Thermal stability and spectroscopic study of Ho³⁺/Yb³⁺ co-doped fluorophosphates glasses

2.1 Glass preparation

Samples of fluorophosphate glasses with general formula (in mol%): (80-x-y) NaPO₃-10SrF₂-10ZnF₂- xHoF₃- yYbF₃ (x = 1 and y = 0.5, 1, 1.5 and 2 mol%) were prepared using high-purity of NaPO₃, SrF₂, ZnF₂, HoF₃ and YbF₃ powder. The doping concentration of the rare earth was set at 1 mol% for Ho³⁺ ions and varying from 0.5 to 2 mol% for Yb³⁺ ions. A stoichiometric mixture of the starting materials (about 10 g) is placed in a platinum crucible and melted at 900 °C for 15 min. The melts were cast on preheated brass molds at a temperature 10 °C below the glass transition temperature. Thermal annealing of the samples is performed for several hours, in order to eliminate the internal stresses of mechanical or thermal origins created during quenching, following by a slow cooling to room temperature. After annealing, the glass samples were cut and polished. The samples obtained after polishing have parallel faces allowing the optical characterization. In the other hand, some local defects, such as scratches and unevenness, are difficult to avoid.

2.2 Measurements

The characteristic temperatures (temperature of glass transition T_g and temperature of onset crystallization T_x) were determined by differential scanning calorimetry (DSC) using DSC TA Instrument with a programmed heating rate of 10 °C/min. The accuracy on the temperature is about ±2 °C. Density was measured by Archimedean method with an accuracy of ±0.001. Refractive index was measured by an Abbe refractometer with a mean error ±0.001. Absorption spectra were recorded at room temperature using a double beam spectrophotometer UV–Vis – Near IR CARY 5G brand operating between 200 and 3000 nm with a spectral resolution of 0.1 nm.

3.1 Glass samples

Fig. 1 shows all the samples: non-doped, doped with Ho³⁺ only and co-doped H³⁺/Yb³⁺. Compositions of glass samples were presented in Table 1.

Close

Fig. 1

Fluorophosphates glass Samples.

Export to PPT

3.2 Thermal properties, density and refractive index

Fig. 2 represents the DSC curves of the samples:j without any doping ions (dash-dotted line), doped with Ho³⁺ only (dotted line) and Ho³⁺/Yb³⁺ co-doped (with 1 mol%YbF₃, solid line). Glass transition temperature T_g for NPSZ_HYy series still situated around 250 °C and the crystallization onset T_x around 450 °C.

Close

Fig. 2

DSC Curves of the non-doped, Ho³⁺ doped and Ho³⁺/Yb³⁺ co-doped glass samples.

Export to PPT

The values of refractive index and the density of glass samples were presented in Table 2. The results show a small almost linear variation of refractive index and density with the variation of YbF₃ content.

3.3 Absorption spectra and Judd-Ofelt analysis

The absorption spectra are shown in Fig. 3. We can count ten absorption bands centered at 1957, 1155, 642, 538, 486, 472, 448, 418, 386 and 360 nm corresponding to the optical transitions of the Ho³⁺ ion from its ground state ⁵I₈ to the different excited states: ⁵I₇, ⁵I₆, ⁵F₅, ⁵F₄, ⁵F₃, (⁵F₂, ³K₈), ⁵G₆, ⁵G₅, ⁵G₄ and ³H₆ respectively. We can also observe an absorption band at 980 nm correspond to the transition ²F_7/2 to ²F_5/2 of the Yb³⁺ ion in the four co-doped samples.

Close

Fig. 3

Absorption spectra of NPZS_HY samples.

Export to PPT

The different spectroscopic parameters of the trivalent rare-earth ions in various hosts can be calculated by the application of the theory proposed by Judd (Judd, 1962) and Ofelt (Ofelt, 1962). The detailed assumptions of the theory have been described in original articles. A brief summary of the theory will present below taking into account only the formulas necessary to determine the different spectroscopic parameters.

