Efficient ∼2 μm emission and energy transfer mechanism of Ho³⁺ doped barium gallium germanate glass sensitized by Tm³⁺ ions

Abstract

Optical absorption and emission properties of Ho³⁺ ions in barium–gallium-germanate glass are investigated. A significant result is that the emission intensity of Ho³⁺:⁵I₇→⁵I₈ excited by 808 nm through the energy level Tm³⁺:³H₄ is stronger than that of the direct excitation of Ho³⁺. The emission cross section of Ho³⁺ is calculated to be 9.02×10⁻²¹ cm² at the wavelength of 1.994 μm. Comparatively, the larger emission cross section and efficient energy transfer process reveal that the prepared glass could be a kind of promising material for laser emission at 2 μm.

1 Introduction

Solid-state lasers emitting in the 2 μm eye-safe region are promising candidates for various applications in laser medicine surgery, monitoring of atmospheric pollutants, and high-resolution spectroscopy of low pressure gases, etc. [1–4]. Ever since Johnson has reported the first laser action at 2.0 μm from Ho³⁺ doped CaWO₄ crystal in 1962 [5], an intense level of research has been focused on the spectroscopic and lasing properties of Ho³⁺ doped various crystals and glasses [6, 7]. Among the rare-earth ions, Ho³⁺ ion is one of the most important active ions applied to the infrared emission, because of its favorable energy level structure. The Ho³⁺ ion has a relatively long-lived ⁵I₇ level which can act as a good population reservoir for infrared luminescence process. It is reported that Ho³⁺ doped fibers can be pumped by semiconductor disk laser around 1.16 μm and the passively mode-locked fiber laser has been demonstrated [8, 9]. Ho³⁺ can also be pumped by sensitizers such as Tm³⁺, Er³⁺ or Yb³⁺ ions. Ho³⁺ doped fiber laser has been achieved when pumped by an Yb³⁺ doped fiber laser at 1147 nm [10]. In our previous studies, the spectroscopic properties of Ho³⁺ sensitized by Yb³⁺, and co-doped with Er³⁺/Tm³⁺ have been studied [11, 12]. The Yb³⁺ ion has large absorption coefficient and emission cross section at 980 nm, but the energy utilization efficiency is relatively low. These studies have shown that the energy transfer microparameter from Yb³⁺ to Ho³⁺ is much lower than that of Tm³⁺/Er³⁺ to Ho³⁺. What’s more, strong upconversion luminescence band be observed in the Yb³⁺/Ho³⁺ codoped glasses, so Tm³⁺/Er³⁺ ions have become the main choice as sensitizers of Ho³⁺.

In this research, we would like to analyze the mechanism of Ho³⁺ fluorescence and Tm³⁺ singly sensitized Ho³⁺ in glass host. The ratio of number of photos emitted by Tm³⁺ ions to the number of photons absorbed, namely, quantum efficiency, can reach as high as 200 % (see Fig. 1). After the ground state absorption (GSA) of one pump photon, two ground level Tm³⁺ ions can be excited to the upper laser level through a cross relaxation (CR) process ³H₄,³H₆→³F₄,³F₄. The pumped energy of the excited level of Tm³⁺:³F₄ transfers to the level of Ho³⁺:⁵I₇ through a resonant or nonresonant energy transfer process [13]. The nonresonant process can be assisted by local phonons and energy difference must be given to the lattice vibrations. Germanate glass has been chosen as the matrix material for its good thermal stability and high transparency in a wide wavelength range [14, 15]. Compared with silicate and fluorophosphates glass, its low phonon energy and broad transmission range allow the observation of laser emission from rare earth ions in a wide optical range. There are quite few studies on infrared emission around ∼2 μm from Ho³⁺ singly or Tm³⁺ codoped of germanate glasses. The present paper mainly aims to report on the studies of ∼2 μm emission characteristics of Ho³⁺ ions both by direct excitation and by sensitized excitation through energy transfer from Tm³⁺ ions in the codoped barium gallium germanate glasses. Additionally, a systematic characterization of spectroscopic properties of Ho³⁺ ions in the present glass by applying the Judd–Ofelt theory has been dealt along with the physical and structural analysis.

