Composition-dependent evolution toward multi-phase crystallization in lanthanum tellurite glasses

Tellurite-based glass systems have attracted considerable attention owing to their promising thermal and optical characteristics for advanced glass and glass-ceramic applications. The present study investigates the influence of La 2 O 3 addition on the thermal behavior, crystallization kinetics, and phase evolution of TeO 2 -based binary glasses. TeO 2 -La 2 O 3 samples containing 5–30 mol% La 2 O 3 were synthesized via melt quenching and characterized using physical, structural, and non-isothermal DSC analyses. XRD results confirmed that the predominantly amorphous character was retained up to 20 mol% La 2 O 3 , whereas 30 mol% La 2 O 3 content promoted crystallization and the formation of lanthanum-rich crystalline phases. Increasing La 2 O 3 content increased glass density from 5.22 to 5.51 g.cm -3 and T g from 640 to 741 K, while simultaneously decreasing the glass stability parameter ΔT from 37 to 25 K. Non-isothermal kinetic analyses suggested a composition-dependent evolution in crystallization behavior. The low-La composition (95T5L) exhibited relatively sequential crystallization characteristics, whereas the high-La formulation (80T20L) showed evidence of a more complex phase assemblage after heat treatment at the first crystallization-peak temperature. Comparison of the representative low- and high-La compositions demonstrated a substantial increase in E p1 from approximately 192–194 to 348–353 kJ.mol -1 , accompanied by a decrease in E p2 from approximately 261–274 to 220–235 kJ.mol -1 with increasing 5 and 20 mol% La 2 O 3 content, respectively. This suggested the redistribution of the apparent activation energies associated with the crystallization events. Post-heat-treatment XRD analyses were consistent with the kinetic findings by revealing the coexistence of multiple La-rich crystalline phases in the heat-treated high-La composition. In conclusion, the present results suggest that increasing La 2 O 3 content is associated with a composition-dependent evolution from relatively sequential crystallization toward increasingly complex crystallization behavior. These findings provide useful insight into crystallization control and composition-dependent phase evolution in tellurite-based glass and glass-ceramic systems and establish a scientific basis for future studies of their thermal processing and potential optical and photonic applications.

The developments in glass science and technology have steadily progressed in parallel with advancements in a wide range of applications. Starting with traditional glass types, such as silicates, and extending to advanced compositions, such as tellurites, glass materials have been extensively utilized to meet increasing technological demands (Shelby, 2005a; Mauro, 2014). The high flexibility of compositional design in glass systems has attracted considerable scientific interest, leading to continuous developments in glass formulations (Zhao et al., 2026). In oxide glass systems, structural networks are generally formed through interconnected polyhedral units linked via bridging oxygen (BO) atoms. Accordingly, compositional modifications directly influence critical properties, including thermal stability, optical transmission, and chemical durability (Mauro, 2018; Mauro et al., 2014). Therefore, understanding the role of different oxides within glass networks remains essential for controlling glass properties in advanced technological applications, such as optics and photonics.

Among alternative glass systems, tellurium oxide (TeO 2)-based glasses (TG) have gained considerable attention over the last few decades owing to their superior optical and thermal properties (El-mall and awany, 2018; D’Souza et al., 2023; Uosif et al., 2023). Compared with conventional silicate or borate glasses, TG systems exhibit enhanced rare-earth solubility, high refractive index, wide infrared transparency, low melting temperatures, and considerable dielectric constants, making them promising candidates for advanced optical, optoelectronic, and glass-ceramic applications (Yakhkind, 1966; El-Mallawany, 1998; El-Mallawany et al., 2008). These characteristics have positioned tellurite-based materials among the promising candidates for photonic technologies, optical amplifiers, and transparent glass-ceramic systems. Structurally, tellurite glasses are mainly composed of TeO 4 trigonal bipyramidal and TeO 3 trigonal pyramidal units (Verger et al., 2024; Sekiya et al., 1995; Umar et al., 2023). The relative distribution of these units strongly influences the structural and thermal behavior of the glass network (Rada et al., 2009). In particular, the transformation from TeO 4 to TeO 3 units is generally associated with the formation of non-bridging oxygen (NBO) atoms, which significantly affect structural connectivity, thermal stability, and crystallization tendency within tellurite glass systems (Gupta et al., 2017). As a result, TG systems can potentially be designed to achieve the desired glass properties for related application areas.

While TG systems have many advantages, they, on the other hand, exhibits limited glass-forming ability because of its intermediate glass-former character (Musgraves et al., 2019; Kavaklioğlu et al., 2015). That means, TG systems generally require the incorporation of additional oxides to stabilize the glass structure and maintain structural connectivity between Te-O units (Joshi et al., 2013). Depending on the field strength, ionic radius, and coordination characteristics of the incorporated cations, such compositional modifications may either improve or deteriorate the thermal stability and crystallization behavior of TG systems (Inaba et al., 1999; Irannejad et al., 2013). From the viewpoint of optical and photonic glass processing, crystallization behavior becomes particularly critical because uncontrolled devitrification may induce optical scattering, structural heterogeneity, and reduced transparency in thermally processed glass and glass-ceramic systems. Thus, understanding the structural role of dopant oxides remains a critical factor for controlling the overall stability and phase evolution behavior of tellurite glasses.

