 0.2 mm. 2.2 Measurements Absorption spectra of the glasses were measured by a Perkin-Elmer Lambda 950 UV-VIS spectrophotometer over the spectral range of 300-2000 nm. Infrared (IR) transmission spectra were measured by a Fourier transform infrared (FTIR) spectrophotometer (NICOLET NEXUS firm: Thermo Scientific). Emission spectra were measured by an Edinburgh FS920 fluorescence spectrometer equipped with a liquid-nitrogen cooled steady state InSb detector and luminescence decay was recorded by a TDS3000C digital phosphor oscilloscope. Table 1. MT (°C), Density (g/cm3, ± 0.04), refractive Index at 1.55 μm (n), concentrations of Er3+ (NEr3+ ± 0.01, 1020 ions/cm3) and free OH- groups (NOH-, 1019 ions/cm3) of the glass samples Glass samples T1 T2 T3(H3) T4 T5 H1 H2 H3(T3) H4 H5

Bi2O3 (mol. %) 45 45 45 45 45 30 40 45 50 60

MT 950 1050 1150 1250 1350 1150 1150 1150 1150 1150

Density 6.39 6.19 6.10 5.89 5.76 5.34 5.75 6.10 6.28 6.77

n 1.95 1.95 1.95 1.95 1.95 1.83 1.92 1.95 1.98 2.04

NEr3+ 1.54 1.49 1.47 1.42 1.39 1.70 1.51 1.47 1.40 1.31

NOH2.82 3.14 2.95 3.19 3.18 2.92 2.90 2.95 3.06 3.34

NEr3+ × NOH4.34 4.68 4.35 4.53 4.42 4.96 4.38 4.35 4.28 4.38

Refractive index was measured by using variable angle spectroscopic ellipsometry (VASE, J. A. Woollam Co., Inc ellipsometry) with rotating analyzers. Cauchy model was used for the fitting. Sample density was obtained by Archimedes method with alcohol as the immersion liquid. Er3+ concentration was calculated from the measured density and initial composition. Table 1 summarizes the refractive indices (1530nm), densities, and Er3+ concentrations of the glass samples.

#248120 © 2016 OSA


3. Results 3.1 Effect of MT on SPs of the glasses Figure 1 shows absorption spectra of the samples prepared at different MT. The cut off edge of the samples shifts from 400 to 500 nm (red shift) with MT. The sample gets darkened when prepared at higher MT (inset photo). The absorption bands at 488, 522, 544, 652, 798, 976 and 1530 nm are attributed to the electronic transitions of Er3+ ions from the 4I15/2 ground state to 4F7/2, 2H11/2, 4S3/2, 4F9/2, 4I9/2, 4I11/2 and 4I13/2 excited states, respectively. The absorption cross-section σa(λ) of the samples was calculated according to σa(λ) = α(λ)/NEr3+, where α(λ) is the absorption coefficient at wavelength λ. σa(λ) of the 1.55 μm emission band slightly decreases with MT (inset).

Fig. 1. Absorption spectra of the samples prepared at different melting temperatures. Insets: digital photo of the samples and absorption cross section of the 4I15/2 → 4I13/2 transition of Er3+.

The influence of MT on UC luminescence spectra is illustrated in Fig. 2(a). The UC spectra consist of four bands locating at 486, 524, 545 and 660 nm, respectively. The corresponding transitions are indicated in the inset. The UC decreases with MT especially for the 486 nm band which is hardly seen when MT is above 1250 °C. The UC may stem from the exited state absorption (ESA) as indicated in the inset.

Fig. 2. Visible up-conversion luminescence (a) and 1.55 μm emission spectra (b) of the samples prepared at different MT. The excitation wavelength is 980 nm. Inset in (a): schematic of up-conversion mechanism (by excited state absorption, ESA). Inset in (b): decay curve and fit to the experimental data of the T5 sample taken as an example.

Figure 2(b) shows variation of 1.55 μm emission spectra with MT. Upon the 980 nm excitation, an emission band covering S (1460 - 1530 nm), C (1530 - 1565 nm) and L (1565 ~1625 nm) fiber communication bands can be observed. As MT increases, the emission intensity weakens gradually. The effective bandwidth Δλeff was calculated according to 1700

Δλ eff =

 I ( λ ) dλ / I

peak

, where I(λ) and Ipeak are the emission intensities at wavelength λ and

1450

emission peak, respectively. The results are listed in Table 2, Δλeff becomes smaller (from 44

