In the present work, tellurite 20Li 2 O–80TeO 2 glasses were prepared with identical nominal composition under different glass-forming histories to produce a stressed and stress-free samples. X-ray Diffraction (XRD) and Differential Scanning Calorimetry (DSC) techniques were used to study the effects of the glass-forming histories on the thermal and structural properties of these glasses. The γ-TeO 2 (metastable), α-TeO 2 and α-Li 2 Te 2 O 5 phases were identified during the controlled devitrification in these glasses. The mestastable character of the γ-TeO 2 phase was clearly observed in the glass under stress but this effect is not so clear in the stress-free glass. The γ-TeO 2 and α-TeO 2 phases crystallizes during the initial stages of crystallization in both studied glasses while the α-Li 2 Te 2 O 5 phase crystallize in the final stages of the crystallization. The activation energies and Avrami exponent were calculated for both studied glasses with different particle size leading to E 3 > E 2 > E 1 for stressed glass and E 3 > E 2 ≈ E 1 for stress-free glass, where E 1 , E 2 and E 3 were associated to the γ-TeO 2 , α-TeO 2 and α-Li 2 Te 2 O 5 phases, respectively. The observed distinct n ¯ 1 < n ¯ 2 < n ¯ 3 in both glasses is an indicative that nucleation and growth takes place by more than one mechanism in the early stages of the crystallization. Keywords Nucleation Crystallization Tellurite glasses 1 Introduction TeO 2 -based glasses are known for their excellent nonlinear optical properties and large nonlinear refractive indices. Depending on the composition they present a third harmonic ( χ 3 ) generation almost one order greater than some other important oxide glasses [1,2] . Nowadays, the crescent use of tellurite glasses in some areas of technology [3,4] justify the studies on fundamental aspects to obtain a high performance in a specific device. The physical properties of binary TeO 2 –Li 2 O glasses have been extensively studied in past two decades to understand different properties of this important glass matrix. These studies were centered on the linear [5,6] and non-linear [7,1] optical properties, and structural [8,9] , thermal [10] and electrical properties [11,12] . Despite the central interest in optical non-linear properties, are crescent the studies on the structural and thermal properties of tellurite glasses due to the crescent necessity of a clear comprehension of the nucleation and crystal growth mechanisms in these systems. Physical properties of glasses depend strongly on the thermal histories. As a result, the glass formation, residual stresses, the number of nuclei formed during glass preparation and the relaxation process have an intrinsic dependence on the thermal history of a glass [13,14] . Adequate knowledge of the correlation between physical properties and the thermal history is indispensable for a correct interpretation of several processes in glasses, in particular the crystallization kinetics. In some cases, studies of these properties are vital for preparing high quality glass for use in various solid-state and optoelectronic devices. Thus, the purpose of the present work is to study the effects of the thermal history on structural and thermal properties of 20Li 2 O–80TeO 2 glasses by using X-ray Diffraction (XRD) and Differential Scanning Calorimetry (DSC) as investigative tools. 2 Experimental The 20Li 2 O–80TeO 2 glasses examined in this work were prepared by using the conventional melt quenching method starting from commercial powder reagents Li 2 CO 3 (CBMM, 99.9%) and TeO 2 (Alfa Aesar, 99.95%). After mixing in appropriate proportions using a 15 g batch weight, the powder was melted inside a platinum crucible at 90 °C for 30 min in an electric furnace. Two glasses were prepared with identical 20Li 2 O–80TeO 2 nominal composition but with distinct thermal histories. In the present article, these samples were referred as S1 and S2 glasses. For S1 glass, the melt was poured directly into a polished metal mould at room temperature and removed from the mould after 12 h. So, this glass was not submitted a heat annealing after quenching to remove residual stresses. For S2 glass, the melt was poured into the same mould pre-heated in an electric furnace at 250 °C and immediately placed back again in the electric furnace and maintained at this temperature for 12 h when the furnace was cooled down at 1 °C min −1 to the room temperature. Then, the glass was finally removed from the mould. The S1 and S2 glasses were screened at four different particle size (19 μm, 41 μm, 54 μm and 69 μm) in average and stored until their use for DSC technique. Glasses with 69 μm particle size in average were used to study the devitrification by XRD technique. To study the crystallization in the glasses, S1 glasses were previously heat annealed in an electrical furnace at 304 °C, 314 °C and 319 °C for 5 min while S2 glasses were heat annealed at 349 °C, 359 °C and 369 °C for 5 min. The extrapolated onset of glass transition temperature ( T g ) and the extrapolated onset crystallization temperature ( T x ) were obtained from DSC curves. To determine activation energy ( E ) for the crystallization, glasses with different particle size were crystallized inside