When reviewing the prior literature on TG systems, one can trace their development back more than a century to the earliest investigations. Progress in raw material purification and glass melting technology has led to significant developments in TG systems, especially from the late 1980s through the 1990s, when their nonlinear optical and infrared transmission properties began to be comprehensively investigated (Mochida et al., 1978). Following extensive glass formulation studies, namely binary, ternary, and quaternary TG compositions, valuable insights have been obtained into the structural, thermal, mechanical, and optical properties. One of the pioneering researchers, R. A. El-Mallawany, has contributed significantly to existing knowledge by addressing critical aspects, including structural units and vibrational modes, elastic behavior, and rare-earth-related optical characteristics (El-Mallawany, 1995; El-Mallawany and Ahmed, 2008). In addition, Sae-Hoon Kim contributed to understanding the linear and nonlinear optical properties of TG systems by analyzing the effects of transition-metal oxides and exploring electronic polarizability and optical susceptibility (Kim, 1999). Beyond these contributions, other researchers have sought to understand the potential of TG systems more broadly so that advanced optical devices, including fiber amplifiers, nonlinear components, and infrared photonic elements, can be developed (Himei et al., 1994; Ersundu et al., 2010; Liang et al., 2024). While many efforts have been devoted to inserting transition metal oxides (e.g., Nb 2 O 5 or Ta 2 O 5) (Fujimoto et al., 2001; Kilic et al., 2020), rare-earth oxides (e.g., Sm 2 O 3 or CeO 2) (El-Deen et al., 2008), or even light metal oxides (e.g., Li 2 O or Na 2 O) (Çelikbilek Ersundu and Ersundu, 2016) into tellurite glass networks, significant research gaps still exist regarding the role of lanthanum oxide (La 2 O 3) in simple binary tellurite glass systems. La 3+ ions, characterized by relatively large ionic radius and high field strength, are known to influence oxygen packing density, structural connectivity, and thermal behavior within oxide glass networks (Saddeek and Elsayed, 2015; Patra et al., 2022). However, the composition-dependent crystallization behavior and phase evolution of binary TeO 2 -La 2 O 3 glasses have not yet been systematically investigated, particularly from the perspective of non-isothermal crystallization kinetics.

Crystallization behavior is one of the most critical factors governing thermal processing conditions, amorphous stability, and the controllable production of tellurite-based glass-ceramic materials (Lahl et al., 2000). In particular, understanding crystallization kinetics is essential for predicting phase evolution pathways, optimizing thermal treatment schedules, and controlling undesired devitrification during high-temperature processing. Such knowledge is particularly important in tellurite-based optical glasses and glass-ceramic systems, where crystallization characteristics directly influence structural stability and processing behavior. Moreover, controlled crystallization is widely recognized as a key parameter in the fabrication and thermal processing of transparent tellurite glass-ceramics intended for optical and photonic applications. Therefore, systematic kinetic investigations are required to establish composition-crystallization relationships in binary tellurite glass systems. In addition, La 2 O 3 has attracted considerable attention owing to its moderate cost, relative accessibility, and strong field-strength characteristics among rare-earth oxides (Schulz et al., 2017; Rare Earth Elements, 2024; Öztürk et al., 2019). The incorporation of La 3+ ions into tellurite networks may substantially alter thermal stability, crystallization tendency, and phase formation behavior depending on composition. Therefore, understanding how La 2 O 3 content influences crystallization pathways and thermal response remain important for the design and development of tellurite-based glass and glass-ceramic materials.

With this in mind, the present study aims to systematically investigate the influence of La 2 O 3 addition on the thermal behavior, crystallization kinetics, and phase evolution of TeO 2 -based binary glasses. To the best of our knowledge, systematic crystallization kinetic analyses addressing binary lanthanum tellurite glass systems remain very limited in the existing literature. Particular attention is given to the evaluation of characteristic temperatures and the interpretation of crystallization mechanisms using established non-isothermal kinetic approaches, including Kissinger, Ozawa, Lasocka, and apparent Avrami analyses. The obtained findings provide useful insight into crystallization control and composition-dependent phase evolution in tellurite glass systems relevant to thermally processed optical and photonic glass-ceramic materials.

Glass synthesis

Tellurium oxide (TeO 2) was purchased from Thermo Scientific with a purity of 99.5%, while Lanthanum oxide (La 2 O 3) was supplied from Solarbio with a purity of 99%. No further processing was applied to either reagent, and they were used as received. Following that, binary glass systems, TeO 2 -La 2 O 3, were designed with respect to the formula given: (100-x) TeO 2 + x La 2 O 3, where x equals 5, 10, 15, 20, and 30 mol%. Both oxides were weighed in small plastic cups using an analytical scale (Precisa, XB220A, ±0.0001) in accordance with the glass design, yielding 5 g per batch. Then, each batch was transferred into an agate mortar and was mixed for 5 min to ensure homogeneity. Afterward, the mixed batch was gently poured into an alumina crucible and was placed in the melting furnace (CLS Scientific, 5 L, 1400 °C max.). Each batch was melted at 1000–1300 K with a heating rate of 10 K.min -1, and held for 15–30 min at the corresponding peak temperature. Once melting was complete, the alumina crucible was held by a pair of metal tongs, removed from the furnace, and water-quenched immediately. Subsequently, water was vacuum-filtered (CLS Scientific, 7 bar, 23L/min) using a 0.45 µm Whatman paper to obtain the glass frits. In the end, glass frits were dried in the oven (WiseVen) at 400 K for 4 h, and finally, glass synthesis was completed. The mentioned steps were applied for each batch separately, and glass frits were successfully obtained. The glass series was designated as 95T5L, 90T10L, 85T15L, 80T20L, and 70T30L, where T and L denote TeO 2 and La 2 O 3, respectively, while numerical values correspond to the insertion ratios in mol%.