#248120 © 2016 OSA


to 40 nm) with MT, while the full width at half maximum (FWHM) remains constant (55 nm). The emission decay fitted by the first-order exponential function (inset) decreases slightly from 1.87 to 1.72 ms with MT (Table 2). Figure 3 shows MT dependence of the 1.55 μm emission gain spectra. The gain coefficient, G(λ), is calculated according to G(λ) = NEr3+[P σe(λ) - (1 - P) σa(λ)], where P is the population inversion given by the ratio between the population of Er3+ ions at the lasing level (4I13/2) and total concentration of Er3+ ions, σe(λ) is the emission cross-section which can λ 4 A rad , where c is the be calculated according to Fuchtbauer-Ladenburg equation: σ e = 8πcn 2 Δλ eff speed of light, n is the refractive index (cf. Table 1) and Arad is the radiative transition rate obtained by applying Judd-Ofelt analysis (see below), the values are shown in Table 2. Positive gain appears when forty percent (40%) of total Er3+ ions are populated to the lasing level. The maximum gain (inset in Fig. 3), quantum efficiency (η = τmd /τrad, where τmd and τrad ( = 1/Arad) are measured and theoretically calculated lifetimes, see Table 2), and the product of σe × τrad (a measure of the laser threshold, cf [11].) decreases with MT. All these results clearly suggest that the low MT is preferable to achieve large and broad gain. However, the gain is also dependent on the ESA of Er3+ as in the case of Er3+-doped silica fibers [12–14], so in reality, the calculated G(λ)-values can suffer from the problem of overestimating.

Fig. 3. 1.55 μm gain spectra of the samples prepared at different melting temperatures (MT). Inset: MT dependence of the maximum gain.

3.2 Effect of concentration ratio of Bi2O3 to SiO2 on SPs of the glasses Figure 4 shows compositional variation of absorption spectra. The absorption edge shifts to longer wavelength and the color of the samples gets darkened with Bi2O3. As a comparison, the glass of the same composition to H3 but free of oxidant is also shown in the figure, which is totally black and of very poor optical quality. The absorption cross section of the samples increases with Bi2O3 (inset in Fig. 4).

Fig. 4. Absorption spectra of the samples prepared at the same MT. Insets: digital photos of the samples and absorption cross section of the 4I15/2 → 4I13/2 transition of Er3+.

#248120 © 2016 OSA


Compositional dependence of UC spectra is illustrated in Fig. 5(a). The spectra consist of three bands locating at 524, 545 and 660 nm, respectively. The UC intensity considerably increases with Bi2O3. Since all the samples were prepared at relatively high MT (1250 °C), the 486 nm UC emission cannot be recorded similar to the result shown in Fig. 2(a).

Fig. 5. Visible up-conversion spectra (a) and 1.55 μm emission spectra (b) of the samples prepared at the same melting temperature. The excitation wavelength is 980 nm. Inset in (a): schematic of up-conversion mechanism (by excited state absorption, ESA). Inset in (b): decay curve and fit to the experimental data of the H5 sample taken as an example.

Changes of the 1.55 emission spectra with the composition are demonstrated in Fig. 5(b). Both the emission intensity and Δλeff increases with Bi2O3 (except for H2), while the FWHM changes irregularly. The largest FWHM is 58 nm, which is larger than that of Al/P codoped silica glasses (~44 nm), comparable to fluoroziconate glasses (~65 nm), but smaller than that of TeO2-based glasses (~76 nm) and bismuth glasses containing larger amount of B2O3 (~80 nm) [1,2]. The emission decay fitted by the first-order exponential function (inset) increases from 1.57 to 2.24 ms with Bi2O3 (Table 2). The lifetime of the studied glasses is in accordance to the published work on the similar glasses, indicating that our results were reliable [1, 15–18], but smaller than that of dehydrated telluride glasses (~3.85 ms) [19] and silica fibers (~10 ms) [12–14]. The present samples were prepared without any effect to remove the OH- [19], but in contrast silica fibers are generally prepared with the great effect to avoid the contamination of water. So we assumed that the presence of a large amount of OH- may account for the observed short lifetime as will be discussed below. Compositional dependence of the 1.55 μm emission gain spectra is shown in Fig. 6. Positive gain appears when forty percent (40%) of total Er3+ ions are populated to the lasing level. The maximum gain decreases with Bi2O3 (inset in Fig. 6), but the quantum efficiency η and the product of σe × τrad increase with Bi2O3 (Table 2), thus compromise has to be made between the laser gain and threshold regarding the optimum concentration ratio of Bi2O3 to SiO2.

Fig. 6. Variation of the 1.55 μm gain spectra with the composition of the samples. Inset: Compositional dependence of the maximum gain.

