1 Introduction

The TeO₂-based glasses form a highly interesting field of recent studies aiming at basic scientific goals as well as research for new promising materials for non-linear optics. However, whereas the exceptional dielectric and optical properties of those materials are usually associated with the E-pairs of Te(IV), the nature of their mechanical and thermal properties remains an open question. Evidently, such properties must be related to the presence of the Te–_eqO_ax–Te bridges, which always exist in the condensed TeO₂, and in tellurite structures, having complex Te_nO_m anions 〚1〛. Those bridges, as a rule, are essentially asymmetric, since the equatorial Te–O bonds are mainly covalent, and the axial ones are mainly electrostatic 〚2〛.

Thus, it can be thought that the properties in question can be controlled by the polymerisation (homogenisation) of the Te–O–Te bridges. In particular, this can be done by introducing the fluorine atoms into the coordination spheres of tellurium. In the cases of the Te–O–F system, the atoms of fluorine (being more electronegative than oxygens) would accept some electrons of tellurium, thus allowing each Te atom to form the two simple Te–O bonds 〚3〛, which can polymerise into the Te–O chains, having Te–O–Te bridges similar to Si–O–Si bridges. In those chains, the TeO₂ fragments can be readily distinguished, like the SiO₄ units can be distinguished in the SiO₂ framework.

Thus, whereas the pure tellurium dioxide systems would have quasi-isolated TeO₂ units that cannot polymerise 〚4〛, it appears that the existence of polymerised Te–O chains is intrinsic for Te–O–F systems. Such a sharp difference in the electronic structure and crystal chemistry of those objects must be reflected on their fundamental physical properties. At present, a number of works have been devoted to the investigations of tellurium–oxyfluoride mixed systems 〚5–12〛.

Our previous study of the TeO₂–TeF₄ system has clearly shown the influence of F atoms as a factor enlarging the glass forming domain and favouring the structural evolution from a three-dimensional (but weakly polymerised) network (TeO₂) into a well-pronounced chain-like one (TeOF₂) via a sheet-like framework (Te₂O₃F₂) 〚3, 13〛. From these points of view, we are now investigating the TeO₂–Tl₂O–TlF system. Our previous study of the TeO₂–Tl₂O system clearly demonstrated that non-linear refractive indices of thallium tellurite glasses could be up to 100 times as large as those of SiO₂ and that increasing the thallium content leads to the transformation of TeO₄ disphenoids into isolated TeO₃ trigonal pyramids and to the decrease of the number of Te–O–Te linkages 〚14〛. In this work, we have evidenced the glass-forming domain within the TeO₂–TlF and Tl₂Te₃O₇–TlF systems. The densities and the thermal behaviour of the glasses (glass transition and crystallisation temperatures) have been determined. A structural study of thallium fluorotellurite glasses has also been approached.

2 Experimental

Glassy samples were prepared by first melting at 800 °C for 20 min in sealed gold tubes and then air-quenching intimate mixtures of TlF (Aldrich, 99.9%) and TeO₂ (for the study of the TeO₂–TlF system) or TlF and Tl₂Te₃O₇ (for the study of the Tl₂Te₃O₇–TlF system). TeO₂ was prepared by decomposition at 550 °C under flowing oxygen, of commercial H₆TeO₆ (Aldrich). Tl₂Te₃O₇ was obtained by heating, at 350 °C for 18 h then at 250 °C for 24 h, in a gold crucible under pure flowing nitrogen, various mixtures of high purity Tl₂CO₃ (Aldrich, 99%) and TeO₂. Glass-formation domains were determined by using X-ray diffraction (Guinier–De Wolff camera, Cu Kα radiation). Glass transition (T_g) and crystallisation (T_c) temperatures were measured by heat flux differential scanning calorimetry (Netsch STA 409). The powdered samples (≈ 30 mg) were introduced into non-sealed platinum crucibles and the DSC curves were recorded between 20 and 800 °C maximum, using a heating rate of 10 °C min^–1, under pure nitrogen atmosphere. The densities of glassy samples were measured on finely ground powders by helium pycnometry (Accupyc 1330). A structural approach of the glasses was realised using Raman scattering. The Raman spectra were recorded in the 80–1200 cm^–1 range using a Dilor spectrometer (XY model), equipped with a CCD detector and an Ar⁺ laser (514.5 nm exciting line) in a backscattering geometry. With such a detector, a good signal/noise ratio needed one or at most two scans (during 100 s). The samples were visually controlled through a microscope (× 100), which has enabled us to analyse a surface with diameter of about 2 μm. The diameter of the laser spot focused on the sample was about 1 μm. Reproducibility of the results thus obtained was a proof of identity and homogeneity of the samples. Measurements were made at low power (< 100 mW) of the excitation line, in order to avoid any damage of the samples. The spectral resolution was about 4.5 cm^–1 at the exciting line.