Characterizations

After glass synthesis, each series was ground in an agate mortar until the particle size was less than 106 microns. These powders were then used for characterization analyses.

Density (ρ glass) measurements were carried out using the liquid pycnometer technique with a 25 mL pycnometer bottle (V exact: 25.35 mL) and its corresponding cap. Initially, the density of deionized water (ρ water) was determined by weighing the empty pycnometer bottle with its cap, and followingly filling it with deionized water up to the calibration mark. The ρ water was calculated to be 0.9966 g.cm -3 at room temperature (∼20 °C). Following that, each glass sample was measured using at least 3 g of powder. All measurements were performed in triplicate to ensure reproducibility. The experimental uncertainty associated with the ρ glass measurement was found to be within ±3%.

Using the experimentally measured ρ glass values, several structural parameters of the glass network, including molar volume (V m), oxygen packing density (OPD), and average crosslink density (n c), were evaluated.

The molar volume (V m) was calculated to obtain information about the distribution of the structural units within the glass network and the overall compactness of the glass structure. V m of each glass composition was calculated according to Equation 1 (Shelby, 2005b). where AMW represents the average molecular weight of the glass composition.

The oxygen packing density (OPD) was determined to evaluate the packing efficiency of oxygen atoms in the glass matrix and to provide insight into the compactness of the glass network. The oxygen packing density of the glass compositions was calculated using Equation 2. where y denotes the number of oxygen atoms available in the glass composition.

The average crosslink density (n c ¯) was estimated to assess the degree of connectivity and crosslinking within the glass structure. The average crosslink density of the glass network was calculated using Equation 3. where n c ¯ is the average number of crosslinks per cation in the glass composition, x i is the molar fraction of the i th oxide component, (n c) i is the number of crosslinks per cation assigned to the i th oxide, (N c) i is the number of cations in the formula unit of the i th oxide.

X-ray diffraction (XRD) was used to observe the structural changes as a function of La 2 O 3 insertion. A Bruker D8 Advance device was used with a Cu-K α source operating at 40 kV and 30 mA, and the diffraction data were recorded between 2θ = 3°–80° with a 0.02° step size. The crystalline phases were identified qualitatively by matching the experimental diffraction patterns with the corresponding ICDD reference patterns. No Rietveld refinement or other quantitative phase analysis was performed.

Fourier transform infrared (FTIR) spectroscopy was conducted via a Perkin Elmer UATR Two device over 4000 to 400 1/cm wavenumber range to highlight the structural units and vibrational modes.

Differential scanning calorimetry (DSC) was performed using a Netzsch STA 449 Jupiter at heating rates of 10, 20, 40, and 50 K.min -1 to determine the thermal characteristics of TL glass systems. The characteristic temperatures, namely the glass transition temperature (T g), crystallization onset temperature (T x), crystallization peak temperature (T p), melting temperature (T m), and liquidus temperature (T l), were determined from the obtained thermograms.

In addition, glass formation ability (GFA) was evaluated using the glass stability parameter (ΔT), Hruby parameter (K H), and reduced glass transition (T rg), as given in Equations 4–6, respectively (Weinberg, 1994; H and rubý, 1972; Turnbull, 1969).

Furthermore, the crystallization behavior of the synthesized glass series at different heating rates was analyzed using the Kissinger and Ozawa methods, as expressed in Equations 7, 8, respectively (Kissinger, 1957; Ozawa, 1965). Also, Lasocka’s relation was employed to examine the heating-rate dependence of characteristic temperatures, as presented in Equation 9 (Lasocka, 1976). where β is the heating rate, R is the universal gas constant, E p is the apparent crystallization activation energy, T p is the crystallization peak temperature, C is the intercept constant in the Ozawa relationship, T represents the characteristic temperature (e.g., T g, T x, or T p) used in the Lasocka relationship, and A and B are the Lasocka constants corresponding to the intercept and slope of the linear relationship between the characteristic temperature and ln(β), respectively.

In the present study, the activation energies calculated using the Kissinger and Ozawa methods are interpreted as apparent activation energies associated with the overall crystallization event represented by each DSC peak. These values, therefore, reflect the combined effects of nucleation, crystal growth, atomic diffusion, and other kinetic contributions under non-isothermal conditions rather than elementary atomic reaction barriers.

The crystallization behavior was further examined using the logarithmic form of the Johnson-Mehl-Avrami-Kolmogorov (JMAK) expression. Because temperature varied continuously during the DSC measurements, this treatment does not represent a conventional isothermal Avrami analysis. The resulting slope is therefore reported only as an apparent Avrami-type exponent (n app) used for comparative evaluation of the crystallization responses at different heating rates in Equation 10 (Avrami, 1941; Matusita and Sakka, 1980). where α, K, and t represent crystallization fraction, crystallization rate constant, and time, respectively.

Because the present study employed continuous-heating DSC experiments, the calculated Avrami exponent (n app) should be regarded as an apparent non-isothermal kinetic parameter rather than an intrinsic mechanistic constant. The crystallized fraction (α) was determined from the ratio of the partial exothermic peak area to the total integrated peak area after baseline correction between the onset and end temperatures of each crystallization event using the DSC analysis software (NETZSCH Proteus Thermal Analysis). The first two crystallization peaks were sufficiently resolved to permit independent kinetic analysis. Therefore, no peak deconvolution procedure was required.

