1. Introduction

Solid state lasers operating at approximately 2.0 μm have been the subject of several spectroscopic investigations for their potential applications in eye-safe laser radar, medical surgery and atmospheric monitoring [1–3]. In this respect, Ho³⁺ has been recognized as suitable laser active ion for its high effective stimulated emission cross section and long-lived ⁵I₇ level related to Ho³⁺: ⁵I₇ → ⁵I₈ transition [4]. With the advent and rapid development of AlGaAs laser diodes (LD), diode-pumped Tm³⁺-Ho³⁺ system has been proposed as a promising way to achieve efficient 2.0 μm laser. In this context, the drawback of Ho³⁺, i.e., lack of convenient bands for diode pumping can be circumvented via pumping Tm³⁺: ³H₄ state followed by resonant energy transfer (ET) from Tm³⁺ to Ho³⁺. Moreover, a pump quantum efficiency of 200% can be expected from the cross relaxation (CR) between two neighboring Tm³⁺ ions: ³H₄ + ³H₆ → ³F₄ + ³F₄ [5]. Indeed, efficient laser operation in the vicinity of 2.0 μm has been demonstrated in Ho³⁺/Tm³⁺ co-doped various fluoride crystals and fluoride fibers [6–8], benefited from low phonon energy, high transparency and good rare-earth (RE) ion solubility [9]. However, the poor chemical and mechanical stability of fluoride fibers has limited their applications [10]. On the other hand, single crystals are often difficult and costly to produce.

Since Wang and Ohwaki reported of greatly enhanced Er³⁺ upconversion emission in transparent oxyfluoride glass ceramic (GC) in 1993 [11], RE doped transparent oxyfluoride GC as a promising optical material, has attracted much attention for the excellent combined properties of high chemical and mechanical stability related to oxide glass and low phonon energy of the precipitated fluoride nanocrystals [12]. It has been reported that the spectroscopic properties of RE ions were greatly enhanced with the incorporation of RE into fluoride nanocrystals during the creaming process [12,13]. In 2001, Samson et al. reported efficient laser operation in Nd³⁺-doped GC fiber and further highlighted the potential of such material for a new range of optical devices [14].

Herein, this paper investigates the spectroscopic and gain properties of 2.0 µm emission in transparent GC containing BaF₂: Ho³⁺, Tm³⁺ nanocrystals to assess its potentiality as 2.0 µm laser host. We have demonstrated intense 2.0 µm fluorescence of Ho³⁺ upon excitation of 808 nm LD. ET processes between Tm³⁺ and Ho³⁺ were analyzed in terms of ET constants $C_{Tm\to Ho}$ and $C_{Ho\to Tm}$calculated by Dexter’s model for dipole-dipole interaction. The results show that transparent GC is a good candidate for 2.0 µm laser.

2. Experimental

Glasses of molar composition (59-x)SiO₂-20ZnO-20BaF₂-1TmF₃-xHoF₃ (denoted as SZBTx) (x = 0, 0.1, 0.3, 0.6, 1.0, and 1.5) were prepared by melting mixtures of raw materials in covered corundum crucibles at 1460 °C for 1 h in air. The melts were then quenched on a preheated stainless steel plate followed by annealing at 480 °C for 2 h. The obtained glasses were polished and heat-treated at 580 °C for 6 h to fabricate transparent GC. Microstructures of GC were characterized by X-ray powder diffractometer (X’ Pert PROX, Cu Kα) and transmission electron microscope (TEM, JEM-2010). Optical absorption spectra measurements were carried out on a Perkin-Elmer Lambda 900/UV/VIS/NIR spectrophotometer. The fluorescence spectra were recorded on a computer-controlled Triax 320 spectrofluorimeter (Jobin-Yvon Corp.) upon excitation of an 808 nm LD. Lifetimes were determined from the first e-folding time of emission intensities in the decay curves recorded with a digital oscilloscope, a mechanical chopper and an InSb detector.

3. Results and discussion

Figure 1 (a) shows the XRD patterns of Ho³⁺/Tm³⁺ co-doped glass and GC. It is clear that the precursor glass is completely amorphous without any diffraction peak. After thermal treatment at 580 °C for 6 h, the XRD pattern exhibits several broad peaks, which can be indexed to the cubic BaF₂ phase (JCPDS Card No. 00-004-0452). Moreover, one can observe that the diffraction peaks move towards the higher angle side as compared with the XRD pattern of standard BaF₂ crystals, indicative of the occurrence of lattice shrinkage in BaF₂ nanocrystals resulted from the replacement of Ba²⁺ with smaller ions (Tm³⁺, Ho³⁺). The lattice parameter of BaF₂ nanocrystals is found to be 6.08 Å, which is smaller than that of the standard BaF₂ crystal (6.20 Å). TEM image and corresponding selected area electron diffraction pattern shown in Fig. 1(b) demonstrate the composite structure of GC with nanocrystals sized 7-19 nm (see Fig. 1(c)) distributing homogenously among glass matrix. The high-resolution TEM image in Fig. 1(d) shows the detailed lattice structure of an individual BaF₂ nanocrystal.

