1. Introduction

Owing to the rich energy level structure, trivalent rare-earth ions can achieve luminescence with wavelengths ranging from ultraviolet (UV), visible to near/mid-infrared (NIR/MIR) bands under an appropriate optical excitation. Consequently, the study of luminescence characteristics of glass materials doped with rare earths has been the hotspot in developing optoelectronic devices like fiber-optic amplifiers, illumination displays and solid-state lasers. Meanwhile, because of the rich energy level structure, certain energy levels of various rare earth ions have comparable energy heights, and/or the energy gap between two energy levels of one ion is close to that of another ion. These characteristics make the co-doping of various rare earth ions become an efficient technical method to improve the luminescence performance of glass materials. With the co-doping of two rare earth ions for instance, it can make an enhancement in the luminescent intensity of the acceptor ion through the sensitization mechanism, or broaden the luminescent spectrum through the spectral overlapping of two adjacent luminescent bands from the co-doping ions. In addition, due to the selection rule of energy level transition, rare earth ions have strong absorption at some bands such as Er³⁺ ions at the 1480 nm band [1], Yb³⁺ ions at the 980 nm band [2], and Tm³⁺ ions at the 800 nm band [3], just corresponding to the working wavelengths of commercial laser diodes (LDs). Therefore, under the excitation of appropriate commercial lasers, co-doping with these rare earth ions can be expected to improve the pumping efficiency.

As a band between the MIR and the NIR regions, 2.0 µm band laser has wide and potential applications in environmental detection, biomedicine, remote sensing communication and other areas [4–6]. In the aspect of environmental monitoring, for instance, the main components of the atmosphere, like CO₂, H₂O, CO and N₂O, have characteristic absorption peaks in the range of 2∼3 µm, so using this band laser can monitor the change of these gas contents in the environment. In the biomedicine, water molecules have a high absorption coefficient (∼100 cm^-1) near the 2.0 µm band. When using 2.0 µm band laser to act on biological tissues, a large amount of heat will be generated so that the diseased tissues can be quickly removed. It has the advantages of high cutting accuracy, small postoperative wound, and good hemostasis. While in the remote sensing communication, laser in the 2.0 µm band possesses high penetration capability to the atmosphere because it is located in the one of atmospheric transmission window (1∼3 µm), so it can be directly applied to coherent Doppler wind radar and laser ranging; In addition, because of the inherent Rayleigh scattering effect of materials, if 2.0 µm band is used as the laser working wavelength instead of 1.55 µm band, the fiber loss caused by Rayleigh loss can be effectively reduced and the distance of relay free communication can be increased accordingly.

Currently, Ho³⁺ is among the most efficient rare earth ions in producing lasing emission at 2.0 µm band. Its ⁵I₇→⁵I₈ radiative transition with 2.0 µm band emission has a long fluorescence lifetime and great stimulated emission cross section. Nevertheless, so far, no commercial LD is available on the market that can excite Ho³⁺ directly to produce fluorescence at the 2.0 µm band. Amongst the trivalent rare earth ions, Tm³⁺ can be pumped by commercially 808 nm LD to emit strong 1.83 µm band lasing emission through the ³F₄→³H₆ radiative transition, and also the pump energy absorbed can be transferred to Ho³⁺:⁵I₇ level from the Tm³⁺:³F₄ level to make the latter produce fluorescence [7]. In recent years, there are many reports on the adopting of Tm³⁺/Ho³⁺ co-doping scheme to achieve ∼2.0 µm band broadband luminescence. For example, J. Kang et al. [8] prepared Tm³⁺/Ho³⁺ co-doped lanthanum aluminosilicate glass and photonic crystal fiber (PCF), and a broadband luminescence in the range of 1600 to 2200 nm was observed in the glass while an optical-to-optical efficiency 6.37% of this broadband output spectrum was obtained in the PCF. The broadband luminescence characteristics and laser performance were investigated. SK. Taherunnisa et al. [9] synthesized Tm³⁺/Ho³⁺ co-doped sodium-sulfo lead phosphate glasses, and the NIR and MIR photoluminescence spectra were recorded at 797 nm excitation. The studies of luminescence properties of Tm³⁺:1.8 µm and Ho³⁺:2.0 µm bands with doped concentrations suggested that Tm³⁺/Ho³⁺ co-doped phosphate glass is advantageous to generate 1600–2200 nm broadband luminescence. N.M. Ty et al. [10] reported the broadband flat NIR/MIR emissions in Tm³⁺–Ho³⁺ co-doped and Tm³⁺–Ho³⁺–Yb³⁺ tri-doped zinc silicate glasses under excitation of 808 and 980 nm LD respectively. Broadband relatively flat fluorescence emissions in the range of 1600 to 2200 nm in Tm³⁺–Ho³⁺ co-doped and Tm³⁺–Ho³⁺–Yb³⁺ tri-doped glasses were identified, and the influences of Ho³⁺ and Yb³⁺ concentrations on the broadband luminescence properties were investigated. Obviously, the realization of broadband luminescence is mainly attributed to the energy transfer from Tm³⁺:³F₄ level to Ho³⁺:⁵I₇ level. However, an energy difference exists between these two levels, so the energy transfer between them needs to be completed by phonon assistance (Tm³⁺:³F₄ → Ho³⁺:⁵I₇ + phonons), and the phonon energy of glass host has an important influence on the completion of energy transfer.