Glass formation and structural characteristics

Figure 1 presents the X-ray diffraction patterns of TeO 2 -La 2 O 3 glass series. When the diffraction data of the samples having comparatively lower La 2 O 3 amounts, which are 95T5L, 90T10L, and 85T15L, are examined, a characteristic broad hump located between 2θ = 25°–35° appears without any sharp peaks. This indicates that the amorphous structure is preserved up to 15 mol% La 2 O 3 addition. When higher amounts of La 2 O 3 are incorporated into the TeO 2 matrix, for example 80T20L sample, slight changes are observed, revealing that the structure largely preserves its non-crystallinity, but very weak crystalline peaks with low intensities begin to form. When the highest La 2 O 3 content, 70T30L, is introduced into the glass network, sharp crystalline peaks become more prominent. These peaks are identified as a: Te 4 O 9 (PDF 01-071-0355), b: La 2 Te 6 O 15 (PDF 00-037-1394), c: La 2.33 Te 1.67 O 6.85 (PDF 00-041-0260), and d: La 2 (TeO 3) 3 (PDF 00-051-0279) crystalline phases. The formation of such crystal phases at the highest La 2 O 3 insertion rate indicates that partial crystallization occurs within the glass matrix. This behavior suggests that increasing La 2 O 3 content promotes the transformation of the TeO 2 network toward energetically more stable crystalline lanthanum tellurate phases. Similar phase evolution has been reported in La 2 O 3 -containing tellurite glass systems, where the formation of lanthanum-rich phases becomes favorable at higher La 2 O 3 concentrations (Patra et al., 2022). Therefore, the TL glass system preserves its amorphous structure up to 20 mol% La 2 O 3; however, a further increase to 30 mol% clearly leads to the formation of significant crystalline phases.

Table 1 shows the measured glass density (ρ glass) values and several structural parameters of the glass network, including molar volume (V m), oxygen packing density (OPD), and average crosslink density (n c). According to the results, ρ glass increases from 5.22 to 5.51 g.cm -3 as the La 2 O 3 content increases. This increase is primarily attributed to the replacement of TeO 2 by the considerably heavier La 2 O 3 component rather than to a more compact glass network. This increase is consistently observed for each compositional step in the sample series. In addition, increasing La 2 O 3 doping rate causes V m to increase from 32.17 to 38.02 cm 3.mol -1. That means, the increase in bulk density occurs simultaneously with an expansion of the molar volume, indicating that the two parameters reflect different compositional and structural effects. A closer examination of OPD reveals that it decreases from 63.73 to 60.50 mol.L -1 with the increasing La 2 O 3 content, indicating that the compositional modification in the TL glass series directly influences the packing density of oxygen atoms within the glass network. A similar tendency also exists for the n c ¯ values. That is, 95T5L sample has a n c ¯ value of 19.05, whereas increasing La 2 O 3 amount through 70T30L sample decreases this parameter to 15.39. Considering all structural parameters listed in Table 1, the gradual substitution of TeO 2 by La 2 O 3 in the glass network leads to noticeable variations in the structural characteristics of the TL glass system.

Fourier transform infrared spectra of TeO 2 -La 2 O 3 glass series from 1600 to 400 1/cm wavenumbers is shown in Figure 2. All samples exhibit a sharp band between 600 and 800 1/cm, indicating a strong relation with Te-O vibrations modes, which is commonly observed in similar glass systems (Gangwar et al., 2019). When the spectra are reviewed in detail, a strong band identified around 650 to 700 1/cm may be associated with Te-O stretching vibrations contained by TeO 4 trigonal bipyramids (Tagiara et al., 2017). Besides that, the bands observed around 750 to 800 1/cm may be linked to vibrational modes TeO 3 trigonal pyramids (Kalampounias, 2012). The spectra of the samples largely preserved their shapes as a function of La 2 O 3 insertion levels; however, there are some alterations in the band intensities and areas. Particularly, the bands around 600 to 700 1/cm relating to the samples with higher La 2 O 3 contents reveals slight intensity decreases, whereas the bands around 700 to 800 1/cm becomes more prominent. Considering these results, it can be stated that the fundamental tellurite network vibrations are largely preserved in different compositions; yet, La 2 O 3 addition affected peak intensities and areas to a certain extent.

When structural parameters of the TeO 2 -La 2 O 3 glass system are evaluated together with XRD patterns and FTIR spectra, it can be stated that La 2 O 3 addition significantly affects the tellurite glass network. The observed gradual increase in ρ glass with increasing La 2 O 3 doping ratio can be associated with the higher molar mass of La 2 O 3 (325.81 g.mol -1) compared with TeO 2 (159.6 g.mol -1). Similar density enhancement following La 2 O 3 incorporation in tellurite glass systems as well as in the other glass compositions has been reported in the literature (Lakshminarayana et al., 2020; Karthikeyan and Manikandan, 2024). Although La 2 O 3 addition increases the measured bulk density, the simultaneous increase in V m indicates an expansion of the average molar volume rather than a more compact structural packing. In other words, La 2 O 3 addition may induce structural modifications by reorganizing the average interatomic distances within the glass network. Some studies have also reported a parallel variation of ρ glass and V m in different glass formulations (Al-Buriahi et al., 2020). FTIR results provide additional evidence supporting the structural parameters. Specifically, increasing La 2 O 3 content modifies the bands observed around 650 to 700 1/cm and 750 to 800 1/cm, where the increasing intensity of the latter band corresponds to the formation of TeO 3 units at the expense of TeO 4 units. This tendency may be associated with structural rearrangements within the tellurite glass network accompanying La 2 O 3 addition. When all results are considered together, it can be concluded that La 2 O 3 addition into TeO 2 -based glass systems significantly affects the structural characteristics and thermal behavior of the glass system.