Fig. 1

Energy level diagram of Tm³⁺ and Ho³⁺ ions, indicating the transition of cross relaxation, infrared emission, and the quantum efficiency

2 Experiments

The Ho³⁺/Tm³⁺ codoped, Ho³⁺ singly doped and undoped barium gallium germanate having chemical composition (mol %) 65GeO₂-10Ga₂O₃-20(BaO+BaF₂)-5La₂O₃-xTm₂O₃-yHo₂O₃, (x=0, 2 and y=0, 0.5) were prepared by traditional melt-quenching method with high purity reagent as raw materials. Before melting the powders in a SiC-resistance electric furnace, the molar masses were weighted, mixed and dried. The mixtures were stirred 1 h with a platinum rod and then homogenized at 1400 °C in a platinum crucible in furnace. The homogeneous melt was poured onto a preheated brass mould and annealed for several hours around 600 °C in a muffle furnace, and then allowed to cool slowly inside the furnace by turning the power supply off. The two-face polished sample has plate shape with 15×15×1 mm³ for optical measurements.

The refractive index of the host glass was measured by the prism minimum deviation method at four wavelengths, 656, 588, 486, and 1053 nm, and they are 1.7292, 1.7348, 1.7473, and 1.7195, respectively. The resolution of the instrument is ±0.5×10⁻⁴. The standard deviation in refractive index at different points of the same glass is around ±1×10⁻⁴. The refractive index dispersion curve was calculated by Cauchy’s formula n(λ)=a+b/λ ²+c/λ ⁴, where a, b and c are found to be 1.7061, 1.0222×10⁴ and 1.1171×10⁸ respectively. The density of the glass was measured by Archimedes method using distilled water as buoyancy liquid at room temperature. The analysis of Ho³⁺ ions absorption was carried out using a Perkin-Elmer-Lambda 900UV/VIS/NIR spectrophotometer in the range of 350–2200 nm. The near infrared emission spectra and excitation spectra were recorded on a Traix 320 type spectrometer (Jobin-Yvon Co., France). A Xenon flash lamp was employed as an excitation source. All the measurements were performed at room temperature.

3 Results and discussion

The Raman spectra of Ho³⁺/Tm³⁺ codoped and Ho³⁺ singly doped barium gallium germanate are shown in Fig. 2. For both glasses, the spectra comprise identical nature suggesting that the incorporation of Tm³⁺ in glass network does not bring significant change in the glass structure. The spectra revealed a broad band with superimposition of two reflectance bands centered at 846 and 756 cm⁻¹ which are generally assigned to the stretching mode contributions not only from Ge–O structure units but also from Ga–O units in the environment of bridging and nonbridging oxygen [16]. The frequencies higher than 700 cm⁻¹ are generally assigned to localized vibrational modes within the tetrahedral units. Intermediate frequency region are dominated by a complicated envelope of overlapping peaks e (487 cm⁻¹), f (546 cm⁻¹) and g (598 cm⁻¹), which have been assigned to various vibrational modes involving Ge–O–Ge and Ga–O–Ga bending [17]. The Raman peaks at lower wave number a (126 cm⁻¹), b (156 cm⁻¹), c (204 cm⁻¹) and d (284 cm⁻¹) are due to the vibration of network modifying cations in large interstitial sites or to acoustic modes involving motions of the tetrahedral cations in large clusters.

Fig. 2

Raman spectra of Ho³⁺/Tm³⁺ codoped and Ho³⁺ singly doped barium gallium germanate glasses. The plots show the measured spectrum (black line), Gaussian fitting (red line) of the measured data and the sum of the peaks. The dash line stands for the Gauss fit-peaks

The thermal parameters are based on characteristic temperature including the glass transition T _g , onset of devitrification T _x , the peak crystallization peak T _p temperatures and the melting temperature T _m . Dietzel introduced the glass criterion [18], ΔT=T _x −T _g , which is often an important parameter for evaluating the glass forming ability. Large ΔT means strong inhibition to processes of nucleation and crystallization [19]. The glass formation factor of the materials is given by parameter k _gl =(T _x −T _g )/(T _m −T _g ) and is more suitable for estimating the glass thermal stability than ΔT. The higher values of the criterion parameters reflect greater thermal stability of the glass. The measured densities of Ho³⁺/Tm³⁺ codoped and Ho³⁺ singly doped barium gallium germanate glasses along with the average molecular weight of the respective glasses have been used to estimate the rare earth ion concentration. Table 1 shows the results of characteristic temperatures, rare earth concentrations and densities of the samples. It is clear from the data that T _g , T _x , T _p , along with the thermal stability factor ΔT and k _gl increase with thulium addition. The increase in glass stability with thulium addition is due to its conditional forming ability.