#248120 © 2016 OSA


4. Discussion

4.1 Effects of MT and the addition of oxidant on SPs of the glasses It is known that bismuth ions can exist in a multiple of valence states in glasses, such as Bi5+, Bi3+, Bi2+, Bi+, Bi0 and ion/metal clusters [20]. In the present case, with the addition of oxidizing agents Ba(NO3)2 and Sb2O3, the following reactions occurred during the melting [8– 10]: Ba ( NO3 )2 + Sb 2 O3  > 450  BaO + Sb 2 O5 + 2NO x ↑

(1)

Sb 2 O5  > 700°C  Sb 2 O3 + O 2 ↑

(2)

Bi 0 + O 2  950 ~ 1350°C  Bi 2 O3

(3)

Bi 2 O3  > 700°C  Bi 0 + O2 ↑ (4) Normally, the thermal dissociation or vaporization of Bi2O3 (Eq. (4) occurs at a low oxygen partial pressure. However, it can also occur under atmospheric pressure at high temperatures (even starts at 700 °C [21]). It has been demonstrated in the binary Bi2O3-SiO2 glass system that the vaporization results in the formation of Bi0 species, in which manner Bi0 also forms in the present ternary glass system. On the other hand, the oxidation of Bi0 is preferable from the thermodynamic point of view because Eq. (3) is accompanied by a minus Gibbs free energy [22]. Therefore, the dissociated Bi0 is readily oxidized to Bi2O3 if enough O2 bubbles form in the glass melts according to Eq. (2). However, as Eq. (2) and (3) go, there may not be enough O2 bubbles available for completing Eq. (3). As a result, the scattering in the visible light wavelength range increases, leading to the darkening (Fig. 1). Since the phonon energy barely changes with MT, it is likely that the quenching of UC luminescence (Fig. 2(a)) is related to the decreased transparence in the visible light region (Fig. 1), viz., to the increased background absorption of the host glasses with MT. 4.2 Effect of concentration ratio of Bi2O3 to SiO2 on SPs of the glasses

The red shift of the absorption edge with Bi2O3 (Fig. 4) can be understood based on the concept of optical basicity (OB) which represents the ability of oxygen ions donating electron to surrounding metal ions [23]. The OB for Bi2O3, B2O3 and SiO2 extracted from refractive index data are 1.19, 0.46 and 0.53, respectively. Obviously, OB increases with Bi2O3, indicating that the overlap between oxygen 2p and bismuth 6p orbitals decreases with Bi2O3 [23]. Since the band gap is related to the degree of the orbital overlap between oxygen and bismuth ions, it shrinks with Bi2O3. Generally, the inverse of the 1.55 μm emission lifetime (1/τmd), that is, the total transition rate is mainly subject to the radiative transition rate (1/τrad), multiphonon decay rate (WMP), energy transfer (ET) rate between Er3+ ions (WEr3+), and ET rate between Er3+ and OHimpurities (WOH-). Previous study on concentration quenching of Er3+ doped BBS glasses has shown that significant ET between Er3+ ions occurs only when the concentration of Er3+ is over 1 mol.% [15], accordingly, WEr3+ has been neglected in the present study. On the other hand, the multiphonon decay rate strongly depends on the phonon energy of host materials, which can be estimated by using Raman spectroscopy. Previous study has shown that the glass network of BBS glasses mainly consists of [BiO6], [SiO4] and [BO4] groups with the vibrational energies being approximately 700 cm−1, 1000 cm−1 and 1300 cm−1, respectively [16, 24]. Thus, the decreased average phonon energy with the increase in Bi2O3 is one of the causes leading to the enhanced NIR (Fig. 5(b)) and particularly UC emissions (Fig. 5(a)). However, the UC depopulates the 4I11/2 manifold of Er3+ (cf. inset in Fig. 5(a)), which may in turn decrease the emission from the lower-lying 4I13/2 manifold. This is likely to be the reason why H3 shows less intense NIR emission than that of H2 sample (Fig. 5(b)). With further

#248120 © 2016 OSA


increase in Bi2O3 (H4 and H5), it appears that the enhancement of NIR emission due to the reduced phonon energy overcomes the depopulation effect of UC, resulting in the overall increase in the NIR emission. In addition, it is well known that OH- groups are serious quenchers of NIR luminescence of Er3+. The OH- groups cause mid-infrared absorption bands at 2500 and 3300 cm−1 in the infrared transmission spectra of the samples (results are not shown here). The concentration of free OH- groups (NOH-) can be estimated according to the expression [19]: NOH- = (NA/εL) ln(1/T), where NA is the Avogadro constant, ε is the molar absorptivity of the free OH- groups (49.1 × 103 cm2/mol), L is the thickness of the glasses, T is the transmittance at 3300 cm−1. Since there is no relevant report about the ε value for bismuth glasses, the value for silicate glasses was adopt, in accordance with some published works, for example, X. Feng [19] and X. Wang [16]. Since WOH- is proportional to the product of NEr3+ and NOH- [19], which varies randomly with MT and concentration ratio of Bi2O3 to SiO2 (last column in Table 1). Nonetheless, H1 and T1 samples which show the shortest and longest lifetimes in the H1-H5 and T1-T5 series respectively have the biggest and smallest values of the product. This seems to suggest that the influence of the free OH- groups on the lifetime of the 1.55 μm emission cannot be overlooked but is assumed not to be the decisive factor. 4.3 Judd-Ofelt (JO) analysis Table 2. Δλeff (nm), FWHM (nm), σe (( × 10−21cm2)), calculated and measured lifetimes (ms) of the 1.55 μm emission, quantum efficiency ( × 100%), σe × τmd ( × 10−24cm2•s) and Judd-Ofelt parameters Ωt (t = 2, 4, 6, × 10−20 cm2) of the samples of the glass samples Glass samples T1 T2 H3(T3) T4 T5 H1 H2 H3(T3) H4 H5