Thermal behavior and glass stability

Figure 3 presents DSC thermograms of TeO 2 -La 2 O 3 glass systems recorded at a heating rate of 20 K.min -1 while, Table 2 lists the characteristic temperature values related to glass transition temperature (T g), onset of crystallization (T x), crystallization peak temperatures (T p), melting temperature (T m), and liquidus temperature (T l), together with the calculated glass-forming ability parameters derived from thermal data, namely glass stability (ΔT), Hruby (K H), and reduced glass transition (T rg). When the results are examined in detail, T g values gradually increase as a function of La 2 O 3 increase. The value of T g for 95T5L sample is 640 K, whereas La 2 O 3 insertion ratio raises it to 659 K and 694 K for 90T10L and 85T15L samples, respectively, and finally the highest T g, 741 K, is obtained for 80T20L sample. The increase in T g indicates that higher thermal energy is required to initiate viscoelastic behavior, reflecting reduced structural mobility within the glass network. This behavior can be attributed to enhanced local structural rigidity arising from strong interactions between La 3+ ions and oxygen atoms. Moreover, when the crystallization behavior of the glass series is considered, exothermic crystallization peaks (Tp 1-3) following T g can be observed. Two exothermic peaks are seen for lower La 2 O 3 contents, whereas three crystallization peaks emerge for La 2 O 3 concentrations greater than 15 mol%. This behavior suggests that the crystallization mechanism becomes more complex, potentially involving the formation of multiple crystalline phases. On the other hand, the glass-forming ability of these glass systems can be assessed through commonly used thermal parameters. The widely used parameter ΔT varies according to the compositional changes. The highest difference, 37 K, is obtained for 95T5L sample while the lowest value, 25 K, is held with 80T20L sample. The decrease in glass stability parameters indicates an increasing susceptibility of the glass network to crystallization with increasing La 2 O 3. Likewise, K H equals to 0.23 for 95T5L sample, where it decreases to 0.07 for 80T20L sample. This trend clearly shows that the crystallization tendency increases significantly in line with the higher amounts of La 2 O 3. These results highlight a critical point: a higher T g does not necessarily correspond to improved glass stability. In the present system, the addition of La 2 O 3 increases the thermal rigidity of the glass network, while simultaneously enhancing its tendency toward crystallization. Therefore, the glass can operate at higher temperatures but exhibit a narrower processing window.

When DSC thermograms are evaluated together with XRD patterns, a consistent relationship emerges. Specifically, the increased crystallization tendency observed for samples with higher La 2 O 3 doping ratios, such as 80T20L, is consistent with the phase formations identified in the XRD analyses. This observation reasonably dictates that the amorphous stability of the glass network decreases while the crystallization tendency increases when La 2 O 3 content exceeds approximately 20 mol%. In fact, 70T30L sample is found to be largely crystalline and therefore was not evaluated by DSC measurements. In this context, detailed crystallization kinetics were investigated for the samples exhibiting the lowest and highest glass stability values, which are 95T5L (ΔT = 37 K) and 80T20L (ΔT = 25 K). Studying these two compositions enables a comparative evaluation of glass formulations representing the upper and lower limits of glass stability with respect to crystallization behavior. The intermediate compositions (90T10L and 85T15L) also exhibited progressive changes in their DSC thermograms, particularly through the emergence of an additional crystallization event with increasing La 2 O 3 content. Although these observations qualitatively indicate increasing crystallization complexity, detailed kinetic analyses were intentionally limited to the representative compositions exhibiting the lowest and highest glass stability (95T5L and 80T20L), enabling a direct comparison of crystallization behavior at the compositional extremes.

Non-isothermal crystallization kinetics

To further evaluate the crystallization behavior of TeO 2 -La 2 O 3 glass system, DSC measurements were performed on 95T5L and 80T20L samples at different heating rates of 10, 20, 40, and 50 K.min -1. For consistency in the comparative kinetic evaluation, only the first two crystallization peaks (T p1 and T p2) were considered in the Kissinger, Ozawa, and apparent Avrami analyses. Although a third crystallization peak (T p3) appeared in the higher La 2 O 3 compositions, it was not included because this crystallization event was absent in the 95T5L composition. Restricting the kinetic analyses to the crystallization events common to both representative glasses enabled a direct comparison of their crystallization behavior under identical analytical conditions, whereas the occurrence of T p3 was considered qualitative evidence of increasing crystallization complexity with increasing La 2 O 3 content.

The resulting thermograms are presented in Figure 4, and the characteristic temperatures derived from these analyses are summarized in Table 3. For both compositions, the characteristic temperatures systematically shift toward higher values with increasing heating rate. This behavior is a typical feature of kinetically controlled transformations in glass systems under non-isothermal conditions (Cholin et al., 2022). In other words, as the heating rate increases, the time available for structural relaxation, nucleation, and crystal growth decreases. Thus, higher temperatures are required for the same transformation phenomena to occur. Similar trends have also been reported for tellurite-based glass systems in the literature (Çelikbilek et al., 2011). This observation is not merely a general DSC feature; it also provides direct insight into the atomic-scale rearrangement capability of the glass structure. Phenomena such as glass transition and crystallization are governed by the ability of the structure to reorganize under thermal stimulation. If the structure is more rigid, less mobile, or associated with higher energy barriers, these transformations occur at higher temperatures. From this perspective, the relatively higher T g and T p temperatures observed for 80T20L sample indicate that increasing La 2 O 3 content restricts structural mobility within the glass network. This interpretation is consistent with the increase in local structural rigidity discussed in the previous section. However, it should be emphasized that this shift toward higher temperatures does not necessarily imply improved glass stability, since higher crystallization temperatures do not always correspond to a lower tendency for crystallization.