DSC equipment and DSC curves were recorded at four different heating rates (=2.5, 5.0, 7.5 and 10.0 °C min −1 ). Data from DSC measurements were analyzed by using the Kissinger [15] equation ln ( T p 2 / ϕ ) = E / RT p + constant , where T p is a temperature corresponding to the maximum of the DSC crystallization peak and R is the gas constant. The activation energy was calculated from the slopes of the linear fits to the experimental data from a plot of ln ( T p 2 / ϕ ) versus 1/ T p for different particle size. The Avrami exponent n could be determined at glasses with different particle size from DSC values of the crystallization rate ( dx/dt ) p considering the following equation [16,17] : n = ( dx / dt ) p RT p 2 ( 0.37 ϕ E ) - 1 . In this equation, x is the volume fraction which has been crystallized after time t behind starting crystallization, the subscript p denotes the magnitude values corresponding to the maximum crystallization rate, R is the gas constant, φ is the DSC heating rate and E is the activation energy for the crystallization. 3 Results The structure of the crystalline phases induced in the glasses during the devitrification was studied by XRD technique. Fig. 1 shows XRD patterns of the S1 and S2 glasses. For S1 glass, a typical characteristic of amorphous materials can be observed for the as-quenched glass ( Fig. 1 A) and heat annealed glass at 304 °C for 5 min ( Fig. 1 B). On the other hand, the XRD patterns of the glasses heat annealed at 314 °C ( Fig. 1 C) and 319 °C ( Fig. 1 D) shows a glass–ceramic characteristic. For the S2 glass a similar crystallization study was performed. For this glass, a typical characteristic of an amorphous material was observed for the as-quenched glass ( Fig. 1 E) and heat annealed glass at 349 °C during 5 min ( Fig. 1 F), while the XRD patterns of the glasses heat annealed at 359 °C ( Fig. 1 G) and 369 °C ( Fig. 1 H) exhibit a typical glass–ceramic characteristic. The DSC technique was used in the present work to understand the thermal properties of the studied glasses. Fig. 2 shows DSC crystallization peaks as a function of the temperature, recorded at different heating rates and different particle sizes, for both S1 and S2 glasses. The insets in this figure refer to DSC curves around glass transition temperature. Table 1 summarizes the glass transition temperature ( T g ) and onset of crystallization temperature ( T x ) of the S1 and S2 glasses as a function of particle size at 10 °C min −1 heating rate from DSC curves shown in Fig. 2 . In order to evaluate the crystallization as a function of temperature, the DSC crystallization peaks were fitted using a computational least-square method considering three lorentzian functions, each one representing a different crystalline phase in the glasses during the devitrification. Then, Fig. 3 shows fitting for selected DSC crystallization peaks recorded at recorded at 10 °C min −1 for S1 and S2 glasses with 69 μm particle size in average. The results from these fittings were used to determine the activation energies for individual peaks and the determination of Avrami ( n ¯ ) parameters associated to these peaks. Fig. 4 shows plots of ln ( T p 2 / ϕ ) versus 1/ T p for the three crystallization peaks. Based on Fig. 4 the activation energies for S1 and S2 glasses were calculated and summarized in Table 2 as a function of different particle size. Based on DSC crystallization peaks fittings, the mean Avrami ( n ¯ ) parameters were calculated for both S1 and S2 glasses at different particle size and also summarized in Table 2 . 4 Discussion The structure of the 20Li 2 O–80TeO 2 glasses was initially studied using the XRD technique. Fig. 1 shows XRD patterns of both S1 and S2 glasses for 69 μm particle size heat annealed at different temperatures for 5 min in an electric furnace. As we can see, the XRD patterns of the S1 as-quenched glass ( Fig. 1 A) and heat annealed at 304 °C for 5 min ( Fig. 1 B) shows a typical amorphous characteristic. However, when the glass is heat annealed at 314 °C and 319 °C for 5 min ( Fig. 1 C and Fig. 1 D, respectively) a partially crystallized material is obtained. The indicated peaks in Fig. 1 C were attributed to α-TeO 2 (paratellurite) and γ-TeO 2 crystalline phases permeating the glass matrix. For glass heat annealed at 319 °C for 5 min ( Fig. 1 D), the amorphous phase tends to disappear and an additional Li 2 Te 2 O 5 crystalline phase was also observed in coexistence with both α-TeO 2 and γ-TeO 2 crystalline phases. Fig. 1 E and F shows a typical characteristic of amorphous materials for the S2 as-quenched and heat annealed glass at 349 °C, while glasses heat annealed at 359 °C and 369 °C for 5 min exhibit a typical glass–ceramic characteristic, as shown in Fig. 1 G and H, respectively. At this moment is pertinent discuss about the nature of the crystalline phases observed in the glasses. The crystalline α-TeO 2 and γ-TeO 2 phases are two polymorphs of tellurium dioxide TeO 2 , but the γ-TeO 2 phase is considered a metastable structure [18,19] . Both structures are essentially built up from similar basic TeO 4 units interconnected by the Te–O–Te simple bridges but present small differences. While the structure of the α-TeO 2 are formed by a