In this paper, trivalent Ce³⁺ ions and WO₃ component with moderate phonon energy were introduced into Tm³⁺/Ho³⁺ co-doping tellurite glass, and the improved effect of Ce³⁺ tri-doping and WO₃ introduction on the broadband luminescence of ∼2.0 µm band was reported. The broadband luminescence mechanism of Tm³⁺/Ho³⁺ co-doping, and the cross-relaxation process between Ce³⁺ and Tm³⁺ ions as well as the variation of host phonon energy induced by WO₃ component which are responsible for the luminescence properties were studied. In comparison to traditional glasses like silica glass, silicate glass, aluminosilicate glass and borate glass, tellurite glass possesses some unique advantages [11–13], such as high refractive index (∼2.0), large rare earth solubility (several time of silica glass), wide visible-to-infrared transmission window (∼0.4–5 µm) and low maximum phonon energy (∼750 cm^-1). High refractive index can bring large emission cross-section and transition probability, large rare earth solubility can provide high gain per unit length, and low phonon energy can improve radiative quantum efficiency. All these make tellurite glass a potential host material for trivalent rare earth ions.

2. Experimental procedures

In this work, the melt-quenching technique was applied to prepare the doping tellurite glasses, and the detailed component amounts and corresponding sample names were listed in Table 1. To prepare glass sample, accurately weigh 15 g oxide powders as a batch according to the proportion in the table with a precision balance, pour it into the platinum crucible and stir it evenly. The platinum crucible was then heated to about 320 °C for 30 min with an aim of removing residual moisture from the powders. It was next transferred to a resistance furnace and melted in a dry atmosphere at 900 °C for 45 min, stirred by a quartz glass rod for 10 min in melting process to remove air bubbles generated and also to make molten liquid more homogeneous. Afterwards, the molten liquid was poured into a preheated silica mold and transferred instantly to a 350 °C annealing furnace for 2 hours for the release of thermal stress. Finally, the glass sample was polished into thin disc-shaped block with 1.6 mm thickness and 10 mm diameter for physical and spectroscopic tests after cooling to the ambient temperature.

Table 1. The compositions (mol %) of tellurite glass and corresponding name

View Table | View all tables in this article

The density of glass sample was determined based on the Archimedes principle by using the distilled water as an immersion solution. Refractive index was measured at 632.8 nm by prism coupler of Sialon Technology. X-ray diffractogram (XRD) pattern was obtained by Bruker D8 Advance power diffractometer from 10° to 70° (2θ) under a Cu-Ka radiation. Raman spectrum was obtained by Raman microscope of Renishaw inVia. Thermal stability was evaluated by differential scanning calorimeter (DSC) of TA Q2000 instrument. Absorption spectrum between 400 and 2200 nm was acquired by Perkin Elmer Lambda 950 UV-Vis-NIR spectrophotometer. Fluorescence spectrum in VIS-NIR range was acquired by FLSP920 fluorescence spectrometer of Edinburgh Instruments at excitation of 808 nm laser diode (LD). A digital oscilloscope was used to obtain fluorescence decay curve under 808 nm pulse excitation utilizing the same testing system. The measured conditions for all glass samples were kept the same so as to get a comparable result, and the physical parameters and the morphology of tellurite glass samples obtained are shown in Table 2.

Table 2. Density ($\rho$), refractive index ($n$), doping concentrations of Tm³⁺ (${N_{\textrm{Tm}}}$), Ho³⁺ (${N_{\textrm{Ho}}}$) and Ce³⁺ (${N_{\textrm{Ce}}}$), and photo of the prepared glass sample.