3 Results and discussion

3.1 Formation and thermal behaviour of glasses

Under our experimental conditions, melting at 800 °C and then air quenching, large glassy domains have been evidenced within the TeO₂–TlF (from 7 to 65 mol% of TlF) and Tl₂Te₃O₇–TlF (from 0 to 75 mol% of TlF content) pseudo-binary systems (Fig. 1). With respect to the glassy domain observed within the TeO₂–Tl₂O system 〚15〛 under the same conditions of elaboration, 15 to 52 mol% of TlO_0.5, it is clearly seen that addition of F atoms allows enlarging the glass-forming domain. Within the TeO₂–TlF system, the glasses are all homogeneous and white; their density increases from 5.7 to 7.06 g cm^–3 with increasing the TlF content. Within the Tl₂Te₃O₇–TlF system, the glasses are also homogeneous. Their colour changes from yellow to white with increasing the TlF content; their density increases from 6.72 to 7.21 g cm^–3 with increasing the TlF content. The evolutions within these two systems of the glass transition and crystallisation temperatures, as a function of the composition, are shown in Fig. 2. As a comparison, the evolution observed with TeO₂-Tl₂O glasses is also given.

Within our two fluorotellurite systems, T_g and T_c temperatures continuously decrease with increasing the TlF content, as previously observed with increasing the TlO_0.5 content. This decrease can be attributed to the thallium addition. It is interesting to note that the difference between T_c and T_g, which characterises the thermal stability of the glass, remains practically constant with increasing TlF content (for all the compositional range within the TeO₂–TlF system and at least up to about 65 mol% of TlF within the Tl₂Te₃O₇–TlF system). However, the thermal stability of TeO₂–TlF glasses is slightly lower than that observed with thallium tellurite glasses. This can indicate that the substitution of O atoms by F ones would decrease the thermal stability of glassy samples. Above 60 mol% of TlF content, Tl₂Te₃O₇–TlF glasses are characterised by a sharp decrease of their T_c temperature, whereas their T_g temperature decreases slowly, thus reducing their thermal stability. However, such glasses exhibit a thermal stability in the range 55–95 °C considerably superior to that of TeO₂–TlF ones. A study of the crystallisation (especially the nature of the first phase that crystallises from glass) of such samples is now in progress using in-situ powder XRD. Preliminary results have confirmed the existence of the metastable γ-TeO₂ polymorph 〚4,16〛 for TeO₂-rich samples.

3.2 Structural aspects of the TeO₂–TlF glasses

The structural aspects of our study were mainly considered by analysing the evolution of the Raman spectra of the x TeO₂/(1 – x) TlF system (0.35 ≤ x ≤ 1), and comparing it with that of the x TeO₂/(1 – x) TlO_0.5 system. Raman spectra of the latter (Fig. 3b) show that with decreasing x, the solid TeO₂ continuously evolves into the Tl₂TeO₃ system, in which all the TeO₂ molecules become the isolated 〚TeO₃〛^2– orthoanions having the two principal vibrations: a stretching pulsation (near 725 cm^–1) and a symmetric bonding of the TeO₃ pyramids (≈ 300 cm^–1). Another type of structural evolution can be proposed from the analysis of the Raman spectra of the x TeO₂/(1 – x) TlF system. Actually, for 0.6 ≤ x ≤ 1, the Raman spectrum keeps its view typical for (pure) glassy TeO₂. Beginning from x = 0.6, the quantity of glassy TeO₂ sharply drops: the intense band near 450 cm^–1 (indicating Te–O–Te bridges) disappears, and a strong band appears above 750 cm^–1, which indicates the formation of isolated groups having short (essentially non-bridging) Te–O bonds; a very weak band around 300 cm^–1 can be attributed to the bonding vibrations of O–Te–O angles.

Thus, no indication of the Te–O chains (i.e. Te–O–Te polymerisation) was observed in the Raman spectra. This fact could be explained by the presence of Tl⁺ cations. To our belief, for 0.6 ≤ x ≤ 1, the two phases, TlF and TeO₂, coexist in the glasses independently. The mutual influence begins when x ≤ 0.6, and it results in a reaction of TeO₂ molecules with fluorine atoms, thus forming isolated groups containing non-bridging Te–O bonds. The concrete structure of those groups will be the object of our further XRD study of the relevant crystalline phases.

Acknowledgements

One of us, C. Lasbrugnas is grateful to the ‘Conseil régional du Limousin’ for financial support.