Table 1 Characteristic temperatures, rare earth concentrations and densities of the Ho³⁺/Tm³⁺ codoped and Ho³⁺ singly doped barium gallium germanate glasses

The room temperature absorption spectra of Ho³⁺/Tm³⁺ codoped and Ho³⁺ singly doped barium gallium germanate glasses have been presented in Fig. 3. The Ho³⁺ doped glass exhibits a number of distinct absorption bands in the Vis–IR range at 361, 418, 448, 486, 537, 642, 1151 and 1950 nm for Ho³⁺ ions and at 685, 790, 1209 and 1648 nm for Tm³⁺ ions. It is observed that the intensity of the Ho³⁺:⁵I₈→⁵G₆ transition is highest among the detected peaks of the Ho³⁺ singly doped sample. This is due to its hypersensitive nature in 4f ¹⁰ electronic configuration of Ho³⁺ ions with invariably obeying ΔS=0, ΔJ=2 and ΔL=2 spectral rule possessing strong dependence on the ligand field environment surrounding the rare earth ions [20]. In general, magnitude of the hypersensitivity of a transition mainly linked to the ligand covalency, crystal field symmetry and it is correlated strongly with the Ω ₂ intensity parameter [5]. It also can be seen from Fig. 3 that the absorption band of Tm³⁺:³F₄ is close to that of Ho³⁺:⁵I₇ and the energy is slightly higher than that of Ho³⁺:⁵I₈→⁵I₇ transition. Efficient energy transfer can be expected from Tm³⁺:³F₄ to Ho³⁺:⁵I₇, as shown in Fig. 1. Meanwhile, Tm³⁺ has strong absorption band ³H₄ around 800 nm which is available using commercial laser diode.

Fig. 3

Absorption spectra of Ho³⁺/Tm³⁺ codoped and Ho³⁺ singly doped barium gallium germanate glasses. The symbols H and T stand for Ho³⁺ and Tm³⁺, respectively

Oscillator strengths obtained from the absorption spectra are used as experimental input for Judd–Ofelt intensity analysis. The Judd–Ofelt model describes the perturbation of the 4f ⁿ states by opposite parity f ⁿ⁻¹ nl (where nl is mainly 5d) on the basis of intensity parameters Ω _λ (λ=2,4,6) [21, 22]. The model assumes the states of the perturbing 4f ⁿ⁻¹ nl configurations to be degenerate and the energy difference between these perturbing states and all the states of the 4f ⁿ configuration to be equal. The model further assumes an isotropic medium, equal frequencies of the crystal field transitions within a |SLJ〉→|S′L′J′〉 multiplet transition, and equal thermal population of the crystal field levels within the |SLJ〉 initial state. These conditions are typically sufficiently well satisfied for rare earth doped glasses at room temperature. By the absorption spectrum measured above, the electric-dipole-included oscillator strength, f _ed, of a transition |SLJ〉→|S′L′J′〉 can be calculated using the expression

(1)

where m is the mass of the electron, c is the speed of the light in vacuum, n is the refractive index of the sample and υ is the transition frequency. The term |〈U ^λ〉|² is the reduced matrix element of the tensor operators; generally, it can be considered independent on the host materials, and the values used in this work are quoted from Ref. [23]. The magnetic-dipole induced isotropic oscillator strength, f _md, which is nonzero for transitions with ΔJ=0,±1, is given by

$$ f_{\mathrm{md}} = \frac{8\pi^{2}m\upsilon n}{3he^{2}(2J + 1)}\vert \mu_{B} \vert ^{2}\bigl \vert \bigl\langle SLJ\Vert L + 2S \Vert S'L'J' \bigr\rangle\bigr \vert ^{2} $$

(2)

where e is the electron charge and μ _B is the Bohr magneton. The intensity parameters Ω _λ (λ=2,4,6) are obtained from a least square fitting of calculated total oscillator strength f _theor=f _ed+f _md (Eqs. (1) and (2)) to experimental oscillator strengths f _exp. With the addition of glass modifiers such as rare-earth ions, linkages in the [GeO₄] network of germanate glass are broken. Consequently, there are bridging oxygen (BO) atoms and nonbridging oxygen (NBO) ions, the latter having a formal charge of −1 and being the only negatively charged species in the glass. In effect, the germanate glass network becomes a polyanion with net negative charge primarily localized at NBO’s. Charge balance is achieved with the presence of the constituent cations in interstitial sites. Depending on the structure of the network and on the size, charge, and field strengths of these cations, they will form associates with network oxygens with varying degrees of ionic/covalent character.