 

Fig. 1 (a). XRD patterns of the precursor glass and GC heat-treated at 580°C for 6 h. (b) TEM bright field image and corresponding selected area electron diffraction pattern (left upper side) of the GC. (c) the size distribution of BaF₂ nanocrystals embedded in glass matrix. (d) High-resolution TEM image of a BaF₂ nanocrystal.

Download Full Size | PDF

Figure 2 (a) illustrates the absorption spectra of GC samples doped with 1 mol.% Tm³⁺ and/or 0.6 mol.% Ho³⁺. Referring to Tm³⁺-doped sample, the absorption spectrum consists of seven absorption bands centered at 354, 467, 657, 680, 790, 1208 and 1669 nm, ascribed to transitions from the ³H₆ ground state to excited states ¹D₂, ¹G₄, ³F₂, ³F₃, ³H₄, ³H₅ and ³F₄, respectively. Moreover, it is noted that the absorption spectrum of Tm³⁺-doped GC is characteristic of two Stark split peaks around 1669 nm due to the partition of Tm³⁺ into BaF₂ nanocrystals, which is similar to the case in GC containing PbF₂ nanocrystals [15]. The absorption spectrum of Ho³⁺ is characterized by thirteen bands at 333, 345, 360, 385, 417, 452, 472, 485, 541, 641, 890, 1169 and 1977 nm, assigned to transitions from ⁵I₈ to ³F₄, ³L₉, ³H₅ + ³H₆, ⁵G₄ + ³K₇, ⁵G₅, ⁵G₆ + ⁵F₁, ⁵F₂ + ³K₈, ⁵F₃, ⁵S₂ + ⁵F₄, ⁵F₅, ⁵I₅, ⁵I₆ and ⁵I₇ levels, respectively. Meanwhile, it is observed that the absorption spectrum of Ho³⁺/Tm³⁺ co-doped sample is simply a superposition of the absorption spectra of Tm³⁺ and Ho³⁺. To investigate the absorption variation of the sample after crystallization, the Judd-Ofelt analysis was applied to the precursor glass and the GC. In general, the Judd-Ofelt intensity parameter Ω₂ is sensitive to asymmetry around RE ions [16] and will decrease with the increase in symmetry between lanthanide ions and the ligand fields. Besides, it is often related to the nature of bonding between RE ions and the ligand anions. For instances, the Ω₂ value decreases with the host changing from oxides to fluorides [17]. From the precursor glass to the GC, the decrease of Ω₂ (in units of 10⁻²⁰ cm²) from 3.33 to 2.79 for Tm³⁺ and 3.65 to 2.23 for Ho³⁺ verifies the partition of RE ions into the BaF₂ nanocrystals. Absorption cross sections corresponding to the Tm³⁺:³H₆ → ³F₄ and Ho³⁺:⁵I₈ → ⁵I₇ transitions were calculated from the measured absorption spectra. Fuchtbauer-Ladenburg formula [18] and McCumber theory [19] were used to calculate the corresponding stimulated emission cross sections of Tm³⁺ and Ho³⁺, respectively. The calculated cross sections for GC are presented in Fig. 2(b). Significant overlap between Tm³⁺ (³F₄ → ³H₆) emission cross section and Ho³⁺ (⁵I₈ → ⁵I₇) absorption cross section suggests the great possibility of ET from Tm³⁺ (³F₄) to Ho³⁺ (⁵I₇).

 

Fig. 2 (a). Absorption spectra of GC samples doped with 1 mol.% Tm³⁺ and/or 0.6 mol.% Ho³⁺. (b) Absorption (solid line) and emission cross sections (dash line) corresponding to the ³H₆ - ³F₄ transition of Tm³⁺ and the ⁵I₈ - ⁵I₇ transition of Ho³⁺ in GC.