View Table | View all tables in this article

3. Results and discussion

3.1 XRD pattern, DSC curve and Raman spectrum

Figure 1 displayed the XRD patterns of tellurite glass samples with Tm³⁺/Ho³⁺ co-doping (T-TH0 and T-TH3) and Tm³⁺/Ho³⁺/Ce³⁺ tri-doping (T-THC0 and T-THC3) with and without WO₃ component. As seen from the figure that in the diffraction angle range (2θ) of 10–70°, the XRD profiles of glass samples looks almost the same which have the common characteristics, i.e. there is no sharp diffraction peaks but an intense hump located at 15–40°, indicating that there is no periodicity of atom arrangement in three dimensional space. This means that the physical structure of synthesized tellurite glasses is in amorphous state [14,15]. As the representative, inset of Fig. 1 displayed the DSC curves of glass samples T-TH0 and T-TH3. Generally, the thermal stability of glass hosts is characterized by difference $\Delta T$ (=${T_\textrm{x}}$−${T_\textrm{g}}$) between the crystallization onset temperature (${T_\textrm{x}}$) and transition temperature (${T_\textrm{g}}$) [16]. If the difference $\Delta T$ is greater than 100 °C, it implies that the glass host has good glass-forming ability and anti-crystallization stability, and can allow a wide working temperature range during fiber drawing. Therefore, it can be concluded from the figure that the present studied tellurite glasses possess good thermal stability.

 

Fig. 1. XRD patterns of the synthesized tellurite glass sample, and inset is the DSC curves of T-TH0 and T-TH3 tellurite glasses.

Download Full Size | PDF

Figure 2 displayed the Raman spectra of four tellurite glass samples T-TH0, T-TH3, T-THC0 and T-THC3 in the spectral range from 200 to 1000 cm^-1. The scattering peaks of T-TH0 and T-THC0 glass samples without WO₃ component appear at three positions of about 390, 670 and 761 cm^-1. Among which, the 390 cm^-1 scattering peak is caused by the bending vibration of Zn-O bonds from ZnO₄ unit [17,18] together with the symmetric tensile vibration of Te-O-Te or O-Te-O bonds from TeO₄ triangular double pyramid (tbp) unit [19]. The scattering peak at 690 cm^-1 corresponds to the asymmetric stretching vibration of Te-O bonds from TeO₄ tbp unit [20] and the 761 cm^-1 scattering peak arises from the stretching vibration of non-bridged Te and O atoms from TeO_3 + 1 polyhedron and TeO₃ triangular pyramid units [21]. It is seen that the scattering peak at 390 cm^-1 in T-TH3 and T-THC3 tellurite glass samples with WO₃ component becomes stronger and this is attributed to the addition of tensile and flexural vibration of W-O-W bonds related to the octahedral unit of WO₆ [22]. Besides these three scattering peaks, in both glass samples a new scattering peak is found at about 916 cm^-1, and this scattering peak originates from the bending vibration of W = O bond from WO₄ tetrahedron unit together with the tensile vibration of W-O^- bond from WO₆ octahedron unit [16]. Obviously, the vibrational phonon energy of the tellurite glass becomes large after the introduction of WO₃ component, but is still smaller compared with that of the traditional oxide glasses like silicate glass (∼1050 cm^-1) [23], phosphate glass (∼1250 cm^-1) [24] and borate glass (∼1300 cm^-1) [25].

 

Fig. 2. Raman spectra of tellurite glass samples with and without WO₃ component.