three-dimensional network TeO 4 units sharing oxygen corners by symmetric Te–O–Te bridges, the γ-TeO 2 structure can be considered as a chain system, where TeO 4 units are alternatively linked by nearly and highly symmetric Te–O–Te bridges [19] . Our XRD results, particularly observed for the S1 glass, suggest a crystallization hierarchy on 20Li 2 O–80TeO 2 glass by controlled heating. It is clear for the S1 glass ( Fig. 1 C) that α-TeO 2 and γ-TeO 2 phases crystallization occurs before the Li 2 Te 2 O 5 phase in the studied glass. On pure TeO 2 glass reported in the literature, the γ-TeO 2 appears as the first crystalline phase [20] . Results obtained for S2 glass shows a slightly different behaviour, indicating that crystallization of the α-TeO 2 and γ-TeO 2 phases occurs practically simultaneously, which could not be detected by the resolution of XRD technique and used conditions to induce the crystallization ( Fig. 1 G and H). However, based only on XRD results, at this moment it was not possible to determine if α and γ-phases crystallize simultaneously or at distinct onset crystallization temperatures in the studied glass. Based on XRD patterns of the best crystallized S1 and S2 glasses, Fig. 1 D and Fig. 1 H, respectively, the cell parameters of the observed crystalline phases were calculated by using the Powder4 program, written by N. Dragoe of the Université Paris Sud, Orsay, France. For the α-TeO 2 , the cell parameters of the tetragonal phase were calculated considering the selected (1 1 0), (1 0 2), (2 0 0), (0 0 4), (2 1 2), (2 2 0), (2 2 1), (3 1 0) and (3 1 1) reflections. For S1 glass these observed ( h k l ) reflections are located respectively at the following 2 θ values 26.34°, 30.08°, 37.38°, 47.74°, 48.70°, 53.86°, 55.36°, 60.90° and 62.24° ( Fig. 1 D) and for S2 glass 2 θ values were 26.18°, 29.96°, 37.38°, 47.74°, 48.70°, 53.86°, 55.36°, 60.90° and 62.24° ( Fig. 1 H). The calculated cell parameters for the α-TeO 2 tetragonal phase were a = 4.810(0) Å and c = 7.615(1) Å for S1 glass and a = 4.801(6) Å and c = 7.616(0) Å for S2 glass. For the crystalline γ-TeO 2 , the cell parameters of the orthorhombic phase were calculated by using the (0 1 1), (1 2 0), (0 2 1), (1 1 1) and (0 1 2) reflections at the observed 2 θ around 22.90°, 27.78°, 28.90°, 31.28° and 43.02° for S1 glass ( Fig. 1 D) and at 22.96°, 27.68°, 28.76°, 29.38° and 42.90° for S2 glass ( Fig. 1 H). The calculated cell parameters of the orthorhombic metastable phase were a = 4.327(9) Å, b = 9.150(5) Å and c = 4.313(1) Å, for S1 glass, and a = 4.805(8) Å, b = 8.785(3) Å and c = 4.344(7) Å, for S2 glass. Similarly, for the α-Li 2 Te 2 O 5 orthorhombic phase were used the (1 1 1), (0 2 0), (4 2 0) and (9 3 0) reflections associated to the observed 2 θ around 20.76°, 21.92°, 25.38° and 47.04° in the XRD pattern of the S1 glass ( Fig. 1 D) and at 20.48°, 21.82°, 25.70° and 46.94° for S2 glass ( Fig. 1 H). At this last one phase, the calculated cell parameters were a = 24.267(1) Å, b = 8.312(4) Å and c = 5.099(4) Å, for S1 glass, and a = 24.509(6) Å, b = 8.258(2) Å and c = 5.209(3) Å, for S2 glass. The obtained cell parameters of the paratellurite are in good agreement with those values reported in the literature for α-TeO 2 single crystals [18] , as well the parameters obtained for the tellurium dioxide γ-TeO 2 in the present work are in good agreement with those results reported in the literature [19] . Finally, the calculated cell parameters for α-Li 2 Te 2 O 5 also agrees with results reported in the JCPDS 25-1381 data card, where a = 24.1770 Å, b = 8.1820 Å and c = 5.1910 Å. Fig. 2 shows DSC crystallization peaks for both S1 and S2 glasses as a function of the temperature and different particle size, recorded at different heating rates ( φ ). The inset refers to the DSC curves at around glass transition temperature ( T g ). Some fundamental differences can be observed on shapes of the crystallization peaks for both glasses. As can be observed in Table 1 , at S1 glass the T g temperature is independent particle size and presents essentially the same value at around 265 °C. At S2 glass, the T g temperature presents essentially the same behaviour but a value at around 267 °C, slightly higher than observed for S1 glass. The T x temperature trend to a mean value at around 333 °C for S1 glass, except for glass with 41 μm particle size, which present a value at around 337 °C. For S2 glass, T x shifts from 332 °C to 326 °C when particle size decreases. These variations, more pronounced in S2 glass, may be associated with particle size effects on heat transfer during DSC scans, since larger particles would have greater heat transfer resistance when compared to small particles, and with the thermal history of this stress-free glass. The observed asymmetry at the crystallization peaks for the S1 glass in Fig. 2 suggests the existence of at least two distinct phase transformations in the glass matrix. This effect is more pronounced at smaller particle size (19 μm) in the S1 glass than glass with greater particle size (69 μm). At smaller particle size in the S1 glass the asymmetry became more evident and splits in two distinct peaks at around 344 and 356 °C for DSC curve recorded at 10 °C min −1 . At a fix particle