1. Introduction

The product of large emission cross-section (σ) and long metastable lifetimes (τ) of the lasing level determines the characteristic gain slope and gain in lasers and light amplification devices. The magnitude of emission cross-section is dependent on spectral broadening, determined by the strength of dipoles in a dielectric medium. However, from the radiative rate theory [1], the spectral broadening increases with the strength of electric dipoles, which the dopant ions experience in the medium. The spectral line width increases with the refractive index (n) of the dielectric medium [2,3]. By comparison, the radiative rate, which is inversely proportional to the lifetimes of lasing level, increases with the 5^th power of the refractive index. The product of lifetimes and emission cross-section follows the uncertainty principle on energy and time differential: $\Delta E.\Delta \tau \ge \frac{\hslash }{2}$, showing that the emission bandwidth (ΔE) is inversely related with the lifetime (Δτ). In this article, however we demonstrate how the glass-to-crystal phase transformation process in an Er³⁺-doped tellurite glass, which has a refractive index ~2.0, leads to simultaneous enhancement in the emission cross-section (determined by the strength at each wavelength across the spectral width) and prolonged lifetime of metastable state, ⁴I_13/2 level due to a significant change in the crystal-field effects.

The effect of large refractive index in tellurite glass on spectral width and its lifetime (τ) is very well known in the literature [2]. Since the first demonstration of light amplification using tellurite optical fibre as a medium, the search for larger bandwidth material with longer lifetime has continued for important device engineering applications in optical communication for multichannel transmission, widely tunable lasers, ultra-short light pulse generation, chemical sensing, and coherent tomography for medical imaging.

The presence of characteristic multi-structural units [TeO_x], where x=3 to 4, in tellurite glasses were first discovered by Stanworth in 1952 [4], and was successfully exploited for large bandwidth host for Nd³⁺ laser experiment by Wang and co-workers [5]. They showed that the multiple structural sites in tellurite glass can provide a means for dissolving large concentrations of rare-earth oxide than in the silicates, phosphates and fluoride [1, 6,7]. The structural aspects of large solubility of Er₂O₃ were further identified in tellurite glasses for engineering broadband amplifiers [3]. These evidences point out to potential for making ultrashort tellurite gain medium using traditional lithography and reactive ion etching, provided the mechanism for enhancing pump efficiency can be addressed [8]. The ⁴I_13/2→⁴I_15/2 level transition in Er³⁺-doped tellurite glass occurs radiatively with a metastable lifetime of 3.5 to 4.5 ms, depending upon the composition of hosts and dopant concentrations. In this transition in a waveguide shorter signal wavelengths are reabsorbed and re-emitted at longer wavelengths along the direction of propagation, which when combined with the intrinsic spectral broadening effects may contribute to enhanced tuning range for engineering both continuous wave and ultra-short pulse lasers for applications in optical communications, multichannel gas sensing, lidar technology, and materials processing.

As discussed above, the strength of electrostatic field around rare-earth dopant ions in a crystalline or glassy host affects the emission line shape and cross-section [1]. Furthermore in a glassy host there is an opportunity for further altering the local electric dipole strength by promoting devitrification via controlled nucleation and growth, which may then lead to an optically transparent glass ceramic for waveguide engineering. Although the glass-ceramic and its influence on colour development have been known since the Egyptians discovered the colloidal glass, the implementation of technique for device engineering has been few and far between. In this article, an example of Er³⁺-doped tellurite glass ceramic has been considered for demonstrating the combinatorial effects of structure and phase transformation on a major enhancement of photoluminescence line broadening and simultaneous lengthening of lifetime of ⁴I_13/2 level. The results are highly significant and can be applied for engineering a majority of ground state transitions, namely in Yb³⁺(1020–1080 nm), Tm³⁺(1900–1995nm), and Ho³⁺(2000–2100 nm) for engineering efficient lasers, both at small (10s mw) and large (>1 W) output powers via micro-scale fabrication and engineering.