A detailed examination of Table 3 reveals that increasing the heating rate from 10 to 50 K.min -1 for 95T5L sample leads to a systematic increase in T g (from 635.39 to 647.09 K), T x (from 663.95 to 697.04 K, T p1 (from 698.98 to 735.46 K), and T p2 (from 765.47 to 795.53 K). A similar trend is observed for 80T20L sample, where T g, T x, T p1, and T p2 increase from 736.79 to 747.89 K, from 757.34 to 777.93 K, from 781.34 to 805.20 K, and 845.56–888.75 K, respectively. This systematic shift confirms that the thermal phenomena in the glass system are highly sensitive to heating rate, reflecting the kinetic nature of the transformations and the reliability of the dataset for kinetic analysis. Furthermore, the consistently higher characteristic temperatures observed for 80T20L composition compared to 95T5L indicate that increasing La 2 O 3 content shifts structural rearrangement processes to higher temperatures.

To quantitatively evaluate the effect of heating rate, the variation of characteristic temperatures as a function of ln(β) was analyzed using the Lasocka relationships, and the results are plotted in Figure 5. For both compositions, the characteristic temperatures, namely T g, T x, T p1, and T p2, exhibit approximately linear relationships with ln(β), indicating that the glass system follows a consistent kinetic behavior under non-isothermal conditions. This linear dependence confirms that the DSC data are not randomly distributed but evolve according to well-defined kinetic relationships. In particular, the high correlation coefficients obtained for crystallization-related temperatures support the reliability of the dataset for activation energy calculations. It should be noted that the linear correlation obtained for T g is relatively weaker compared to those of crystallization-related temperatures. This behavior is somewhat expected since the glass transition is not a discrete phenomenon like crystallization but rather a distributed progression associated with structural relaxation over a broad temperature range. Consequently, the strong linear relationships observed for T x, T p1, and T p2 in Figure 5 provide a strong advantage for reliable kinetic analysis.

The activation energies (E p) associated with the crystallization peaks of 95T5L and 80T20L samples were determined using the Kissinger and Ozawa methods, and the corresponding results are summarized in Table 4. The close agreement between E p values obtained from both methods confirms the internal consistency and reliability of the calculations. For 95T5L sample, E p1 corresponding to the first crystallization peak ranges between 192 and 194 kJ.mol -1, while that of E p2 lies between 261 and 274 kJ.mol -1. This indicates that the second crystallization event is characterized by a higher apparent activation energy than the first, consistent with a greater apparent kinetic resistance for the overall process represented by the second crystallization peak. In contrast, a sharply different trend is observed for 80T20L sample. The E p1 is significantly higher, ranging between 348 and 353 kJ.mol -1, while E p2 exhibits lower values in the range of 220–235 kJ.mol -1. This contrast in the activation energies suggests that, in La-rich compositions, the first crystallization peak exhibits the highest apparent activation energy. In other words, the onset of crystallization in 80T20L sample is consistent with a greater overall kinetic resistance associated with the initial crystallization event. The lower apparent activation energy obtained for the second crystallization peak suggests that the overall crystallization process represented by this peak proceeds with a lower apparent kinetic resistance. This behavior suggests that crystallization in the high-La 2 O 3 composition is not a simple structural ordering within the tellurite network, but likely involves the formation of chemically distinct La-rich phases. Such an interpretation is consistent with the XRD patterns obtained after heat treatment, where lanthanum tellurate phases are prominently observed in 80T20L samples (refer to Figure 6). Overall, the activation energy results indicate that the apparent crystallization kinetics are strongly influenced by La 2 O 3 content. The representative low- and high-La compositions exhibit markedly different distributions of the apparent activation energies associated with the first and second crystallization events. Rather than representing elementary reaction barriers, these values reflect differences in the overall kinetic responses of the corresponding crystallization events under non-isothermal conditions.

The crystallization mechanism was further evaluated not only in terms of activation energy barriers but also based on the transformation pathway using crystallization fraction (α) curves, as presented in Figure 7. The corresponding apparent Avrami exponent (n app) values and correlation coefficients are summarized in Table 5. The obtained n app, ranging approximately between 2 and 4, suggest that the crystallization process cannot be adequately described by a single invariant transformation mode under the present non-isothermal conditions. Within the framework of non-isothermal crystallization analysis, this range is commonly interpreted as being consistent with mixed nucleation and crystal-growth contributions, although the obtained n app values should not be interpreted as unique mechanistic descriptors. Accordingly, the n app values are consistent with contributions from both nucleation and crystal growth, but they do not permit an unambiguous distinction between surface- and volume-controlled crystallization under the present non-isothermal conditions. For 95T5L sample, the n app values corresponding to the T p1 vary between 2.44 and 3.49, indicating that the initial crystallization step is more sensitive to heating rate and exhibits a relatively variable mechanistic character. In contrast, the n app values for the T p2 lie within a narrower range, 2.21 to 2.57, suggesting a more consistent growth behavior during the later stages of crystallization. This observation becomes particularly meaningful when considered together with the higher activation energy associated with E p2. In this case, the second crystallization event is characterized by a higher apparent activation energy but proceeds through a more defined and stable phase evolution pathway once activated. On the other hand, a different behavior is observed for 80T20L sample. The n app values for T p1 range between 2.46 and 3.72, reaching relatively higher values. This comparatively broad range of n app values suggests a more variable non-isothermal crystallization response for the first peak; however, it does not uniquely define a specific nucleation or growth geometry. This interpretation is consistent with the significantly higher E p1 determined for this peak. Accordingly, the initial crystallization phenomenon in 80T20L sample can be considered both characterized by a comparatively high apparent activation energy and a variable non-isothermal crystallization response. In contrast, the n app values for T p2 remain within a narrower range, indicating that the second crystallization step proceeds in a more controlled and consistent manner following the initial transformation. A critical point to emphasize is that the n values are not constant with respect to heating rate. If the crystallization process were governed by a single-step, invariant mechanism, similar n app values would be expected across different heating rates. Nevertheless, the observed variations suggest that the apparent crystallization response is sensitive to the applied heating rate. When considered together with the presence of multiple crystallization peaks in the DSC curves, these results are consistent with sequential or partially overlapping crystallization processes. In particular, the increased number of crystallization peaks and the variability in kinetic parameters for 80T20L sample suggest increasing crystallization complexity with increasing La 2 O 3 content.