size, decreasing the heating rate at 2.5 °C min −1 in DSC scan both peaks became more evident in S1 glass. In addition, considering the S1 glass in Fig. 2 , the shoulder observed for the peak located at lower temperature, at smaller heating rate, became more evident when the particle size increase from 19 μm to 69 μm, suggesting in fact the existence of a third crystalline phase. These results are agreement with results obtained from XRD technique shown in Fig. 1 . Thus, the combinations of particle size and heating rate on recorded DSC curves have evidenced in the S1 glass at least three distinct possible phase crystallizations in the studied glass. On the other hand, the crystallization peaks for S2 glasses are much more symmetric than observed for S1 glasses, as shown in Fig. 2 , which shows also a slight asymmetry observed on DSC crystallization peak at the smallest heating rate for all particle sizes in S1 glass. This asymmetry suggests the existence of multiple crystalline phases during DSC scans. This is an important result because we can observe in Fig. 2 that the effect of thermal history is much more pronounced than the effect of particle size in the studied glasses. The symmetric crystallization peaks observed in DSC curves suggest that the three crystalline phases crystallizes at very close temperatures in the S2 glass and this can be a direct effect from released stress during thermal annealing applied to this glass. The inset in Fig. 2 shows DSC curves at around the glass transition temperature. The DSC heating curves of the S1 glass demonstrate a typical glass transition steps for all particle size while DSC curves of the S2 glass shows an evident endothermic peak immediately above T g temperature. This result provide a clear evidence that the observed endothermic peak close to T g have no direct correlation with residual stress in the tellurite glasses studied in the present work, considering the additional fact that S2 is a stress-free sample glass. The nature of the endothermic peak close to T g in the DSC heating curves can be associated to different factors. In glassy polymers, the magnitude of this endothermic peak is proportional to the enthalpy loss and increases with the annealing time [21,22] . Frequently the behaviour of DSC endothermic peak in glasses has been also associated to residual stresses [23] and to molecular motions [24] . However, the DSC endothermic peak above T g can also be interpreted in terms of the enthalpic relaxation, whose kinetics is based on the thermodynamic temperature and the instantaneous structure of the glass [14] . Under this point of view, the enthalpic peaks observed in DSC curves above T g in the present work depend on thermal history of glasses. In the present work, an intrinsic dependence of DSC endothermic peak above T g with the heat annealing could be observed, since the S2 glass was annealed at temperatures very close to T g (250 °C for 12 h) to release stresses while S1 glass was not submitted to a heat annealing to release stresses. Nevertheless, it is necessary explore another possibilities before exclude the direct relationship between residual stresses and the observed endothermic DSC peaks at the studied glasses. To discuss the possibility of molecular motions influences, the endothermic DSC peaks cannot be interpreted separately. The thermal histories certainly will reflect on the crystallization peaks and consequently the interpretation of both endothermic and exothermic peaks should be studied in a common context. The thermal history of glass clearly influences the shape and position of the crystallization peaks in DSC curves. As we can see in Fig. 2 the DSC crystallization peak of S1 glass is more asymmetric than S2 glass, as discussed above, but the peak heights are essentially the same. Assuming that the DSC crystallization peak height is proportional to the concentration of nuclei in the glass [25] , it is clear that nuclei concentration in both S1 and S2 glasses are essentially the same. The density of nuclei can be expressed as the sum of the number of nuclei formed during glass preparation and the number of nuclei per unit volume that is formed during the heating in the DSC scan. If a transformation occurs essentially by internal nucleation and growth, no change will be observed on the crystallization peak height. Under these considerations, it is difficult to explain how two glasses with different thermal histories presents essentially the same nuclei concentration but a radically different behaviour at around glass transition temperature, as observed in Fig. 2 . Thus, these results suggest that the link between the crystallization peaks and endothermic peaks immediately above T g can be associated to molecular motions into the glass matrix. As a consequence, it is expected that heat annealing in the S2 glass produces smaller ordered regions than S1 glass, without necessarily increasing the nuclei number in both glasses. This can explain why the initial stages of the crystallization first occur in the S1 glass. The activation energies E were determined considering the presence of three crystalline phases in the glass, as indicated in DSC curves in Fig. 3 , from the linear fitting of ln ( T p 2 / ϕ ) versus 1/ T p plots in Fig. 4 for both S1 and S2 glasses. The calculated values were summarized in Table 2 as a function of different particle size. The magnitude of activation energies obtained in the present work is comparable to those reported in the literature for 30Li 2 O-70TeO 2 glass [8,10] . Based on our previous discussions the energies E 1 , E 2 and E 3 can be respectively associated to the following γ-TeO 2 , α-TeO 2 and α-Li 2 Te 2 O 5 crystallization in the 20Li 2 O–80TeO 2 glass. The calculated energies for S1 glass showed distinct values such as E 3 > E 2 > E 1 , while for S2 glass the tendency E 3 > E 2 > E 1 ( E 1 ≈ E 2 ) was verified independent of the particle size. The energies values obtained corroborate the selective crystallization observed for S1 glass and is an indicative that both γ-TeO 2 and α-TeO 2 phase crystallization occurs at very close temperatures for S2 glass. The crystallized volume fractions x ( T ) were calculated for S1 and S2 glasses for all particle size, considering each DSC deconvoluted crystallization peak, as shown in Fig. 3 . Based on ( dx/dt ) versus temperature plots, not shown here, the maximum crystallization rates ( dx/dt ) p were obtained. From these results and the activation energies, the mean Avrami n ¯ exponent were calculated for glasses with different particle size and finally summarized in Table 2 . As we can see, the n ¯ values obtained for S1 glasses with 69 μm and 19 μm particle sizes were such as n ¯ 1 < n ¯ 2 ≈ n ¯ 3 . However, S1 glasses with intermediate particle size (54 μm and 41 μm) the obtained Avrami parameters were slightly different such as n ¯ 1 < n ¯ 3 < n ¯ 2 . On the other hand, distinct n ¯ values were obtained for S2 glasses with 69 μm and 54 μm particle sizes such as n ¯ 1 < n ¯ 2 < n ¯ 3 . In the same glass with particle size 41 μm and 19 μm the following behaviour n ¯ 1 ≈ n ¯ 2 < n ¯ 3 was observed. The Avrami exponent obtained from non-isothermal measurements may be correlated to the crystallization mechanism in the glass [26,27] . The n exponent obtained in the present work was slightly greater than n = 1.4, if compared to results obtained for 30Li 2 O–70TeO 2 glasses [10] . However, differently the usual procedure, in the present work the crystallization peaks were deconvoluted in three peaks, which were associated to the different crystalline phases formed in the glass. The observed differences at the mean Avrami exponent, for glasses with different particle size, are a clear indicative of the thermal histories effects in the glasses. The observed distinct n ¯ 1 < n ¯ 2 < n ¯ 3 in the S1 or S2 glasses is an indicative that nucleation and growth takes place by more than one mechanism in the early stages of the crystallization. The Avrami exponent n ≈ 2 indicates one-dimensional crystallization with a constant bulk nucleation rate, n ≈ 3 indicates the existence of the volume nucleation and the two-dimensional growth mechanism, while n ≈ 4 indicates volume nucleation and three-dimensional growth mechanism [26] . For n > 4, it is expected a three-dimensional crystallization with increasing nucleation rate followed by a controlled crystal growth. Thus, results shows that the three deconvoluted crystallization peaks exhibit a distinct nucleation and growth for S1 glasses with 54 and 41 μm particle size and S2 glasses for 69 and 54 μm particle size. Consequently, the increasing in n ¯ exponent from n 1 to n 3 in these both glasses suggests that growth mechanism changes from the initial one-dimensional crystallization with a constant nucleation to the three-dimensional crystal growth with nucleation rate increasing. In addition, for S1 glasses with 69 and 19 μm particle size the result n ¯ 1 < n ¯ 2 ≈ n ¯ 3 suggests that both second and third deconvoluted crystallization peaks presents a nucleation and three-dimensional growth mechanism. On the other hand, for S2 glasses with 41 and 19 μm particle size the both first and second deconvoluted peaks exhibit similar nucleation and crystal growth uni-dimensional while the third peak indicate a nucleation and three-dimensional growth mechanism. These results for the Avrami parameter confirm that S1 and S2 glasses, prepared under different thermal histories, exhibit different crystallization kinetic. 5 Conclusion In summary, 20Li 2 O–80TeO 2 glasses were prepared under different thermal histories producing stressed and stress-free samples. XRD results obtained reveal the crystallization of the γ-TeO 2 , α-TeO 2 and α-Li 2 Te 2 O 5 phases in the studied glasses. The mestastable character of the γ-TeO 2 phase was clearly observed in the glass prepared under stress (S1) but this effect is not so clear in the stress-free glass (S2). The results also confirm that the γ-TeO 2 and α-TeO 2 phases crystallizes during the initial stages of crystallization in both studied glasses while the α-Li 2 Te 2 O 5 phase crystallize in the final stages of the crystallization. The non-isothermal measurements suggested that heat annealing immediately after quenching at temperatures around T g is favorable to produce local ordered regions in the glass without necessarily increasing the nuclei number. Considering the deconvolution of the crystallization peaks, the activation