Download Full Size | PDF

3.2 Absorption spectrum and Judd-Ofelt analysis

Figure 3 displayed the absorption spectra of tellurite glass samples in the range of 400–2000nm and the samples selected as the representative are Ho³⁺ and Tm³⁺ single-doping (T-Ho and T-Tm), Tm³⁺/Ho³⁺ co-doping (T-TH0 and T-TH3) as well as Tm³⁺/Ho³⁺/Ce³⁺ tri-doping (T-THC3) respectively. The results obtained are similar to those of other glass hosts [26]. All absorption bands exhibited in the spectra originate from the transitions of f-f electrons from the ground states to respective excited levels. The bands observed at wavelengths 470, 685, 795, 1212 and 1665 in the Tm³⁺ single-doping sample are attributed to the absorption transitions from the ground state ³H₆ to excited levels ¹G₄, ³F₃(³F₂),³H₄, ³H₅ and ³F₄ of Tm³⁺ respectively [27]. The bands located at wavelengths 1955, 1150, 645, 540, 485, 458 and 420 nm in the Ho³⁺ single-doping sample originate from the absorption transitions from the ground state ⁵I₈ to excited levels⁵I₇, ⁵I₆, ⁵F₅, ⁵S₂ (⁵F₄), ⁵F₃, ⁵F₁ (⁵G₆) and ³G₅ (⁵G₅) of Ho³⁺ respectively [28]. Because the absorption band of Ce³⁺ (²F_5/2→ ²F_7/2) is positioned at MIR region of about 4650 nm [29], thus the absorption peak of Ce³⁺ does not displayed in the figure. However, owing to the strong absorption induced by the inter-configuration transition of 4f→5d (²F_5/2→⁵D₁) of Ce³⁺ at about 400–500 nm [30], the UV absorption edge in the glass sample with Ce³⁺ is red-shifted to a much longer wavelength. Moreover, it is noted that the shapes and positions of all absorption peaks are almost unchanged in these samples. This observation implies that the three ions are homogeneously distributed in tellurite glass network.

 

Fig. 3. Absorption spectra of tellurite glass samples doped with Tm³⁺, Ho³⁺ and Ce³⁺.

Download Full Size | PDF

Besides the absorption transitions observed, the radiative transition characteristics of doped rare earth ions can be evaluated from the absorption spectrum. Based on the J-O (Judd-Ofelt) theory [31,32], some significant spectral parameters, for instance, the fluorescence branching ratio, radiative transition probability and radiative lifetime for the transition from upper level to lower level can be determined. The detailed description of J-O theory can be found in Refs. [33,34]. First, applying the least mean square fitting algorithm to the theoretical and the experimental oscillator strengths, the three J-O intensity parameters (${\Omega _2}$, ${\Omega _4}$ and ${\Omega _6}$) are determined and the obtained results of Tm³⁺ ions in Tm³⁺ single-doping glass sample (T-Tm) are 5.26, 1.45 and 1.84 × 10⁻²⁰ respectively, while those of Ho³⁺ ions in Ho³⁺ single-doping glass sample (T-Ho) are 5.21, 2.81 and 1.72 × 10⁻²⁰ respectively. The parameter ${\Omega _2}$ mainly reflect the short-range coordination effect [35], because it is associated with the covalent nature of chemical bonds between the ligand anions and rare earth cations, together with the local environment symmetry around rare earth sites. When the value of ${\Omega _2}$ increases, the covalent bond rises in number and the local environment of rare earth ions becomes more asymmetrical. The parameters ${\Omega _4}$ and ${\Omega _6}$ reveal the long-range effect of doped glass [36], wherein the hardness and viscosity of glass matrix are mainly associated with ${\Omega _6}$. In this work, the relatively large ${\Omega _2}$ values of Tm³⁺ and Ho³⁺ ions imply that the local symmetry around rare earth ions is lower while the covalent character of the chemical bonds between the rare earth cations and ligand anions is relatively higher.

Further, using the relevant formulas introduced in Refs. [33,34], the radiative transition probability, fluorescence branching ratio and radiative lifetime can be determined based on the values of three intensity parameters. The spectral parameters calculated for the partial lower-excited levels of Tm³⁺ and Ho³⁺ ions are summarized in Tables 3 and 4 respectively. It is well known that high radiative transition probability is necessary to obtain strong fluorescence emission. For the two NIR band emissions at 1.83 and 2.0 µm studied in this paper, that is, the Tm³⁺:³F₄→³H₆ and Ho³⁺:⁵I₇→⁵I₈ radiative transitions, the radiative probabilities are about 505.11 and 232.34 s^-1 respectively. Among them, the radiative probability of Tm³⁺:³F₄→³H₆ transition is higher than that of the germanate glass (∼398.29 s^-1) [37], gallate glass (∼457.25 s^-1) [38], while that of Ho³⁺:⁵I₇→⁵I₈ transition is higher in comparison to that of germanate glass (∼177.70 s^-1) [39] as well as fluorate glass (∼103.02 s^-1) [40]. This indicates that tellurite glass doped with Tm³⁺ and Ho³⁺ ions in this work has great potential application prospect in the ∼2.0 µm band.

Table 3. The electric- and magnetic-dipole radiative transition probabilities (${A_{\textrm{ed}}}$, ${A_{\textrm{md}}}$), branching ratio ($\beta$) and lifetime (${\tau _{\textrm{rad}}}$) of partial energy levels in Tm³⁺ doped glass sample.

