The spectral properties of Pr3+ doped BaInGaGeSe chalcogenide glasses are reported. Absorption spectra, emission spectra, and lifetimes of the lower lying manifolds have been measured. Radiative transition rates are calculated and compared with measured experimental lifetimes. The strong mid-IR emission and spectral properties of this glass make this glass a strong candidate for lasers, amplifiers, and high brightness sources in the mid-IR.
©1997 Optical Society of America
Pr3+ has a rich spectrum of transitions in the IR wavelength region from 1 to 7 μm (Figure 1). The large number of bands in the 3–5 μm offers the promise of lasers, amplifiers, and high brightness sources for remote sensing, countermeasures, and spectroscopic applications. Unfortunately, many of these transitions are inactive or very inefficient in conventional oxide and fluoride glass and crystalline hosts due to the high multiphonon quenching of these materials. While several of these transitions have been observed in low phonon crystalline hosts [1–4], these hosts are hygroscopic and not mechanically durable.
In this paper, we report on the characterization of the mid-IR transitions of Pr3+ in Barium Indium Gallium Selenide glass (BIGGSe) chalcogenide glass. The low phonon chalcogenide glasses are nonhygroscopic and offer the advantages of chemical and mechanical durability. The solubility of the rare-earth ions is high in these glasses. Furthermore, fabrication of these glasses into fibers may be possible which is useful for fabricating fiber amplifiers and fiber lasers.
2. Sample fabrication
Bulk samples of Pr doped selenide glasses were prepared from mixtures of elemental praseodymium, barium, indium, germanium, gallium, and selenium. The glass composition was (BaSe)32.5(In2Se3)6.25(Ga2Se3)6.25(GeSe2)55.0 with Pr additions of up to 10 wt. % to the base composition. The glass components were placed in carbon crucibles which were then sealed in quartz ampoules under vacuum. The mixtures were gradually heated to 900 °C and air quenched and then annealed at 380 °C.
Incorporation of the rare earth into the glass matrix was confirmed by lack of any residue in the melt ampoules and absorption data. The glasses appeared homogenous with no segregation of any component in the glasses. X-ray diffraction showed no signs of crystallization in the glasses up to 10 wt.% concentration of Pr3+, confirming the solubility of the rare earth dopant in these glasses. The glass transition temperature, Tg, of the doped BIGGSe glasses is 385 °C while the crystallization temperature, Tx, is 525 °C. Thermal stability as measured by the difference between crystallization temperature and glass transition temperature, (Tx-Tg), is greater than 100 °C, indicating that fiberization of these glasses is promising.
For measurement of the optical properties, sections of 0.1 wt.% , 0.2 wt.% and 2 wt.% bulk glass samples of Pr doped BIGGSe glass were cut and polished. Pr ion concentration in these samples were ~ 2 × 10-19 cm-3, 4 × 10-19 cm-3, and 4 × 10-20 cm-3, respectively. Samples were typically 1 cm long for absorption and 1 mm thick for fluorescence measurements.
Absorption spectra of the Pr doped BIGGSe samples was taken a Cary 5 spectrometer in the wavelength region of 1-2.5 μm. The spectra for the 0.1 wt.% sample is shown in Figure 2 along with spectra where the Urbach edge and weak absorption tail (WAT)  have been subtracted out to show only the rare earth contribution to the absorption. Strong lines at 1.0 μm, 1.5 μm, 1.6 μm, 2.0 μm, and 2.3 μm corresponding to transitions from the ground state, 3H4, to the 1G4, 3F4, 3F3, 3F2 and 3H6 states are visible. These strong absorption lines are suitable for populating the lower lying manifolds of Pr3+.
We note that the 3F4 and 3F3 absorption lines and the 3F2 and 3H6 absorption lines overlap strongly. For the crystalline host material, LaCl3, it has been shown that the 3F4 and 3F2 states rapidly thermalize with the 3F3 and 3H6 states , respectively, resulting in a coupling of these levels. Measuring the peak of these lines, we find that the line peaks of the 3F2, 3H6 and 3F4, 3F3 are separated by 600 cm-1 and 450 cm-1 respectively. From this energy separation of the absorption peaks and Boltzmann statistics, we estimate that 95% of the excited state population will reside in the 3H6 state of the coupled 3H6, 3F2 levels while 89% will reside in the 3F3 state of the coupled 3F3, 3F4 levels at room temperature. Since this population distribution is similar to that found in LaCl3, this assumption is justified. Consequently, we will indicate the thermally coupled levels by (3F3,3F4) and (3H6,3F2) in the text.