energies and Avrami exponent were calculated for both S1 and S2 glasses with different particle size. The energies E 1 , E 2 and E 3 could be respectively associated to the following γ-TeO 2 , α-TeO 2 and α-Li 2 Te 2 O 5 crystallization in the 20Li 2 O–80TeO 2 glass, where the calculated energies for S1 glass showed distinct values such as E 3 > E 2 > E 1 , while for S2 glass the tendency E 3 > E 2 ≈ E 1 was verified independent of the particle size. In accordance with XRD results, the energies obtained corroborate the selective crystallization observed for S1 glass and is an indicative that both γ-TeO 2 and α-TeO 2 phase crystallization occurs at very close temperatures for S2 glass. The thermal history also reflected on the Avrami parameters. The observed distinct n ¯ 1 < n ¯ 2 < n ¯ 3 in the S1 or S2 glasses is an indicative that nucleation and growth takes place by more than one mechanism in the early stages of the crystallization. Acknowledgements We would like to express our gratitude to Brazilian agencies CNPq (Research Grant No. 307607/2009-7), CAPES and FAPESP for their financial support. References [1] K. Tanaka K. Kashima K. Hirao N. Soga A. Mito H. Nasu J. Non-Cryst. Solids 185 1995 123 [2] H. Nasu O. Matsushita K. Kamiya H. Kobayashi K. Kubodera J. Non-Cryst. Solids 124 1990 275 [3] J. Zhang S. Dai G. Wang H. Sun L. Zhang L. Hu J. Lumin. 115 2005 45 [4] X. Feng T.M. Monro V. Finazzi R.C. Moore K. Frampton P. Petropoulos D.J. Richardson Electron. Lett. 41 2005 835 [5] K. Shioya T. Komatsu H.G. Kim R. Sato K. Matusita J. Non-Cryst. Solids 189 1995 16 [6] L.M. Tong L.L. Hu J.J. Zhang J.R. Qiu Q. Yang J.Y. Lou Y.H. Shen J.L. He Z.Z. Ye Optics Express 14 2006 82 [7] K. Tanaka K. Kashima K. Hirao N. Soga A. Mito H. Nasu Jpn. J. Appl. Phys. 32 1993 L843 [8] I. Avramov G. Guinev A.C.M. Rodrigues J. Non-Cryst. Solids 271 2000 12 [9] T. Taniguchi S. Inoue T. Mitsuhashi A. Nukai J. Appl. Cryst. 33 2000 64 [10] P. Balaya C.S. Sunandana J. Non-Cryst. Solids 162 1993 253 [11] A. Pan A. Ghosh Phys. Rev. B 60 1999 3224 [12] J.M. Rojo J. Sanz J.M. Reau B. Tanguy J. Non-Cryst. Solids 116 1990 167 [13] J. Holubová Z. Černošek E. Černošková J. Optoelectron. Adv. Mater. 7 2005 2671 [14] Z. Černošek J. Holubová E. Černošková M. Liška J. Optoelectron. Adv. Mater. 4 2002 489 [15] H.E. Kissinger J. Res. Natl. Bur. Stand. 57 1956 217 [16] J. Vázquez C. Wagner P. Villares R. Jiménez-Garay Acta Mater 44 1996 4807 [17] J. Vázquez P.L. López-Alemany P. Villares R. Jiménez-Garay Mater. Lett. 38 1999 423 [18] P.A. Thomas J. Phys. C: Solid State Phys. 21 1988 4611 [19] J.C. Champarnaud-Mersjard S. Blanchandin P. Thomas A. Mirgorodsky T. Merle-Mérjean B. Frit J. Phys. Chem. Solids 61 2000 1499 [20] O. Noguera T. Merle-Mérjean A.P. Mirgorodsky M.B. Smirnov P. Thomas J.C. Champarnaud-Mersjard J. Non-Cryst. Solids 330 2003 50 [21] S.E.B. Petrie J. Polym. Sci. A-2 10 1972 1255 [22] Y. Wang D. Shen Polym. J. 29 1997 860 [23] J.W.P. Schmelzer E.D. Zanotto I. Avramov V.M. Fokin J. Non-Cryst. Solids 352 2006 434 [24] K.L. Ngai J. Chem. Phys. 109 1998 6982 [25] C.S. Ray D.E. Day J. Am. Ceram. Soc. 73 1990 439 [26] K. Matusita T. Komatsu R. Yokota J. Mater. Sci. 19 1984 291 [27] J. Rocherullé M. Matecki B. Baron P. Verdier Y. Laurent J. Non-Cryst. Solids 211 1997 222

Nucleation process and crystal growth for three samples of the (20-x)Li2O–80TeO2–xWO3 glass system were studied using X-ray diffraction and differential scanning calorimetry techniques. X-ray diffraction data confirmed the amorphous characteristic of the as-quenched samples and indicated the growth of crystalline phases formed due to the thermal treatment for annealed samples. These results reveal the presence of three distinct γ-TeO2, α-TeO2 and α-Li2Te2O5 crystalline phases in the TL sample, and two distinct α-TeO2 and γ-TeO2 crystalline phases in the TLW5 and TLW10 samples. The activation energy and the Avrami exponent were determined from DSC measurements. The activation energy values X-ray diffraction data of the TLW10 glass sample suggest that γ-TeO2 phase occur before the α-TeO2. The results obtained for the Avrami exponent point to that the nucleation process is volumetric and that the crystal growth is two or three-dimensional.

Tellurite glasses were prepared with identical 20 Li 2 O -80 TeO 2 nominal compositions but with different thermal histories. Differential scanning calorimetry (DSC), X-ray diffraction (XRD) and Raman spectroscopy techniques were used to understand the effects of the thermal histories on the thermal and structural properties of these glasses. It was observed that investigated properties depend strongly on the thermal histories. DSC results suggested that annealing immediately after quenching at temperatures around the glass transition temperature (Tg) and for longer times is favorable for producing local ordered regions in the glass without necessarily increasing the number of nuclei. XRD results revealed the crystallization of the γ- TeO 2, α- TeO 2 and α- Li 2 Te 2 O 5 phases in both studied glasses. Raman spectroscopy revealed the mestastable character of the γ- TeO 2 crystalline phase, while the α- TeO 2 and α- Li 2 Te 2 O 5 crystalline phases persisted up to the final stages of the in-situ crystallization.