The intensities of intraconfigurational f-f transitions of trivalent rare earths observed in the absorption spectra can be described by the oscillator strengths f_mes of each J → J′ transition. The majority of these transitions are induced electric dipole transitions, although a few magnetic dipole transitions are also present. From the absorption spectra, we can calculate the oscillator strength f_mes, which is proportional to the band absorption intensity, from the value of the absorption coefficient α(λ) at a particular wavelength λ according to the formula:

On the other side, the oscillator strength can be given in term of the electric dipole line strength S_ed:

The values of the Ω_k are empirically determined by comparing the computed values starting from the formula (2) with the values obtained from the absorption spectra at ambient temperature of the oscillator strengths (formula (1)). If q is the number of the absorption bands considered in experiments, the resolution of a system of q equation to 3 unknown factors, by a least square approximation, makes it possible to obtain the values of the parameters Ω_k. These parameters are expressed in cm². A measure of the accuracy of the fit is given by the root mean square deviation (RMS):

The values of Judd-Ofelt parameters obtained are used to calculate the line strength of the transitions between the initial state J and the final state J′ using the Eq. (3). The probabilities of the radiative transitions are given by the equation:

The branching ratios can be to obtain from the probabilities of radiative transitions A_rad by the equation:

The radiative lifetime of the level J is given by:

Another important term, integrated emission cross-section (in m), which is particularly useful to determine the possibility of lasing in glass, is defined as:

When the value of the integrated emission cross-section is close to or greater than ∼10⁻²⁰ m, there is a possibility of lasing (Watekar et al., 2008).

The results obtained from the Judd-Ofelt analysis are reported in Tables 3–5.

4.1 Thermal stability

Determination of glass transition temperature (T_g) and onset crystallization (T_x) is interesting for many reasons: in practical terms, T_g is a measure of approximate upper use temperature for a given composition, and annealing of glass samples is carried out somewhat below T_g. In addition, T_g is an important factor for laser glass. It is reported that a glass with high T_g has a good thermal stability to resist thermal damage at high pumping power (Xu et al., 2011a,b). T_x defines a safe upper limit for processing the melt if crystallization is to be avoided. From the DSC curves, one can observe that there is no crystallization peak till T = 450 ± 2°C for all samples, T_g is about 247 ± 2 °C and T_x is about 460 ± 2 °C for the glass sample without any doping ions and for glasses containing HoF₃ and YbF₃, the vitreous transition still situated around 250 °C and crystallization onset around 450 °C. One can notice that the glass transition temperature is shifted upwards high temperatures and the onset crystallization is shifted downwards low temperatures by addition of HoF₃ and YbF₃, results in agreement with previous works (Imanieh et al., 2014). As the changes of T_g and T_x is largely depending on the major mixing elements, because the temperature is closely related to the bonding force among the constituent elements, one can explain the shifting of T_g and T_x by structure arrangement of glass due to participation of Yb³⁺ and Ho³⁺ ions and enhanced glassy network structure (Chen et al., 2014). However, it has been observed that with high concentration of rare-earth, the glass became impossible to pour due to its high instability and during the pouring it could be crystallize to become opaque (Dantelle et al., 2005). Thus, an optimization of rare earth content is necessary to obtain a transparent glass.

The quantity ΔT = T_x − T_g, known as the supercooled liquid region, defined as the temperature gap between T_g and T_x. It has been used as a rough measure of the glass forming ability of a melt, i.e., of resistance against crystallization during casting (Wang, 2008). Large ΔT means strong inhibition to processes of nucleation and crystallization (Park et al., 2010; Xu et al., 2011a,b). For this reason, it is frequently used as an assessment criterion to determine the stability of glass. To achieve a large working range during operations such as fiber fabrication or the preparation of bulk glass articles, it is desirable to have ΔT as large as possible. We find that all glass samples have relatively a large value of ΔT (∼200 °C) close to the value obtained for fluorophosphate glasses (Tian et al., 2010; Chen et al., 2014), which is significantly higher than that of other kinds of rare earth doped fluorophosphate glasses such as Tm³⁺ (Liao et al., 2007). This result means that all our glass samples are stable against crystallization.

From the analysis of DSC curve, we can conclude that thermal properties of our glass samples are good for fiber drawing.

4.2 Absorption spectra and Judd-Ofelt analysis

Absorption spectra of the Ho³⁺ and Ho³⁺/Yb³⁺ co-doped glass samples are shown in Fig. 3. The spectroscopic properties of our glass samples were determined using the Judd-Ofelt model. The f-f transitions are considered to be electric dipole in nature, because the magnetic dipole oscillator strength (f_md) will be relatively small (Sooraj Hussain et al., 2006). For that reason in the present work, these magnetic dipole line strengths have not been considered. Measured and calculated oscillator strength for Ho³⁺ ions in NPSZ_HY samples were presented in Table 3.The low RMS values suggest the good agreement between calculated and experimental oscillator strengths of Ho³⁺ ions in our glasses samples which indicates the validity of the Judd-Ofelt model for predicting the spectroscopic properties of Ho³⁺.