Room temperature fluorescence spectra were obtained in the IR by pumping the samples at 1.064 μm with a Coherent CW Nd:YAG laser. The emission from the samples was passed through a Jarrell-Ash ¼ m monochromator and detected with an InSb detector. Spectra in the range of 1.2 μm to the InSb cutoff at ~5.5 μm were recorded in segments with the appropriate grating for the wavelength range. Appropriate filters were placed in front of the InSb detector to remove second order lines from the spectra. Spectra is shown in Figure 3 for a 0.2 wt.% sample and is uncorrected for system response. Note, due to a grating change, the peak heights of the spectra shown in red are not relative to the spectra shown in blue in Figure 3.
Fluorescence centered at 1.35 μm is associated with the 1G4 ⇒ 3H5 transition and is quite smooth over this band. Fluorescence at 1.6 μm is associated with the (3F3,3F4) ⇒ 3H4 transition. Emission between 1.7 μm and 2.6 μm is associated with the 1G4 ⇒ (3H6,3F2), (3F3,3F4) ⇒ 3H5, and (3H6,3F2) ⇒ 3H4. Emission in the 3–5 μm band is associated with the 1G4 ⇒ (3F4,3F3), 3F4 ⇒ 3F2, (3F4,3F3) ⇒ 3H6, (3H6,3F2) ⇒ 3H5, and 3H5 ⇒ 3H6 transitions. Since emission from these transitions overlap, it is not possible to separate the emission under 1.064 μm pumping where all the initial states are populated. Dips in the fluorescence spectra in this region can be attributed to CO2 absorption and H-Se and Ge-H impurity absorption in the glass.
The fluorescence decay at 1.3 μm was measured by pumping a 0.1 wt.% Pr sample at 1.064 μm with a Q-switched Nd:YAG laser. Decay at 1.6 μm and in the mid-IR was measured by pumping at 1.55 μm with a pulsed IR OPO driven by a Q-switched Nd:YAG laser. In both cases, pulse widths were less than 10 ns. Fluorescence was selected by the appropriate use of band pass filters. The fluorescence decay traces are shown in Figures 4.
Fluorescence lifetime of the 1G4 level was measured by examining the 1.3 μm emission under 1.0 μm pumping. The decay is exponential with a lifetime of 180 ± 10 ms. Decay of the (3F3,3F4) coupled levels was measured by observing the emission at 1.6 μm under 1.55 μm pumping. The decay is again exponential with a lifetime of 100 ± 10 μs.
The fluorescence lifetimes of the 3F2, 3H6, and 3H5 levels were more difficult to determine due to the overlapping emission lines. For the (3H6,3F2) level, samples were pumped at 1.6 μm, and the decay of the emission in the 2.0–2.4 μm band, shown in blue in Figure 3, was observed. Under this pumping scheme and assuming no upconversion to populate the 1G4 level, emission in this band can be attributed to the (3F3,3F4) ⇒ 3H5, and (3H6,3F2) ⇒ 3H4 transitions. Since the (3F3,3F4) ⇒ 3H5 decays much faster than the measured 2.0–2.4 μm emission, the tail of this emission must be pure (3H6,3F2) ⇒ 3H4 transition, barring any refeeding of the (3F3,3F4) ⇒ 3H5 due to upconversion. The emission tail is measured to be 290 ± 20 μs. Similarly, emission at 4.5–5 μm under 1.6 μm pumping, shown in red in Figure 3, is attributed to the 3F4 ⇒ 3F2, (3F4,3F3) ⇒ 3H6, (3H6,3F2) ⇒ 3H5 and 3H5 ⇒ 3H4 transitions. Again, the 3F4 ⇒ 3F2, (3F4,3F3) ⇒ 3H6, and (3H6,3F2) ⇒ 3H5 transition decays faster than the observed 4.5–5 μm emission. Consequently, the tail of this emission is pure 3H5 emission. The lifetime of this tail is 2.5 ± 0.9 μs.
Fluorescence lifetimes were also measured for a 2 wt.% sample. This sample was found to highly concentration quenched with lifetimes of 190 ± 80 μs, 70 ± 30 μs, and <20 μs for the 3H5, (3H6,3F2), and (3F4,3F3) levels, respectively. Emission from the 1G4 level could not be detected.