2. Background

In this paper we have investigated 89.5 mol% TeO₂-10 mol% Na₂O glass with 0.5 mol% Er₂O₃ as a rare-earth oxide dopant. The binary sodium tellurite glass falls in the stable glass-forming region, as reported by Heo and co-workers [9] which were further analyzed in detail by Tagg and co-workers [10] by explaining the importance of Na₂Te₄O₉ (NT4) and Na₂Te₂O₅ (NT2) crystalline complexes. Tagg and co-workers emphasized the importance of ring-like structures, derived from a TeO₄ and TeO₃ combination, in the Na₂Te₄O₉ (NT4) complex. The structure comprises of tellurite sheets inter layered with sodium ions, which on increasing Na₂O concentrations progressively changes into a more ionic Na₂Te₂O₅ (NT2) complex. At the substructural level both NT4 and NT2 complexes have the building blocks of α- and β-TeO₂ crystals [10] and the polymerization of Na₂Te₂O₅ (NT2) complex with α- and β-TeO₂ yields the larger NT4 structure in the TeO₂-rich part of binary phase diagram, as reported by Zhu and co-workers [11]. In the 90 mol% TeO₂-10 mole% Na₂O composition range, a new crystalline complex with Na₂Te₈O₁₇ (NT8) exists [11] and is stable up to 615K below the eutectic temperature at 686K and at 72 mol% TeO₂. The formation of NT8 complex occurs in a similar manner via polymerization of NT4 with α- and β-TeO₂..

The composition with 89.5 mol% TeO₂-10.0 mol% Na₂O with 0.5 mol% Er₂O₃ lies very close to NT8, which implies that on reheating the sodium tellurite glass might devitrify in a polymorphic manner, with a minimal change in the composition between the glass and the nucleating crystal. To our knowledge there is no spectroscopic data in the literature which considers the influence of polymorphic phase transformation on the strength and line shape of an optical transition for device engineering. Since in the binary sodium tellurite system, reported by Zhu and co-worker [11], the NT8 crystals eventually decompose above 615 K by forming TeO₂ and NT4 phases, the changes in the electric dipole environment of Er³⁺-ions can be readily quantified spectroscopically. We show that the observed polymorphic transition literally doubles the bandwidth in a sodium tellurite glass-ceramic, with concomitant enhancement in the lifetime. Since the phase transformation observed herein is of polymorphic nature, it would also help in reducing the scattering loss in the waveguide due to the negligibly small difference in the refractive indices of the glass matrix and crystals.

3. Experimental

The binary sodium tellurite composition: 10 mol% Na₂O-89.5 mol% TeO₂-0.5 mol% Er₂O₃ was prepared by weighing and mixing 99.999 weight percent pure Na₂CO₃, and TeO₂ in the required molar percentages in a glove box. The mixture was then melted in a gold crucible inside a resistance furnace in an atmosphere of flowing air (1.5 litre min-1) at 1073 K at which temperature it was allowed to homogenize for 30 minutes. The homogenized composition was then cast into brass mould, preheated below the glass transition (T_g~550K) temperature and the glass was annealed at T_g for 2 hours, before being cooled down slowly to room temperature at a rate of 0.25 K min^-1. The glass sample was cut into a shape of 5 mm×5 mm size and each surface was polished to a ~1µm finish for heat treatment experiment for 10 hours in air.

In order to determine the dependence of crystallite size on the emission line shapes of heat treated Er³⁺ doped glass samples, we analyzed the X-ray diffraction (XRD) spectrum (Cu K_α radiation) for each sample and determined the average crystallite size (nm) using the Debye-Scherrer formula [12].

In eq.1, the values of k and λ are 0.95 and 0.15406 nm, respectively, β is the full width at half maximum or half width in radians and θ is the peak position of a diffraction peak.

Emission spectroscopic measurements were carried out using the 800 nm line of Ti-Sapphire laser (Schwartz Electro Optic, model CWBB, Orlando, Florida), which was pumped with an Ar-ion (Innova Coherent (Ar⁺) laser) having an output power of 5 Watts. The effect of radiation re-absorption and re-emission was minimized by allowing the pump to fall at one edge, and the emitted signal was collected 1 mm away from the pump launch edge in the orthogonal direction [13,16]. The method adopted for the measurement of lifetimes of the ⁴I_13/2 level has been described extensively in refs [3, 5, 8, 13].

4. Results and Discussion

 

Fig 1. XRD for 0.5 Er₂O₃ 10 Na₂O 89.5 TeO₂ glass crystallized at different temperature for 10 hours and not heat treated (NHT) sample. 593, 613, 653 and 693 K sample showing Na₂Te₈O₁₇ crystal.