View Table | View all tables in this article

Table 4. The electric- and magnetic-dipole radiative transition probabilities (${A_{\textrm{ed}}}$, ${A_{\textrm{md}}}$), branching ratio ($\beta$) and lifetime (${\tau _{\textrm{rad}}}$) of partial energy levels in Ho³⁺ doped glass sample.

View Table | View all tables in this article

3.3 Fluorescence emission and luminescence mechanism

Figures 4 and 5 displayed the fluorescence emission spectra of Tm³⁺/Ho³⁺ co-doping and Tm³⁺/Ho³⁺/Ce³⁺ tri-doping tellurite glass samples with various amounts of WO₃ from 1200 to 2200 nm and from 620 to 740 nm spectral ranges respectively under the excitation of 808 nm LD. It can be seen that the co-doping and tri-doping tellurite glass samples have the similar luminescence properties. On the one hand, 705 nm, 1.47 µm, 1.83 µm and 2.0 µm four bands are observed in the measured wavelength range, and they originate from Tm³⁺:³F_2,3→³H₆, ³H₄→³F₄,³F₄→³H₆ and Ho³⁺:⁵I₇→⁵I₈ radiative transitions [41–44] respectively. It should be noted that the appearance of fluorescence of Ho³⁺ in the 2.0 µm band is caused by the energy transfer from Tm³⁺ to Ho³⁺ ions which will be discussed later, because Ho³⁺ does not have an absorption level that is compatible with the 808 nm LD. Interestingly, this 2.0 µm band fluorescence of Ho³⁺ ions overlaps the 1.83 µm band fluorescence of Tm³⁺ ions, as a result, a broadband and relatively flat fluorescence emission spectrum between 1600 and 2200 nm is formed. On the other hand, after introducing WO₃ component, the fluorescence intensities of 1.47, 1.83 and 2.0 µm bands are enhanced to the different degrees until the amount of WO₃ reaches about 7.5 mol%, while the intensity at 705 nm band fluorescence exhibits an opposite trend. As discussed in section 3.1, the introduction of WO₃ component brings the increase of vibrational phonon energy of glass network, and such increase of phonon energy is responsible for the luminescence phenomena observed. Thirdly, for both NIR bands of 2.0 µm and 1.83 µm that we are interested in, their fluorescence intensity in the tellurite sample with Tm³⁺/Ho³⁺/Ce³⁺ tri-doping is noticeably stronger compared to that in the sample with Tm³⁺/Ho³⁺ co-doping, and this is attributed to the cross relaxation among Ce³⁺ and Tm³⁺ ions.

 

Fig. 4. Fluorescence spectra in the range of (a) 1200-2200 nm and (b) 620-740 nm in Tm³⁺/Ho³⁺ co-doping tellurite glass samples with different WO₃ amounts under 808 nm LD excitation.

Download Full Size | PDF

 

Fig. 5. Fluorescence spectra in the range of (a) 1200-2200 nm and (b) 620-740 nm in Tm³⁺/Ho³⁺/Ce³⁺ tri-doping tellurite glass samples with different WO₃ amounts under 808 nm LD excitation.