Phase evolution and identification after heat treatment

To evaluate the crystallization phenomenon identified by DSC analyses and to correlate the kinetic findings with phase formation, the samples, 95T5L and 80T20L, were subjected to heat treatment processes at their corresponding crystallization peak temperatures (T p1 and T p2) using a heating rate of 10 K.min -1, held for 1 h, and finally cooled inside the furnace to room temperature before XRD analysis. Subsequent to these, XRD analyses were used to identify crystal phases and the related patterns are plotted in Figure 6. For the case of 95T5L sample, the XRD pattern obtained at T p1 reveals the formation of only a: α-TeO 2 (PDF 00-042-1365) and b: La 2 Te 6 O 15 (PDF 00-043-0551) phases. This limited phase assemblage suggests that the initial crystallization step proceeds through a relatively simple crystallization phenomenon associated with structural reorganization within the tellurite network. The relatively intense diffraction reflections assigned to α-TeO 2 suggests that crystallization at this stage is primarily governed by Te-based structural ordering, whereas fewer or less intense reflections were assigned to the La-containing phase. Some researchers have also reported similar phase evolution pathway for TeO 2 -based glass systems, especially the ones doped with lower additive contents similar to our case (Wajhal et al., 2020). In addition, at T p2, more crystal phases including c: TeO 2 (PDF 00-052-0795) and d: La 2.33 Te 1.67 O 6.85 (PDF 00-041-0260) emerge alongside the previously formed phases, (a) and (b). This observation suggests further phase development and diversification exist, confirming that the second crystallization phenomenon involves not only crystal growth but also new phase formation. The formation of these additional phases is consistent with the higher activation energy (E p2), suggesting that more extensive structural rearrangements are required at this stage. Overall, the observed phase evolution in 95T5L is consistent with a relatively sequential crystallization, transforming from limited phase formation, (a) and (b), to a more complex phase assemblage, (a) to (d).

In contrast to 95T5L, 80T20L sample exhibits significantly different crystallization kinetics. At T p1, multiple phases are already present, including a: La 2 O 3 (PDF 03-065-3185), b: La 2 Te 6 O 15 (PDF 00-043-0551), c: La 2 (TeO 3) 3 (PDF 00-051-0279), d: TeO 3 (PDF 00-021-1204), e: La 2 Te 4 O 11 (PDF 01-073-5353), and f: TeO 2 (PDF 00-052-1005). These observations show that multiple crystalline phases were present after heat treatment at T p1. Although this phase assemblage is consistent with a complex initial crystallization event, the ex-situ XRD results cannot establish whether all identified phases formed simultaneously during the DSC peak or developed sequentially during the subsequent 1 h holding period. Unlike 95T5L, crystallization does not proceed through a simple rearrangement of the tellurite network, but rather through a chemically more extensive reorganization including lanthanum containing phases. This situation is consistent with the comparatively higher apparent activation energy (E p1) obtained for the first crystallization peak. That means a substantial structural and chemical reorganization is required to initiate crystallization in the La-rich composition. Besides that, the XRD pattern at T p2 contains a more restricted set of clearly identified reflections, including those assigned to e: La 2 Te 4 O 11 being identified among the major crystalline phases. This situation may suggest that, after the initial crystallization event associated with the comparatively higher apparent activation energy, the system evolves toward a more stable and direct crystallization pathway. The reduced activation energy for the second crystallization step (E p2) is consistent with this behavior. In other words, subsequent transformations proceed more readily once the initial structural reorganization is achieved. Previous studies have reported somewhat similar crystallization behavior in TeO 2 -based glass systems doped with different oxides (Wajhal et al., 2019).

Taken together, the DSC and post-heat-treatment XRD results indicate that La 2 O 3 content substantially influences both the apparent crystallization kinetics and the crystalline phase assemblages obtained after thermal treatment. While the representative low-La composition exhibited comparatively limited phase assemblages after heat treatment, the high-La composition produced a greater diversity of identified crystalline phases. These observations support the interpretation that La 2 O 3 modifies the crystallization response of the glass system. Therefore, the observed phase evolution suggests composition-dependent differences between the relatively simple crystallization behavior of 95T5L and the more complex, multi-stage behavior of 80T20L.