Download Full Size | PDF

In Fig. 1 the XRD patterns of heat treated and as-cast glass samples are compared. The inset in Fig. 1 shows the glass transition temperature at around 550K, a major devitrification exotherm at around 650K, followed by a sharp melting peak at around 920K. All these thermal events agree well with the literature data [9–11]. The X-ray powder diffraction data below 693K confirmed the presence of NT8 and TeO₂ crystals in the glass host. The formation of TeO₂ and NT4 crystals begins above 613 K which is in remarkable agreement with the data presented by Zhu and co-workers [11] and is consistent with the devitrification onset below 650K in the Fig. 1 inset. The characteristic diffraction peaks for NT8 crystals are more evident at 613K. Besides NT8, NT4 and TeO₂ crystals, we also find that there is a minor presence of NaErO₂ (sodium erbate) monoclinic crystals whose characteristic peaks are at 2θ (degrees)=17.27, 17.54, 26.11, 31.15, 32.70, 32.86, 38.61, 46.91, 47.23, 48.05, 49.70, 53.4, 54.4. The transformation of NT8 can be explained by: Na₂Te₄O₉(NT4)+4TeO₂=Na₂Te₈O₁₇(NT8) and the residual Er₂O₃ combines with a proportionate amount of Na₂O to form NaErO₂. Since there is uncompounded TeO₂ available as a result of the decomposition of NT8, the erbium oxide present as a dopant is also likely to react with TeO₂ and form spinel structures Er₂Te₅O₁₃ and Er₆Te₅O_19.2 below 693K, which are difficult to distinguish in the XRD pattern for sample at 693K in Fig. 1. This is because of the small phase volume of erbium tellurites and their strongest peaks being obscured by the presence of dominant phases. Using the X-ray line broadening data in Fig 1 for the strongest peak at 2θ~27.7°, we determined the average crystallite size for NT8 crystals which are summarized in Table 1. The observed growth rate follows the Arrhenius relationship closely in the 583K and 693K range, over which the average crystal size increased from 9±2 nm to 482±2 nm. From the heat treatment and phase analysis, we find that above 613K the dominant phases are NT8 and the residual TeO₂, until 643K when NT4 begins to form.

Table 1. The dependence of FWHM, lifetime of 4I13/2 level with 800 nm pump and crystal size on temperature for heat treated sodium tellurite (0.5 Er2O3 10 Na2O 89.5 TeO2) glass for 10 hours.

View Table

Further confirmation of phase changes were also followed by the Raman spectroscopic analysis using λ_ex=633 nm, which is ideal for Er-doped materials, as there are no absorption peaks for the dopant at this wavelength. A significant difference in the optical phonon structure is evident in Fig. 2 for samples annealed over a range of temperatures. At 593K the vibrational mode structures are less defined [14], which characteristically changes when NT8 and TeO₂ crystals dominate above 613K. The large broad peak decomposes to several sharp peaks falling in the 580–820 cm^-1 range. The sharpness of Raman peaks in the spectra above 643K characterizes the presence of crystalline phases in which dominant vibrational modes have energies centered around 640 cm^-1 and 720 cm^-1 [14]. The Raman spectrum of annealed material at 693K differs significantly from that at 653K, confirming that these changes in the vibrational modes agree well with the phase transformation reaction, proposed by Zhu and co-worker [11].

 

Fig 2. A comparison of Raman spectra for annealed sodium tellurite glass ceramics using excitation wavelength at 633 nm.

Download Full Size | PDF

Figure 3 compares the emission spectra (normalized at 1532 nm) of Er³⁺ ions in the glass and glass-ceramic samples. Each emission spectrum at an isotherm appears similar in shape, however the relative intensities at specific wavelengths 1500, 1560 and above 1600 nm are enhanced in Fig. 3, Fig. 4 confirms that the crystallization is enhancing the magnitude of electric dipole strength significantly, which are consistent with the data, reported elsewhere [3,13]. The samples heat treated above 693 K were opaque to the naked eye and showed strong scattering and were therefore excluded from the investigation. In Fig. 5 we also compare the strength of emission intensities peaked at visible wavelengths at 550 nm and 655nm, which are related to optical transitions from ⁴S_3/2 and ⁴F_9/2, respectively. The strength of visible transitions also confirms that the light scattering is not dominant in materials, heat treated up to 653 K.

 

Fig 3. Er³⁺ ion emission spectroscopy for heat treated sodium tellurite (0.5 Er₂O₃ 10 Na₂O 89.5 TeO₂) glass for 10 hours. NHT is non heat treated glass.

Download Full Size | PDF

 

Fig 4. Er³⁺ ion emission line FWHM and lifetimes for ⁴I_13/2 level heat treated sodium tellurite (0.5 Er₂O₃ 10 Na₂O 89.5 TeO₂) glass for 10 hours.