Download Full Size | PDF

The luminescence observed under 808 nm LD excitation can be understood by means of a schematic diagram which includes the interactions and transitions of Ce³⁺, Ho³⁺ and Tm³⁺ ions as displayed in Fig. 6. First, through the ground state absorption (GSA), Tm³⁺ ions at ³H₆ ground state are excited to the pumping level of ³H₄, where a few of them radiate directly to the level ³F₄ accompanying with a faint 1.47 µm band fluorescence due to the inherent small fluorescence branching ratio. The majority of ions at ³H₄ level quickly return to the lower³F₄ and ³H₅ levels via the cross relaxation (CR1, Tm³⁺:³H₄+ Tm³⁺:³H₆ → Tm³⁺:³F₄ + Tm³⁺:³F₄) and multi-phonon relaxation (MPR) processes. A fraction of Tm³⁺ ions at ³H₅ level keep decaying to the ³F₄ level through the MPR process and meanwhile transfer the energy to the Ho³⁺:⁵I₆ level (ET1, Tm³⁺:³H₅+ Ho³⁺:⁵I₈ → Tm³⁺:³H₆+ Ho³⁺:⁵I₆), while the remainder are excited to a higher level ¹G₄ via the excited state absorption (ESA, Tm³⁺:³H₅+ a photon → Tm³⁺:¹G₄). After they return to the level ³F_2,3 by a process of MPR, the radiative transition to ground state ³H₆ produces a fluorescence in 705 nm band. Lastly, at the level ³F₄, the Tm³⁺ ions radiate to the ground state ³H₆ and generate a strong fluorescence in 1.83 µm band. In the Tm³⁺/Ho³⁺ co-doping, energy transfer can easily take place since the energy differences between the certain energy levels of Tm³⁺ and Ho³⁺ ions are small, for example, from Tm³⁺:³H₅ level to Ho³⁺:⁵I₆ level (ET1) and from Tm³⁺:³F₄ level to Ho³⁺:⁵I₇ level (ET2, Tm³⁺:³F₄+ Ho³⁺:⁵I₈ → Tm³⁺:³H₆+ Ho³⁺:⁵I₇). Therefore, at ⁵I₇ level, the Ho³⁺ ions from the indirect excitation of energy transfer ET2 together with those from the MPR process of Ho³⁺ ions at the ⁵I₆ level, radiate to the ground state ⁵I₈ and produce a fluorescence in the 2.0 µm band. In the case of Tm³⁺/Ho³⁺/Ce³⁺ tri-doping, the introduced Ce³⁺ ions can interact with Tm³⁺ ions through the cross relaxations of Tm³⁺:³F_2,3+ Ce³⁺:²F_5/2 → Tm³⁺:³H₄+ Ce³⁺:²F_7/2 (CR2) and Tm³⁺:³H₅+ Ce³⁺:²F_5/2 → Tm³⁺:³F₄+ Ce³⁺:²F_7/2 (CR3). These two CR2 and CR3 processes promote the Tm³⁺ ions return from ³F_2,3 level to ³H₄ level and from ³H₅ level to ³F₄ level respectively, which is conducive to the intensity increases of the 1.47, 1.83 and 2.0 µm three bands, but is negative for the 705 nm band. Further, with the addition of WO₃ component, the increase in phonon energy of glass matrix facilitates the MPR processes from the high level to the low levels of Ho³⁺ and Tm³⁺ ions, especially the energy transfer process from Tm³⁺ to Ho³⁺ ions. For energy transfer ET2, the difference in energy between the Ho³⁺:⁵I₇ level and Tm³⁺:³F₄ level is approximately 900 cm^-1, which nearly corresponds to the maximal phonon energy of glass matrix (∼916 cm^-1). Thus, one phonon is enough to bridge the difference in this phonon-assisted energy transfer process. Just due to the increased MPR rate and energy transfer efficiency, the 1.83 and 2.0 µm band fluorescence intensities increase further with the introduction of WO₃ component.

 

Fig. 6. Energy level diagram of Tm³⁺, Ho³⁺, Ce³⁺ transitions and interactions among them under LD excitation at 808 nm.

Download Full Size | PDF

To see the influence of Ce³⁺ ions and WO₃ component on the luminescence properties of Ho³⁺ and Tm³⁺ ions more clearly, Fig. 7 displayed the fluorescence emission spectra of T-TH0 glass sample without Ce³⁺ ions and WO₃ component, T-TH3 glass sample with 7.5 mol% WO₃ component and without Ce³⁺ ions, T-THC0 glass sample with 0.1 mol% CeO₂ and without WO₃ component, and T-THC3 glass sample with 0.1 mol% CeO₂ and 7.5 mol% WO₃ component. In comparison to glass sample T-TH0, it displays that the introduction of Ce³⁺ ions or WO₃ component both enhances the fluorescence intensities in the 1.47, 1.83 and 2.0 µm three bands but weakens the 705 nm band emission. Importantly, the simultaneous introduction of Ce³⁺ ions and WO₃ component has a more obvious effect. For the broadband emission band of 1600–2200 nm, an increase of about 103% in the glass sample T-THC3 was observed. In addition, this broadband emission band has a full width at half maximum (FWHM) of ∼360 nm, which is better than that of most reports such as Tm³⁺/Ho³⁺ co-doped antimony-silicate glass (356 nm) [45], lanthanum aluminosilicate glass (343 nm) [46], silicate glass (350 nm) [47], silicate-germanate glass (231.5 nm) [48], Tm³⁺/Ho³⁺/Yb³⁺ tri-doped gallo-germanate glass (343 nm) [49] and Tm³⁺/Ho³⁺/Ce³⁺ tri-doped bismuth tellurite glass (216 nm) [50].

 

Fig. 7. Fluorescence spectra in the range of (a) 1200-2200 nm and (b) 620-740 nm in T-TH0, T-TH3, T-THC0 and T-THC3 glass samples under 808 nm LD excitation.