The Judd–Ofelt parameters Ω _λ (λ=2,4,6) are important for the investigation of the local structure and bonding in the vicinity of rare earth ions. The Ω ₂ parameter is proposed to be a measure of the degree of covalency of the chemical bonds between rare earth ions and its nearest neighbor atoms [24]. Ω ₆ reflects the bulk properties of the host such as rigidity and viscosity [24]. Previous research has also reported that the 2 μm emission transition of Ho³⁺:⁵I₇→⁵I₈ is mainly affected by Ω ₆ intensity parameter due to its larger |〈U ⁶〉|² matrix element and as the radiative spontaneous emission probability is proportionately dependent on Ω ₆ [25]. The Ω ₂,Ω ₄ and Ω ₆ values are calculated to be 7.83×10⁻²⁰ cm²,6.37×10⁻²⁰ cm² and 2.05×10⁻²⁰ cm², respectively. These are larger than those of Ga₂S₃–La₂S₃ glass, whose Ω ₂,Ω ₄ and Ω ₆ values are 6.9×10⁻²⁰ cm², 5.5×10⁻²⁰ cm² and 1.07×10⁻²⁰ cm², respectively [26]. These are also larger than those of ZBLAN glass, whose Ω ₂,Ω ₄ and Ω ₆ values are 2.3×10⁻²⁰ cm²,2.3×10⁻²⁰ cm² and 1.7×10⁻²⁰ cm², respectively [27]. Large Ω ₂ values are probably caused by relatively high covalency of the chemical bond between rare earth and oxygen ions. The Ω ₆ parameter is proposed to increase with increasing rigidity of the medium, i.e., the mean force constant of the Ho³⁺–oxygen/sulfide/fluoride bonds. It can be seen that the Ω ₆ values are higher for germanate than for sulfide and fluoride glasses. The Judd–Ofelt analysis therefore indicates that germanate glass has a higher covalency and a higher force constant of Ho³⁺-oxygen bonds than those of sulfide and fluoride glasses.

Figure 4 shows the near-infrared emission spectra of Ho³⁺/Tm³⁺ codoped germanate glass when excited by 808 nm through ³H₄ excited state of sensitizer Tm³⁺ ions and by direct Ho³⁺ ions excitation through ⁵F₅ level with 644 nm wavelength. These respective excitation wavelength have been chosen from the recorded excitation spectrum of co-doped sample by monitoring 2.0 μm emission of Ho³⁺ ions as seen inset of Fig. 4. The emission peaks detected at 1.8 and 2.016 μm are assigned to the transition of Tm³⁺:³F₄→³H₆ and Ho³⁺:⁵I₇→⁵I₈, respectively. The intensity of the excitation sources keeps the same, so the results are comparable. The interesting observation is that the emission intensity of Ho³⁺:⁵I₇→⁵I₈ had becomes stronger when excited by 808 nm through the energy level Tm³⁺:³H₄ than that of direct excitation of Ho³⁺. This is because of high absorption cross section of Tm³⁺ accompanied with efficient sensitized energy transfer from Tm³⁺ to Ho³⁺ ions.

Fig. 4

Fluorescence spectra of Ho³⁺/Tm³⁺ codoped germanate glass, the excited source of the red and black lines are 808 nm and 644 nm, respectively. The inset shows the excitation spectra of Ho³⁺/Tm³⁺ codoped and Ho³⁺ singly doped germanate glasses for λ _emi =2.0 μm

Figure 5 represents the emission spectra of Ho³⁺ doped germanate glass when excited by a xenon lamp at the wavelength of 658 nm. The spectra depict a broad emission band due to Ho³⁺:⁵I₇→⁵I₈ transition with several stark components ranging from 1850 to 2150 nm. The quality of the curve fit is justified by the best superimposition of the fitted curve over the measured emission peak. The deconvolution has displayed four major transitions between energy level stark components as represented in the figure peaking at 1910, 1954, 2008 and 2070 nm. Upon critical observation of emission peak, it can be seen that the relative intensity of 2070 component is higher than that of 2008 component in the glass indicating there is slight energy re-absorption at the considered dopant ion concentration.