Download Full Size | PDF

From Fig. 4, it is evident that the standard relationship for radiative rate (lifetime) and the emission cross-section [1], as described by the uncertainty principle above, is not followed due to the enhancement in the local field effects at the structural scale. In Fig. 4 we see that both the emission linewidth and lifetimes increase with the annealing temperature and then decrease abruptly at 693 K. In the annealing range of 583 to 653K the average crystallite size increases from 9 nm to nearly 129 nm, as shown in Table 1. Crystallinity is confirmed by XRD pattern and glass-ceramics are transparent to naked eye. The glass-ceramic samples remain transparent and show little evidence for scattering, based on the X-ray diffraction and Raman spectroscopy data in Figs. 1 and 2, respectively.

On the basis of X-ray diffraction, phase and crystal growth analyses, the spectral broadening in the glass-ceramic materials appears to be predominantly due to the presence of NT8 crystals which begin to form above the glass transition temperature (T_g~543K) and dominate the phase constitution up to 640K. In the temperature range of 640–653K, the NT8 crystals begin to decompose and form NT4 and TeO₂ with a small phase volume of NaErO₂. From the structural analysis of NT4, reported by Tagg and co-workers [10], the sheet-like network of [Te₄O₉]^2- is separated by Na⁺ ions (ionic radius=1.02Å [10,14–15]) which are surrounded by either 5 or 6 oxygen ions in a pseudo-octahedral cage. The average Na-O distance varies between 0.2926 nm and 0.2713 nm, indicating that the sodium ions might have two electric dipole environments ranging from octahedron and pseudo tetrahedron geometries. On the other hand the rare-earth ions, e.g. Er³⁺, prefer to reside in an octahedron co-ordination. However the energy consideration due to its 3⁺ charge and size (0.89 Å) in the NT4 crystals, the Er³⁺-ions may be forced to share both octahedral and tetrahedral sites. The oxygen ions form both bridging and non-bridging sites with Te ions, and therefore further add to the complex dipole environment. The non-bridging oxygen sites are shared by both Er³⁺- and Na⁺. Above all it is interesting to note that average co-ordination of Te in the structure is 4.3, as opposed to 4 in the bipyramid structure [10,14–15].

 

Fig 5. A comparison of visible emissions at around 550 nm (green) and 645 nm (red) in heat treated sodium tellurite glass ceramic materials

Download Full Size | PDF

Based on the structural model of Tagg and co-worker and phase equilibrium analysis [9–11], it is evident that the NT8 structure is simply a polymeric extension of NT4 by adding more TeO₄ units from α and β TeO₂ in the ring structure. In NT8, the dopant Er³⁺-ions remain in a similar co-ordination environment as in NT4, indicating that the minimum Er³⁺-Er³⁺ distance will be at least comparable with that in NT4 which varies between 0.3244 and 0.3476 nm. The Te-Te distances for making a ring-shape sheet, depending on their sites, vary between 0.3165 and 0.3865 nm [3, 10]. The NT8 structure, which nucleates and grows between 543 K and 650 K is therefore likely to disperse Er³⁺-ions at least by 0.3476 nm, if not longer due to the lower concentrations of the ion. Since the Er³⁺ and Na⁺ ions compete for the same site in a polymeric structure of NT8/NT4, our proposed estimation is that minimum the Er-Er distance might be determined by the availability of vacant Na-ion sites in the tetrahedron or octahedron site in the polymeric structures. Under this limiting condition for availability of Er-site, the radiation will prefer to migrate from site-to- site and quench only when an OH- ion or Er-Er ion clustering occurs [13], which is why we expect longer lifetime. We also wish to point out there is no structural method available which allows us to determine the Er-Er distance accurately, except an estimation based on structural models. Enhanced probability of radiation migration via coupled dipoles appears to be responsible for such a large broadening and strengthening of emission line. Above 643K the NaErO₂ phase begins to nucleate and after the decomposition of NT8 is completed by 653K, NaErO₂ acts as a potential quenching site for radiation migration, which is why the lifetime is shortened abruptly together with the emission line broadening, due the presence of erbate phase.

5. Conclusions

A 10 mol% Na₂O-89.5 mol% TeO₂ 0.5 mol % Er₂O₃ glass has been converted to glass-ceramics via heat treatment. These samples are studied for the fluorescence and line broadening behavior of Er-ions in the glass-ceramic matrix. The influence of crystal growth on line broadening of ⁴I_13/2 spectra and metastable lifetime of ⁴I_13/2 level i Shows that the full-width-of-half maxima at ⁴I_13/2 level increases from 59 to 137 nm when crystal size increased from 9 to 129 nm. A further increase in crystal size drastically decreased the FWHM to 67 nm. The lifetime of ⁴I_13/2 level also increases from 4.2 ms to 5.4 ms and further increase in crystallite size shortens it to 3.8 ms.