Download Full Size | PDF

The fluorescence transient behavior of 1.47, 1.83 and 2.0 µm three bands in samples with Tm³⁺/Ho³⁺ co-doping and Tm³⁺/Ho³⁺/Ce³⁺ tri-doping was examined. Figures 8 and 9 displayed the measured fluorescence decay curves of corresponding samples. All of the curves deviated from mono-exponential decay law, implying interactions like cross relaxation together with energy transfer exist among the doping rare earth ions. Decay curve is fitted by a double exponential function [51]:

 

Fig. 8. Fluorescence decay curves of (a) Tm³⁺:³H₄→³F₄, (b) Tm³⁺:³F₄→³H₆ and (c) Ho³⁺:⁵I₇→⁵I₈ transitions in Tm³⁺/Ho³⁺ co-doped tellurite glass samples under 808 nm LD excitation, and the solid line is fitting result.

Download Full Size | PDF

 

Fig. 9. Fluorescence decay curves of (a) Tm³⁺:³H₄→³F₄, (b) Tm³⁺:³F₄→³H₆ and (c) Ho³⁺:⁵I₇→⁵I₈ transitions in Tm³⁺/Ho³⁺/Ce³⁺ tri-doped tellurite glass samples under 808 nm LD excitation, and the solid line is fitting result.

Download Full Size | PDF

It is seen that when 7.5 mol% WO₃ component is introduced into samples with Tm³⁺/Ho³⁺ co-doping and Tm³⁺/Ho³⁺/Ce³⁺ tri-doping, the lifetimes of Tm³⁺:³H₄→³F₄,³F₄→³H₆ and Ho³⁺:⁵I₇→⁵I₈ radiative transitions are all increased, which are in good accordance with the fluorescence emission intensities observed. Among them, the lifetime of Tm³⁺:³H₄→³F₄ transition is slightly increased from 0.051 to 0.063 ms in the Tm³⁺/Ho³⁺ co-doping case, that of Tm³⁺:³F₄→³H₆ transition is increased from 1.04 to 1.21 ms, and that of Ho³⁺:⁵I₇→⁵I₈ transition is increased from 1.32 to 1.40 ms. While in the case of Tm³⁺/Ho³⁺/Ce³⁺ tri-doping, the lifetimes of above transitions are increased from 0.064 to 0.076 ms, 1.27 to 1.44 ms and 1.44 to 1.87 ms, respectively. As analyzed above, the addition of WO₃ component brings a raise of vibrational phonon energy in glass network, which enhances the process of energy transfer ET2 as well as the MPR process from the high levels to the low-lying levels, and this is responsible for the increase of radiative transition lifetimes. However, when the amount of WO₃ component in glass continues to increase, the vibrational phonon density at 916 cm^-1 increases accordingly, and the MPR processes of these three NIR emitting levels will become stronger synchronously, which leads to the fluorescence and lifetime of these transitions decrease. Therefore, the amount of WO₃ oxide with high phonon energy introduced into the tellurite glass must be appropriate to ensure that the fluorescence emission tends to the best state. In addition, it is pointed out that the lifetime of the three transitions in samples with Tm³⁺/Ho³⁺/Ce³⁺ tri-doping is larger in comparison to that in the case of Tm³⁺/Ho³⁺ co-doping, and this is attributed to the CR2 and CR3 processes between the Ce³⁺ and Tm³⁺ ions.

From the measured fluorescence lifetime (${\tau _{m}}$) and the calculated spontaneous radiative lifetime (${\tau _{\textrm{rad}}}$) according to the J-O theory, the quantum efficiency (${\eta _{\textrm{rad}}}$) of Tm³⁺:³F₄→³H₆ and Ho³⁺:⁵I₇→⁵I₈ transitions can be estimated by the following [52]:

Table 5 lists the quantum efficiency of Tm³⁺:³F₄→³H₆ (1.83 µm) and Ho³⁺:⁵I₇→⁵I₈ (2.0 µm) transitions including the fluorescence lifetime and spontaneous radiative lifetime of T-TH0 and T-TH3 tellurite glass samples. With the introduction of WO₃ component, the quantum efficiency of Tm³⁺:1.83 µm and Ho³⁺:2.0 µm two bands has been improved to different degrees, which again shows that the introduction of an appropriate amount of WO₃ component has a beneficial effect on these fluorescence emissions. In addition, the quantum efficiency of Tm³⁺:³F₄→³H₆ transition in this work is higher than that of germanate glass (23.9%) [53], but lower than that of fluorotellurite glass (75.93%) [54]. While the quantum efficiency of Ho³⁺:⁵I₇→⁵I₈ transition is higher than that of silica glass (9.38%) [55], but lower than that of low silica aluminosilicate glass (51%) [56]. High quantum efficiency is mainly attributed to the low phonon energy of glass and is conducive to the fluorescence emission.