Download Full Size | PDF

Figure 3 shows the fluorescence spectra of SZBTx glass samples upon excitation of 808 nm LD. For Tm³⁺ doped sample, the spectrum exhibits two emission bands located at approximately 1.46 and 1.8 μm, attributed to the Tm³⁺: ³H₄ → ³F₄ and ³F₄ → ³H₆ transitions. Obviously, a trace addition of Ho³⁺ in the Tm³⁺ doped sample gives rise to 2.0 µm emission band originated from Ho³⁺: ⁵I₇ → ⁵I₈ transition. Compared with Tm³⁺-doped sample, the emission intensity at 1.8 μm decreases monotonously with increasing Ho³⁺ concentration, whereas 2.0 μm emission intensity increases significantly before it reaches the maximum value when 0.6 mol.% Ho³⁺ is doped. These observations indicate that the 2.0 µm emission of Ho³⁺ can be achieved through efficient ET from Tm³⁺ to Ho³⁺. The principal involved ET mechanisms for Tm³⁺-Ho³⁺ ions system have been proposed, as shown in the inset of Fig. 3. Upon excitation at 808 nm, Tm³⁺: ³H₄ state is excited, followed by CR process between two adjacent Tm³⁺ ions (³H₄, ³H₆ → ³F₄, ³F₄), resulting in effective “two for one” excitation of Tm³⁺: ³F₄ level [20]. After that, Ho³⁺: ⁵I₇ level can be populated via ET from Tm³⁺ to Ho³⁺, i.e., Tm³⁺: ³F₄ + Ho³⁺: ⁵I₈ → Tm³⁺: ³H₆ + Ho³⁺: ⁵I₇. Finally, radiative transition of Ho³⁺: ⁵I₇ → ⁵I₈ occurs, yielding 2.0 µm emission. On the other hand, there might exist a direct route to populate Ho³⁺: ⁵I₇ state, i.e., CR process of Tm³⁺: ³H₄ + Ho³⁺: ⁵I₈ → Tm³⁺: ³F₄ + Ho³⁺: ⁵I₇.

 

Fig. 3 HoF₃ concentration dependence of Tm³⁺ and Ho³⁺ fluorescence intensities in SZBTx glass samples under 808 nm excitation. The inset shows the simplified energy level diagram for Tm³⁺-Ho³⁺ ions system.

Download Full Size | PDF

To examine the suitability of GC as potential 2.0 µm laser material, SZBT0.6 was selected to precede controlled crystallization. The fluorescence spectra of the precursor glass and GC are presented in a comparable way in Fig. 4(a) . Obviously, the emission spectrum of GC exhibits Stark split emission peaks around 2.0 µm due to the partition of Ho³⁺ ions into BaF₂ nanocrystals. Moreover, integrated emission intensity of Ho³⁺ in GC increases by a factor of about 6 compared with that of the precursor glass. This can be ascribed to the enhanced ET efficiency from Tm³⁺ (³F₄) to Ho³⁺ (⁵I₇), benefited from the concentration of RE ions in BaF₂ nanocrystals. In this case, the distances between RE ions would be shorter than those of the ions distributed in the precursor glass, resulting in the efficient CR and ET processes of Tm³⁺-Tm³⁺ and Tm³⁺-Ho³⁺ pairs. Meanwhile, the multiphonon relaxation of Tm³⁺: ³H₄ → ³H₅, which competes with CR process (³H₄, ³H₆ → ³F₄, ³F₄), can be limited due to the low phonon energy environment of BaF₂ nanocrystals. Assuming that electrons in Ho³⁺ are either in the ⁵I₇ state or in the ⁵I₈ state, the gain coefficient $G\left(\lambda \right)$of Ho³⁺ near the 2.0 µm wavelength region can be determined by the following equation [21]:

 

Fig. 4 (a). Fluorescence spectra of glass and GC co-doped with 0.6 mol.% Ho³⁺ and 1 mol.% Tm³⁺ under 808 nm excitation. (b) Calculated gain coefficients corresponding to the ⁵I₈ - ⁵I₇ transition of Ho³⁺ for the GC sample.

Download Full Size | PDF

Figure 4(b) shows the calculated gain coefficients versus wavelengths of ⁵I₇ - ⁵I₈ transition in Ho³⁺/Tm³⁺ co-doped GC by assuming a set of p values ranging from 0 to 1. It is noted that the laser performance wavelengths of the maximum gain coefficient shift to longer wavelengths with decreasing the value of p. This behavior is a typical characteristic of the quasi-three level system. To estimate the extent of ET between Tm³⁺ and Ho³⁺, ET constant$C_{D-A}$and critical radius$R_0$are calculated by Dexter theory [22]. In the case of dipole-dipole interaction, they can be estimated by using the following equations:

4. Conclusion

In summary, a transparent GC containing BaF₂:Ho³⁺,Tm³⁺ nanocrystals was prepared. Thanks to the incorporation of RE ions into the precipitated BaF₂ nanocrystals, efficient 2.0 µm fluorescence ascribed to Ho³⁺: ⁵I₇ → ⁵I₈ transition has been demonstrated upon excitation of 808 nm LD. A large ratio of forward ET constant to that of reverse process further revealed that ET from Tm³⁺ (³F₄) to Ho³⁺ (⁵I₇) is quite efficient in GC, resulted from the reduced ionic distances of Tm³⁺-Tm³⁺ and Tm³⁺-Ho³⁺ pairs and limited multiphonon relaxation of Tm³⁺: ³H₄ → ³H₅ in BaF₂ nanocrystals. The results have demonstrated the promising potential of transparent oxyfluoride GC as a host for efficient Ho³⁺ 2.0 µm laser.