Δλeff

FWHM

44 42 41 41 40 34 43 41 43 45

56±2 56±2 56±2 56±2 55±2 55±2 58±2 56±2 56±2 51±2

σe (±0.10) 6.63 6.59 6.10 6.24 5.88 7.23 5.91 6.10 6.04 5.10

τrad

τmd

η

3.80±0.12 3.82±0.13 4.01±0.08 4.10±0.10 4.33±0.02 3.83±0.06 4.23±0.11 4.01±0.09 3.99±0.05 4.28±0.07

1.87 1.84 1.80 1.77 1.72 1.57 1.58 1.80 1.85 2.24

49±1 48±1 45±1 43±1 40±1 41±1 37±1 45±1 46±1 52±1

σe×τmd (±0.10) 12.40 12.13 10.98 11.04 10.11 11.35 9.34 10.98 11.17 11.42

Ω2 *

Ω4 *

Ω6 *

5.05 5.04 4.95 4.98 5.01 6.40 5.01 4.95 4.85 4.41

3.12 3.10 2.67 2.79 2.40 2.96 2.55 2.67 2.54 2.52

1.10 1.08 1.00 1.02 0.96 1.52 1.03 1.00 0.95 0.60

*The error of fitting is no more than 25% for all the samples

JO analysis is the most useful tool in estimating the forced electric dipole transitions of rareearth (RE) ions. According to the JO theory, three intensity parameters (Ωt, t = 2, 4, 6), which depend on the local chemical environments, determine the radiative transition rate (Arad) of RE ions. Of the three parameters, Ω2 is the most sensitive to local structure, and previous studies have suggested that it increases with the degree of asymmetry around RE ions [25,26]. The other two parameters, Ω4 and Ω6 are not so much susceptible to the environmental changes as Ω2. However, Ω6 has been found to increase with the extent of the overlap between the 4f and 5d orbitals of Er3+ ions, or instead, it decreases as the 5d electron density diminishes [27]. JO parameters can be obtained by least-squares fitting the experimental to theoretical oscillator strengths for at least four absorption bands. The results are shown in Table 2. The JO parameters decrease with Bi2O3, similar results have been found elsewhere [17,18]. The decrease in Ω2 stems from the formation of highly polarizable Bi-O bonds with Bi2O3, reducing the diversity of the microsymmetries around Er3+ ions, similar to glasses containing Pb2+ ions that have the same electronic configuration as Bi3+ [28]. On the other hand, the decrease in Ω6 can be explained based on the concept of OB as described above. Since OB of the samples increases with Bi2O3, the electron donating tendency of oxygen ions raises, consequently, the 6s orbitals of Er3+ ions have larger chance accepting electrons donated by oxygen ions, and this may result in decreased 5d electron density as the 6s

#248120 © 2016 OSA


electrons are assumed to shield the 5d orbital or to repulse the 5d electron and to diminish the existing probability of 5d electrons [27]. 5. Conclusion

Our results suggest that both MT and concentration ratio of Bi2O3 to SiO2 are important factors to be considered when preparing Er3+-doped bismuth glasses for fiber amplifiers. The formation of bismuth metal nanoparticles in glasses melted at high MT is the main reason impairing the SPs of the glasses. The variation of UC and NIR emissions is mainly determined by the phonon energy of the glasses although the existence of the free OH- groups cannot be ignored. A good agreement between JO and OB theories is obtained, namely, the decrease in Ω2 stems from the formation of highly polarizable Bi-O bonds with Bi2O3, which reduces the diversity of the microsymmetries around Er3+, and the decrease in Ω6 is related to the increasing electron donating power of the oxygen ions around Er3+ in glasses of larger OB. Acknowledgments

This study was supported by the National Natural Science Foundation of China (51302082, 61377096), Heilongjiang Young Researcher Support Project (1253G018),the 111 project (B13015) to the Harbin Engineering University, and State Key Laboratory of Silicate Materials for Architectures (Wuhan University of Technology) (SYSJJ2014-08).

#248120 © 2016 OSA