Tellurite 20Li(2)O-80TeO(2) glasses were prepared with identical nominal composition but with different glass forming histories. Differential scanning calorimetry and x-ray diffraction techniques were used to study the effects of glass forming histories oil the thermal and structural properties of these glasses. It was observed that the investigated properties depend strongly on the glass forming histories. XRD results reveals the crystallisation of the gamma-TeO2, alpha-TeO2 and alpha-Li2Te2O5 phases in the studied glasses. DSC results suggested that annealing immediately after quenching at temperatures around T-g and longer times are favorable to produce local ordered regions in the glass without necessarily increasing nuclei number.

Structural and thermal properties of the 20Li2O-80TeO2 glass were studied using X-ray diffraction analysis and differential scanning calorimetry techniques to understand and control the crystallization process on this glass. The γ-TeO2, α-TeO2 and α-Li2Te2O5 phases were identified during the crystallization in this glass. Activation energies and Avrami exponent n were calculated from non-isothermal measurements for glasses with different particle size. The mean values \(\bar n\) of Avrami exponent were obtained for glasses with 63–75 and 45–63 μm particle sizes such as \(\bar n_1 < \bar n_2 < \bar n_3 \), but glasses with particle size 38–45 μm and smaller than 38 g,m presented \(\bar n_1 \approx \bar n_2 < \bar n_3 \).

Tellurite 20Li2O-80TeO2 glasses were prepared with identical nominal composition but with different glass-forming histories. Structural and thermal properties of these glasses were studied by using X-Ray Diffraction Analysis (XRD) and Differential Scanning Calorimetry (DSC) techniques to understand and control the crystallization process on this glass matrix. The γ-TeO2, α-TeO2 and α-Li2Te2O5 phases were identified during the isothermal induced crystallization in this glass, where the γ-TeO2, phase is a metastable structure. The measured activation energies showed that γ-TeO2 and α-TeO2 crystallizes at very close temperatures but distinct for the Li2Te2O5 crystallization found in this glass. This result suggests that both γ-TeO2 and α-TeO2 phases first crystallizes in the glass matrix. Physical properties of glasses depend strongly on the glass-forming histories. The adequate knowledge of the correlation between physical properties and the glass-forming history is indispensable for a correct interpretation of several processes in glasses, in particular the crystallization kinetic. In recent years, TeO2-based glasses were extensively studied to understand different properties of these important glass matrixes and improve some specific characteristics of technological interest. Depending on the composition, some TeO2 glasses present very interesting non-linear properties such as a third harmonic (χ) generation almost one order greater than some other important oxide glasses. However, the technological applications of these glasses require a profound comprehension of fundamental characteristics such as structural, thermal, optical and electrical properties. The purpose of the present work was to study the effects of the glass-forming history on structural and thermal properties on 20Li2O-80TeO2 glasses by using the XRD and DSC techniques as investigative tools. Glasses studied in the present work were prepared with nominal compositions 20Li2O-80TeO2 but with different glass-forming histories. Therefore, stress free and stressed glasses were obtained. Powdered glasses with particle size 63-75 μm, 45-63 μm, 38-45 μm and < 38 μm were used for studies reported in this work. The structure of the crystallized glasses was analyzed by using XRD. DSC measurements were performed to study the thermal properties of the studied glasses. Activation energies were evaluated considering that observed DSC crystallization peak is a superposition of three Gaussian functions, associated to three distinct crystallization phases in the glass devitrification. This assumption was supported by XRD results. The XRD results confirm the amorphous state for all as-quenched glasses used in the present investigation and that different thermal histories were not sufficient to produce any crystalline phase in these glasses. XRD results show the γ-TeO2, α-TeO2 and Li2Te2O5 phase crystallization in the glass matrix during devitrification, where the γ-TeO2 phase is considered a metastable structure. The structure of the α-TeO2 are formed by three-dimensional network TeO4 units sharing oxygen corners by symmetric Te-O-Te bridges while the γ-TeO2 structure can be considered as a chain system, where TeO4 units are alternatively linked by nearby and high symmetric Te-O-Te bridges. Based only on XRD it was not possible to determine whether α-TeO2, γ-TeO2 and Li2Te2O5 phases crystallize simultaneously or at distinct onset crystallization temperatures. From DSC results, the obtained glass transition temperature (Tg) values were between 263267oC and the onset crystallisation temperature between 328-347oC. The DSC heating curves of the stressfree glasses shows endothermic peaks close to Tg while DSC curves of the stressed glasses demonstrate a typical glass transition steps. This result provide a clear evidence that the observed endothermic peak close to Tg have no direct correlation with residual stress in the tellurite glasses studied in the present work. In accordance with XRD results, the asymmetry on DSC crystallization peaks suggests a multiphase crystallization in the glass. Under these considerations, each DSC crystallization peak was treated as a convolution of three individual peaks with distinct maxima temperature (Tp) to determine the activation energies. Results shows that 1 2 3 E E E ≈ > (327 > 298 ≈ 296 kJ.mol ) for glass with particle size 63-75 μm, while 3 2 1 E E E ≈ < (246 < 289 ≈ 293 kJ.mol ) is for glass with particle size smaller than 38 μm. Therefore, the energies E1, E2 and E3 were associated to the following γ-TeO2, α-TeO2 and Li2Te2O5 crystallization in the 20Li2O-80TeO2 glass, respectively.