Table 5. The spontaneous radiative lifetime (${\tau _{\textrm{rad}}}$), fluorescence lifetime (${\tau _{m}}$) and quantum efficiency (${\eta _{\textrm{rad}}}$)

View Table | View all tables in this article

3.4 Multiphonon relaxation and energy transfer

The introduction of an appropriate amount of WO₃ can enhance the NIR fluorescence emissions, which is attributed to the increased MPR rates from the high levels of Tm³⁺ and Ho³⁺ ions to their fluorescence emitting levels together with the enhanced energy transfer from the Tm³⁺:³F₄ level to the Ho³⁺:⁵I₇ level. Hence, it is essential to examine the effect of the addition of WO₃ on the MPR between the relevant levels and the energy transfer from Tm³⁺ to Ho³⁺ ions. The MPR rate from a high level to a low level can be estimated from an empirical formula [8,57]:

The energy transfer among rare earth ions is primarily based on the dipole-dipole interaction, and the microscopic probability from the donor (D) to the acceptor (A) was defined by D.L. Dexter [58] as below:

In the case of existing energy difference between the donor and the acceptor levels, the energy transfer between them requires the assistance of host phonons, and the overlap integral in Eq. (7) was revised as the following [61]:

The energy transfer micro coefficients, the number of phonons involved and their contributions to the total probability calculated for the energy transfer from Tm³⁺:³F₄ level to Ho³⁺:⁵I₇ level (ET2) in the Tm³⁺/Ho³⁺ co-doping tellurite glass samples (T-TH0 and T-TH3) with and without WO₃ component are summarized in Table 6. The absorption cross section of acceptor ions, and the emission cross section and emission sideband of donor ions required in the calculation are shown in Fig. 10. It is concluded from the table that the energy transfer ET2 is a non-resonant interaction process, which is mainly assisted by one phonon, and its contribution ratio in T-TH0 and T-TH3 glass sample has reached about 87% and 92% respectively. The increase of one phonon contribution ratio in glass sample after the introduction of WO₃ component indicates that the increased phonon energy is more suitable for bridging the energy difference between the Tm³⁺:³F₄ level and Ho³⁺:⁵I₇ level, thus enhancing the energy transfer, which is conducive to the 2.0 µm band fluorescence of Ho³⁺ ions.

 

Fig. 10. Emission cross-section of Tm³⁺:³F₄→³H₆ transition and absorption cross-section of Ho³⁺:⁵I₈→⁵I₇ transition in (a) T-TH0 and (b) T-TH3 tellurite glass samples, and the dash line is one and two-phonon emission sidebands of Tm³⁺:³F₄→³H₆ transition.

Download Full Size | PDF

Table 6. Microscopic parameters of energy transfer Tm³⁺:³F₄→Ho³⁺:⁵I₇ in T-TH0 and T-TH3 tellurite glass samples.

View Table | View all tables in this article

4. Conclusions

An enhanced broad 1600–2200 nm band luminescence with a FWHM of ∼360 nm was obtained in tellurite glasses tri-doped with Tm³⁺, Ho³⁺ and Ce³⁺ containing WO₃ excited at 808 nm LD. The spectral overlapping of 1.83 µm band fluorescence of Tm³⁺ and 2.0 µm band fluorescence of Ho³⁺ resulted in this broadband luminescence. The tri-doping of Ce³⁺ with Tm³⁺ and Ho³⁺ accelerated the return of Tm³⁺ ions to the fluorescence level ³F₄ through the cross relaxation among Ce³⁺ and Tm³⁺ ions which is beneficial to the 1.83 µm band, while the increase of glass phonon energy induced by the addition of WO₃ component enhanced the MPR rates as well as the energy transfer from Tm³⁺:³F₄ level to Ho³⁺:⁵I₇ level which is beneficial to the 2.0 µm band, all these led to an improvement of broadband luminescence by about 103% in comparison to the tellurite glass only with Tm³⁺/Ho³⁺ co-doping. The present study proposed an experimental scheme to improve the broadband luminescence, which has significant implications for the further development of optoelectronic devices operated in ∼2.0 µm band.