The Judd-Ofelt parameters Ω_k (k = 2, 4, 6) are mainly depend on the host glass composition. These parameters can provide versatile information regarding the rare earth in glass structure. Some empirical correlations of the Judd-Ofelt parameters and the local structure of the rare earth ions have been reported in literature (Jacobs and Weber, 1976; Ebendorff-Heidepriem et al., 1998; Malta and Carlos, 2003). Generally, Ω₂ is an indicator of the covalency of the rare earth-ligand bonds (short-range effect), and it is hypersensitive to the compositional changes in the host materials. Ω₂ is also related with the symmetry of ligand field in the glass host. According to previous studies (Ebendorff-Heidepriem and Ehrt, 1999), Ω₂ increase with the increasing of covalency and asymmetry at the rare earth sites. The values for Ω₄ and Ω₆ provide some information of the rigidity of the host materials (Wang, 2008). They depend on bulk properties such as viscosity and dielectric constant of the media (long-range effects). They are also affected by the vibronic transitions of the rare earth ions bound to the ligand atoms. In the present work, the Judd-Ofelt parameters Ω_λ are calculated and compared with those of various Ho³⁺ doped glasses, the results were presented in Table 4. One can observe, from the values of Ω₂ of our samples, that the covalent environment for Ho³⁺ is slightly decreased with increasing of Yb³⁺ amount in glasses. The observed Ω₂ values of our samples are located between the higher side to the values reported for ionic glasses (for fluoride glasses, Ω₂ ∼ 2 × 10⁻²⁰ cm²) and those of the covalent glasses (for phosphates Ω₂ ∼ 5 × 10⁻²⁰ cm²). Theoretically, the polarizability of oxide ions is higher than that of fluorine ions, resulting in the increasing covalence of the bonds between rare earth ions and surrounding ligand from fluoride to phosphate glasses due to substitution of fluorine ions by oxygen ions. As fluorophosphate glasses has both O²⁻ and F⁻ ions, the value of Ω₂ is larger than that of fluoride glass and smaller than that of phosphate glass (Tian et al., 2010). However, the observed lower values of Ω₂ parameter for Ho³⁺/Yb³⁺ co-doped NPSZ_HY glasses compared by those of phosphate glasses indicates that the ligand asymmetry around the rare earth ions in Ho³⁺/Yb³⁺ co-doped NPSZ_HY glasses is weaker than those in phosphate glasses. Additionally, a larger modifier ion in phosphate glass gives rise to a larger average between P-O-P chains causing the increasing average Ho-O distance to increase. Such increase in the bond lengths produces weaker local field near rare earth ions and lead to lower value of Ω₂ (Rao et al., 2012).

According to the theory of Jacobs and Weber (Jacobs and Weber, 1976), the rare earth emission intensity can be characterized uniquely by Ω₄ and Ω₆ parameters. Thus, we used the so-called spectroscopic quality factor (Ω₄/Ω₆). This factor is important in predicting the behavior of various lasing transitions in a given matrix. Based on this factor, it is found that the NPZS_HYy glasses appear to be better optical glasses. Moreover, it is noticed that the values of Judd-Ofelt parameters of our sample glasses follows the trend Ω₂ > Ω₄ > Ω₆ in consistent with that one observed for halogeno-phosphate, phosphate and germanate glasses; nevertheless, it differs from those of fluoride and tellurite glasses.

Table 5 presents the predicted radiative transition probability A_r, branching ratios β, integrated emission cross-section and radiative lifetimes τ_rad for ⁵I₇ → ⁵I₈ transition of Ho³⁺ ions in NPSZ_HY samples. The results show that the values A_rad of the ⁵I₇ level of Ho³⁺ were 68.4, 70.2, 79.2 and 63.0 s⁻¹, which are similar to the value found for fluorophosphates glass and smaller than that of germanate and fluoride glasses. It is reasonable to obtain such result because the radiative transition probability is proportional to n(n² + 2)²/9 (Eq. (5) in Section 3.3) and the refractive index of fluorophosphates glasses is usually less then germanate and fluoride glasses. It is evident to find also that the radiative lifetimes of the ⁵I₇ level of Ho³⁺ for NPZS_HYy glasses are greater than those of germanate and fluoride glasses. Further, the radiative lifetimes values for Ho³⁺/Yb³⁺ co–doped NPZS_HYy glasses are found to be smaller than that for individually Ho³⁺ doped NPZS_H1 glass. This result, which may be explained by the crucial contribution of phonon subsystem, is in agreement with those found by Gandhi et al. (2010) for the measured lifetime of blue, green and red transitions.

The values of the integrated emission cross-sections obtained are greater than ∼10⁻²⁰ m, which indicates a possibility of lasing with these glasses. But it must be mentioned that lasing can be specifically determined only if emission cross-section and fluorescence lifetime are taken into account.