Use of CsCl to Enhance the Glass Stability Range of Tellurite Glasses for Er 31 -Doped Optical Fiber Drawing Carmen Rosa Eyzaguirre, Eugenio Rodriguez, Enver Fernandez Chillcce, Se´ rgio Paulo Amaral Oso´ rio, Carlos Lenz Cesar, and Luiz Carlos Barbosa w Departamento de Eletroˆ nica Quaˆ ntica, Instituto de Fı´sica Gleb Wataghin, Universidade Estadual de Campinas— UNICAMP, Campinas, SP, Brazil Italo Odone Mazali and Oswaldo Luiz Alves Laborato´ rio de Quı´mica do Estado So´ lido, Instituto de Quı´mica, Universidade Estadual de Campinas—UNICAMP, Campinas, SP, Brazil Tellurite glasses are important as a host of Er 31 ions because of their good solubility and because they present broadband optical gain compared with Er 31 -doped silica, with the potential to in- crease the bandwidth of communication systems. However, the small glass stability range (GSR) of tellurite glasses compro- mises the quality of the optical fibers. We show that the addition of CsCl to tellurite glasses can increase their GSR, making it easier to draw good-quality optical fibers. CsCl acts like a net- work modifier in glass systems, weakening the network by form- ing Te–Cl bonds. We show that the thermal expansion coefficient mismatch is in the right direction for optical fiber fabrication purposes and that the Bi 2 O 3 content can be used to control the refractive index of clad and core glasses. Single- mode and multi-mode Er 31 -doped optical fibers were produced by the rod-in-tube method using highly homogeneous TeO 2 – ZnO–Li 2 O–Bi 2 O 3 –CsCl glasses. I. Introduction T ELLURITE glass optical fibers are important for telecommu- nications because the Er 31 ion fluorescence bandwidth at 1550 nm in this host is much broader than in other glasses at the same time as the solubility limit for the rare earth is much high- er. 1,2 This means that it can be doped up to 70000 p.p.m., al- lowing higher gain per unit length. 3 This fact can be understood considering that Er 2 O 3 is a network intermediate/modifier and not only a dopant in this kind of glass. The glass could even be called TeO 2 –Er 2 O 3 glass. Used in optical amplifiers, these fibers should show broadband optical amplification for wavelength division multiplexing systems, increasing the number of wave- lengths of the optical channels from those of the usual Er 31 - doped silica fibers. 3,4 Moreover, due to the higher doping level and attenuation, these devices should require only a few centi- meters of fiber length, instead of the usual tens of meter scale length that has been used up to now. 5–7 However, optical fiber production with this glass has been a challenge mainly due to its low glass stability range (GSR), which leads to crystallization processes at the moment of optical fiber drawing. Two other problems one has to face are how to control the difference in the refractive index and the expansion coefficient mismatch between the core and the clad glasses of the optical fibers. Since the possibility of large amplification band- width has been recognized, several tellurite glass compositions have been tried; among them, the TeO 2 –ZnO–Na 2 O–Er 2 O 3 and TeO 2 –WO 3 –Er 2 O 3 families have generated international pat- ents 8,9 and use the Bi 2 O 3 content to control the core/clad re- fractive index and viscosity. The aim of this paper is to show that the GSR of tellurite glasses can be enhanced by adding CsCl, without losing any of the other characteristics of the tellurite glasses that made it promising for Er 31 -doped optical amplifiers. To show this, we present the thermo-physical results obtained with bulk TeO 2 – ZnO–Li 2 O–Bi 2 O 3 –CsCl (TeZnLiBiCsCl) glass samples and the Er 31 -doped optical fiber produced with them by the rod-in-tube method. II. Experimental Procedure To synthesize the TeZnLiBiCsCl glass samples, we melted pow- dered chemical compounds with 99.999% purity in gold cruci- bles at 1023 K in a resistance furnace in an oxygen-controlled atmosphere over 2 h. After fusion, the glass was quenched in stainless-steel molds, followed by a thermal treatment (anneal- ing) at 513 K for 2 h to avoid the formation of internal stresses, and then cooled naturally to room temperature. Tables I and II show the nominal compositions of the several samples synthesized. The chemical composition of the synthesized TeZnLiBiCsCl glasses was evaluated by energy dispersive X-ray spectroscopy (EDX) microanalysis in a JEOL (Tokyo, Japan) JSM 6360-LV scanning electron microscope coupled to a Noran System (Waltham, MA) SIX Model 6714A micro- analyzer (operating with a beam energy of 20 kV), X-ray fluo- rescence (XRF) spectroscopy (EDX 700, Shimadzu, Tokyo, Japan) and by an ion selective electrode potentiometric method. For the last analysis, the TeZnLiBiCsCl-d glass sample was dis- solved in 75 mL of 2.0 mol/L aqueous HNO 3 solution at 298 K. The non-crystallinity of the samples was confirmed by X-ray diffraction analysis using CuKa radiation with a Shimadzu XD3A diffractometer. Typical temperatures related to glass sta- bility: T g (glass transition temperature), T x (onset of crystalliza- tion temperature), and T m (melting temperature) were measured by differential thermal analyses (DTA, TA-50WS, Shimadzu), while T d (softening dilatometric point) and a (thermal expansion coefficient) were measured by thermomechanical analyses (TMA, TA 50WS, Shimadzu). The DTA measurements were performed with both powdered (particle diameter around 64 mm) and bulk glass samples at a heating rate of 10 K/min under an argon atmosphere (20 mL/min). To measure the thermal ex- pansion coefficient by TMA, we used cylindrical samples with polished parallel faces. The refractive indices (n) were measured using a prism-cou- pling method at 632.8, 1305.4, and 1536.0 nm on polished glass D. Johnson—contributing editor This work was financially supported by FAPESP and CNPq. w Author to whom correspondence should be addressed. e-mail: oalves@iqm.unicamp.br Manuscript No. 22265. Received September 18, 2006; approved February 12, 2007. J ournal J. Am. Ceram. Soc., 90 [6] 1822–1826 (2007) DOI: 10.1111/j.1551-2916.2007.01705.x r 2007 The American Ceramic Society 1822