In this work, X-ray diffraction, Raman spectroscopy and differential scanning calorimetry techniques were used to understand the crystallization process on 20Li2O–80TeO2 glass. X-ray diffraction results reveal the presence of three distinct alpha γ-TeO2, α-TeO2 and α-Li2Te2O5 crystalline phases in the glass matrix. The Raman spectroscopy band structure of this glass is similar to the one observed in glassy TeO2. Raman results clearly reveal the metastable character of the γ-TeO2 phase in the 20Li2O–80TeO2 glass, whose associated vibration modes disappear completely at temperatures higher than 315°C. On the other hand, the Raman modes associated to α-TeO2 and α-Li2Te2O5 phases persists up to temperatures close to the final stages of the crystallization in the studied glass (around 420°C). From DSC measurements, the activation energies 296±3 and 298±1kJmol−1 were associated to γ-TeO2 and α-TeO2 phases crystallization, indicating that these phases crystallizes at temperatures very close in the studied glass.

Crystallization kinetics and structure of 20Li(2)O-80TeO(2) glasses are studied using x-ray diffraction, Fourier transform infrared spectroscopy and differential scanning calorimetry techniques. XRD results show gamma-TeO2, alpha-TeO2 and Li2Te2O5 phase crystallization in the glass matrix. The infrared band structure of this glass is similar to that observed in glassy TeO2. Activation energies were evaluated from Lorentzian three-peak deconvolution of the DSC crystallization peak recorded at different particle sizes. As the obtained activation values were very close, it was not possible to establish a hierarchy on the crystallization of this glass. In addition, the height of the Lorentzian peaks ( delta T-P1, delta T-P2 and delta T-P3) for each sample was measured and plotted against the nucleation temperature. The presence of two maxima at around 284 and 304 degrees C suggests at least two maximum nucleation temperatures in the studied 20Li(2)O-80TeO(2) glass.

O presente trabalho reporta estudos sobre a cristalização em vidros teluretos 20Li2O-80TeO2 induzida a partir de tratamentos térmicos realizados sobre vidros com tamanho de partículas entre 38 µm e 75 µm. Estes estudos foram conduzidos em duas matrizes vítreas tratadas e não tratadas termicamente para aliviar as tensões após o quenching, utilizando-se de forma combinada às técnicas de difração de raios X, calorimetria diferencial de varredura e espectroscopia no infravermelho. Os resultados revelaram a presença de três fases cristalinas distintas durante o processo de cristalização e apontaram para uma hierarquia tal que as fases alfa-TeO2 e gama-TeO2 cristalizam-se antecipadamente à fase Li2Te2O5 no vidro 20Li2O-80TeO2 sujeito a tensões mecânicas induzidas durante a síntese. No vidro 20Li2O-80TeO2, livre de tensões mecânicas, não foi possível discriminar esta hierarquia de cristalização.This work report crystallization studies on tellurite 20Li2O-80TeO2 glasses induced from heat thermal annealing on glasses with particle size between 38 µm and 75 µm. These studies were conducted on two glass matrix heat thermal annealed and non-annealed to remove the stress after the quenching, by using the X-Ray diffraction analysis, Fourier transform infrared spectroscopy and differential scanning calorimetry techniques. The results shown the presence of three distinct crystalline alpha-TeO2, gamma-TeO2 and Li2Te2O5 phases during the crystallization process, suggesting a crystallization hierarchy on the glass matrix under stress, since the gamma-TeO2 and alpha-TeO2 phases crystallization occurs before the Li2Te2O5 phase. On the glass stress free, this crystallization hierarchy was not clearly determined.

O presente trabalho reporta estudos sobre a cristalização em vidros teluretos 20Li2O-80TeO2 induzida a partir de tratamentos térmicos realizados sobre vidros com tamanho de partículas entre 38 µm e 75 µm. Estes estudos foram conduzidos em duas matrizes vítreas tratadas e não tratadas termicamente para aliviar as tensões após o quenching, utilizando-se de forma combinada às técnicas de difração de raios X, calorimetria diferencial de varredura e espectroscopia no infravermelho. Os resultados revelaram a presença de três fases cristalinas distintas durante o processo de cristalização e apontaram para uma hierarquia tal que as fases alfa-TeO2 e gama-TeO2 cristalizam-se antecipadamente à fase Li2Te2O5 no vidro 20Li2O-80TeO2 sujeito a tensões mecânicas induzidas durante a síntese. No vidro 20Li2O-80TeO2, livre de tensões mecânicas, não foi possível discriminar esta hierarquia de cristalização.

20Li2O–80TeO2 glasses were heat annealed at different temperatures between Tg and Tx and studied by using XRD, FTIR spectroscopy and DSC techniques to understand the crystallization kinetics in this glass matrix. The infrared band structure of this glass is similar to what was observed in glassy TeO2. XRD results reveal the presence of three distinct crystalline γ-TeO2, α-TeO2 and Li2Te2O5 phases during the crystallization process. This is a first report of γ-TeO2 phase crystallization in this glass matrix. DSC results confirm the crystallization of three distinct structures in the glass. In summary, our results suggest a crystallization hierarchy on this glass matrix since the γ-TeO2 and α-TeO2 phases crystallization occurs before the Li2Te2O5 phase crystallization.