Fig. 5

Deconvolution of emission spectra of Ho³⁺ doped germanate glass using symmetric Gaussian functions

The stimulated emission cross section σ _e can be calculated by using the Fuchbauer–Ladenbury expression [28, 29]:

$$ \sigma_{e} = \frac{A_{\mathrm{rad}}\lambda_{p}^{4}}{8\pi cn^{2}\Delta \lambda_{\mathrm{eff}}} $$

(3)

where λ _p is the fluorescence peak wavelength, Δλ _eff is the effective band width, n is refractive index (1.73), c is velocity of light and A _rad is the radiative spontaneous emission probability (165 s⁻¹). The emission cross section of Ho³⁺ is calculated to be 9.02×10⁻²¹ cm² at the wavelength of 1.994 μm. The value is larger than that of fluoride glass (5.3×10⁻²¹ cm²) [30] but smaller than that of tellurite glass (10.0×10⁻²¹ cm²) [31]. This may properly be due to the refractive index of the host glass, the refractive index of the germanate (1.73) is larger than that of the fluoride glass (∼1.4) [32] while lower than that of tellurite glass (∼2) [31]. It is worthwhile to notice that the value of effective band width is found to be 170 nm, which is highly useful for laser tuning.

Figure 6(a) shows the emission spectra (σ _e ) of Tm³⁺:³F₄→³H₆ and absorption spectra (σ _a ) of Ho³⁺:⁵I₈→⁵I₆, it can be seen that the overlap between the σ _e and σ _a is quite large, so efficient energy transfer can be expected from Tm³⁺ to Ho³⁺. According to Dexter’s model, the extent of energy transfer depends on the spectral overlap of donor’s emission and acceptor’s absorption [33]. In the case of resonant energy transfer between Tm³⁺ and Ho³⁺, the model predicts a linear relationship between the transfer probability (P _ET) and the spectral overlap between donor emission and acceptor absorption bands:

$$ P_{\mathrm{ET}} \propto\int\frac{f_{D}(E)f_{A}(E)}{E^{2}}dE $$

(4)

where f _A (E) and f _D (E) are the line-shape functions of the acceptor absorption and donor emission cross sections, respectively. The energy gap between Tm³⁺ and Ho³⁺ is about 712 cm⁻¹, as the phonon energy of the germanate glasses is 846 cm⁻¹, about one or less phonon is required to bridge the energy gap. Previous studies have shown that the direct transfer is found to be a quasiresonant process with nonphonon 79.5 %, having a participation of one phonon 20.5 % [12, 34]. The efficient energy transfer requires the participation of one phonon [13], the Dexter model can be generalized to the nonresonant phonon assisted energy transfer case by taking into account the energy of the phonon involved (E _PH) as well as the phonon density [35]. Thus, the transfer probability can be estimated from the phonon modified spectral overlap integral I(E _PH). If we assume that only one phonon is involved,

(5)

where k _B and T are Boltzmann constant and absolute temperature, respectively. By using the cross section of Fig. 6(a) we have calculated the I(E _PH) function at room temperature for phonon energy up to 1000 cm⁻¹, as shown in Fig. 6(b). It can be seen that the phonon assisted overlap integral increases with an increase in phonon energy and it reaches a maximum for phonons with energy of about 400 cm⁻¹. For further increase in the phonon energy, the energy transfer probability decreases and diminishes. Though the lower phonon energy matrices, such as tellurite and fluoride glass is suitable to promote the energy transfer from Tm³⁺ to Ho³⁺ with one phonon, but it also possess higher nonradiative relaxation of ⁵I₇ level of Ho³⁺ leading to less probability of 2 μm emission.

Fig. 6

(a) Tm³⁺ emission cross section and Ho³⁺ absorption cross section. (b) Phonon assisted overlap I(E _PH) as a function of phonon energy

4 Conclusions

A detailed spectroscopic analysis of Ho³⁺ ions in barium-gallium-germanate glass has been investigated. The 2 μm emission characteristics of Ho³⁺ both by direct excitation and through Tm³⁺ sensitization in barium gallium germanate glass have been reported. The emission intensity of Ho³⁺:⁵I₇→⁵I₈ is much stronger when excited by 808 nm through the energy level Tm³⁺:³H₄ than that of the direct excitation of Ho³⁺. The Judd–Ofelt theory and Fuchbauer–Ladenbury formula have been used to calculate the intensity parameters and emission cross sections. The Judd–Ofelt analysis indicates that germanate glass has a higher covalency and a higher force constant of Ho³⁺-oxygen bonds than those of sulfide and fluoride glasses. The emission cross section of Ho³⁺ is calculated to be 9.02×10⁻²¹ cm² at the wavelength of 1.994 μm. Dexter’s model and multiphonon relaxation theory have been used to analyze the energy transfer process between Tm³⁺ and Ho³⁺ ions on the energy gap ΔE. A comparative assessment of parameters of the emission cross section and energy transfer process suggests that the germanate glass could be identified as a promising optical material at 2 μm.