The obtained results demonstrate that La 2 O 3 addition substantially modifies the structural characteristics of TeO 2 -based glasses. Although the measured glass density increased progressively with increasing La 2 O 3 content, the simultaneous increase in molar volume and decreases in oxygen packing density and crosslink density indicate that the density increase primarily originates from the greater molecular weight of La 2 O 3 rather than from more efficient packing of the glass network (Almuqrin et al., 2021), (Dorofeev et al., 2011). Moreover, the FTIR spectra showed composition-dependent changes in the Te-O vibrational bands while the fundamental tellurite framework was largely preserved. The increasing contribution of the bands associated with TeO 3 units may indicate compositional rearrangement within the tellurite network. Accordingly, the structural parameters and spectroscopic observations collectively indicate that La 2 O 3 modifies the arrangement and connectivity of the tellurite network without replacing its principal structural framework.

From a thermal perspective, increasing La 2 O 3 content systematically increased T g, T x, and the crystallization peak temperatures, indicating that structural relaxation and crystallization shifted toward higher temperatures. Similar increases in characteristic temperatures have been reported for rare-earth-containing tellurite glasses and are generally associated with reduced structural mobility and stronger interactions between the modifier cations and oxygen atoms (Saddeek and Elsayed, 2015), (Kalampounias, 2012). Nevertheless, the simultaneous decreases in ΔT, K H, and T rg show that the increase in T g did not correspond to improved resistance to crystallization. Instead, the La-rich compositions exhibited a narrower processing window and greater susceptibility to devitrification. This result confirms that thermal rigidity, represented by the temperature required for structural relaxation, and glass stability against crystallization are related but distinct characteristics (Oermann et al., 2009).

The non-isothermal kinetic analyses further demonstrate that the representative low- and high-La compositions exhibit markedly different apparent crystallization responses. The close agreement between the Kissinger and Ozawa methods confirms the internal consistency of the calculated activation energies (Lopes et al., 2014a), (Lopes et al., 2014b). Nevertheless, these activation energies should be regarded as apparent kinetic parameters describing the overall crystallization events represented by the DSC peaks rather than elementary atomic reaction barriers. Together with the apparent Avrami analysis, the results indicate that crystallization behavior depends on both composition and heating conditions under the present non-isothermal experimental framework (Ray et al., 2007). The observed differences between the representative compositions therefore reflect changes in the overall kinetic response of the crystallization process rather than direct evidence of specific atomic-scale mechanisms. Such behavior is consistent with multi-stage crystallization processes involving overlapping and sequential transformations in high-content rare-earth-containing oxide glasses (Gupta et al., 2018; Shi et al., 2025).

The post-heat-treatment XRD analyses revealed clear compositional differences in the crystalline phase assemblages obtained at T p1 and T p2. The 95T5L glass produced a comparatively limited number of identified phases, whereas several La-containing tellurate phases were detected in the heat-treated 80T20L composition. This difference is consistent with the greater compositional and phase complexity of the La-rich glass and with the additional crystallization event observed in its DSC curve. Nevertheless, the XRD patterns were recorded after 1 h of isothermal holding at the selected peak temperatures and therefore represent the phase assemblages remaining after heat treatment rather than the instantaneous phase constitution at the DSC peak. Overall, the combined XRD and kinetic results support the interpretation that increasing La 2 O 3 content modifies the crystallization behavior, leading to changes in both the apparent kinetic parameters and the resulting phase evolution.

The present study demonstrated that La 2 O 3 content strongly governs the crystallization behavior and phase evolution of TeO 2 -based glasses. XRD results confirmed that the amorphous structure was largely preserved up to 20 mol% La 2 O 3, whereas further increase to 30 mol% promoted crystallization and the formation of Te 4 O 9, La 2 Te 6 O 15, La 2.33 Te 1.67 O 6.85, and La 2 (TeO 3) 3 phases. Density increased from 5.22 to 5.51 g.cm -3 with increasing La 2 O 3 content, while oxygen packing density and crosslink density gradually decreased, indicating compositional modifications within the tellurite glass network. Thermal analyses revealed that T g increased from 640 to 741 K with increasing La 2 O 3 addition, whereas the glass stability parameter ΔT decreased from 37 to 25 K, demonstrating enhanced thermal rigidity but increased crystallization tendency. Non-isothermal kinetic analyses further suggested composition-dependent differences in crystallization behavior. In the low-La composition (95T5L), crystallization appeared to proceed through relatively sequential transformations with E p1 and E p2 values of approximately 192–194 and ∼261–274 kJ.mol -1, respectively. In contrast, the high-La composition (80T20L) exhibited a markedly different kinetic response, where E p1 increased to approximately 348–353 kJ.mol -1 while E p2 decreased to approximately 220–235 kJ.mol -1, suggesting a redistribution of the apparent activation energies associated with the crystallization events. Post-heat-treatment XRD analyses were consistent with these kinetic findings. While 95T5L exhibited a comparatively limited assemblage of identified crystalline phases after heat treatment at T p1, 80T20L exhibited a more complex assemblage containing multiple La-rich crystalline phases under the corresponding heat-treatment condition. In summary, the present results suggest that increasing La 2 O 3 content is associated with a composition-dependent evolution of crystallization behavior, progressing from relatively sequential crystallization in the low-La composition toward increasingly complex crystallization involving multiple crystalline phases in the high-La composition. These findings provide useful insight for researchers working on tellurite-based glass and glass-ceramic systems, particularly in studies focusing on crystallization control, phase-selective glass-ceramic design, and thermal stability optimization. Furthermore, the observed composition-dependent differences in crystallization behavior provide a scientific basis for future investigations of controlled thermal processing and the potential optical and photonic applications of tellurite-based glass-ceramic materials.