4. Judd-Ofelt Analysis
A Judd-Ofelt analysis [6,7] was performed on the 0.1 wt.% Pr doped BIGGSe composition to determine radiative rates, branching ratios, and cross-sections. The electric dipole line strengths, , were calculated from absorption spectra by
where k(λ) is the absorption coefficient at wavelength λ,λ̅ is the mean wavelength associated with the transition, n is the index of refraction at the mean wavelength, 2J+1 is the number of levels in the upper J manifold and N is the rare-earth ion density. is the magnetic dipole line strength which was calculated from intermediate coupled wavefunctions. For the doped BIGGSe glasses, n ~ 2.4. When the absorption spectra overlap, as in the case of the (3F3,3F4) and (3H6,3F2) lines, the absorption spectra was deconvoluted by fitting the absorption lines to gaussian distributions centered around the peaks in the absorption spectra. Without a detailed knowledge of the Stark structure of Pr3+ in the host glass matrix, there is some ambiguity in the deconvolution of the lines by this methodology. However, it was found that the final results varied by less than 20%. Since this is within the total error estimated for our Judd-Ofelt calculation, this procedure is justified.
In order to determine the Judd-Ofelt parameters Ω2, Ω4, Ω6 ., the electric dipole line strengths were fit by a least squares fit to
Values of reduced matrix elements, <∥ Ut ∥>, were taken from Weber . The calculated Judd-Ofelt intensity parameters were found to be
The large value of Ω2 found for this material is indicative of the large degree of covalent bonding in the glass system .
From the Judd-Ofelt parameters, line strengths were calculated for all transitions and electric dipole and magnetic dipole intermanifold spontaneous emission rates, and , were determined by
The total radiative rate of a manifold is the sum of the intermanifold spontaneous emission rates, or
From this, we can define the fluorescence branching ratio as
The effective cross section, Σ, which is the stimulated emission cross section integrated over the total band, was calculated from
This effective cross section gives a measure of the peak stimulated emission cross section for transitions. Table 1 lists the results of these calculations.
Error estimates for the rates and cross-sections of Table 1 are ± 35% and are based upon rms deviation between calculated and experimental line strengths in the Judd-Ofelt least square fit, error in the measurement of absorption cross-sections, and error in Pr concentration determination.
The total fluorescence lifetime of a state includes both the radiative and nonradiative contributions to the decay. The fluorescence lifetime τ is then written as
where τrad is the radiative lifetime and τnr is the nonradiative lifetime. Nonradiative contributions to the decay include multiphonon emission, cross-relaxational energy transfer, and direct or migrational energy transfer to quenching centers. The radiative quantum efficiency of a level is fraction of the population of the state that decays radiatively. This can be written as
By comparing calculated radiative rates with experimentally determined lifetimes, one can estimate the radiative quantum efficiency. The results are tabulated in Table 2 for the 0.1 wt.% Pr doped BIGGSe sample. The radiative rates of the (3H6,3F2) and (3F3,3F4) coupled levels are calculated as a weighted pair of the rates of the two levels. Weighting is based upon the population distribution calculated from absorption peak separation and Boltzmann statistics detailed above. Error for the radiative quantum efficiency is determined from the error in the experimentally determined values and the ± 35% error in the Judd-Ofelt values above.
The estimated radiative quantum efficiency of the 1G4 level is 0.7 ± 0.3. Based upon phonon spectra of commercial AMTIR selenide glass which has a maximum phonon energy of ~260 cm-1 and the energy separation of ~3000 cm-1 between the 1G4 and 3F3 level, the multiphonon contribution to the decay rate of this level should be negligible. Consequently, we would expect the radiative quantum efficiency of the level to be closer to 1. The variation between the experimental and calculated rates seen in Table 2 can then be attributed to the accuracy of the Judd-Ofelt methodology. This value for the radiative quantum efficiency of the 1G4 level of Pr in BIGGSe glass is to be compared to the 60–70 % radiative quantum efficiency found in 500 ppm Pr doped GLS glass  and the ~3% radiative quantum efficiency found in Pr doped ZBLAN glass .
The radiative quantum efficiencies of the 3H5 and (3H6,3F2) levels are quite low. This cannot be accounted for by multiphonon relaxation, since the (3F3,3F4) level is expected to be more strongly multiphonon quenched due to the smaller energy gap between the 3F4 and 3F2. One explanation would be migrational energy transfer to H-Se centers which are resonant with the energy gaps of these states. This, however, has not been verified.
The short lifetime of the highly doped 2 wt.% sample indicates strong concentration quenching effects at this concentration. Such effects could include cross-relaxational energy transfer, upconversion and energy transfer to defect sites or impurity centers. The radiative quantum efficiency of all levels at this concentration is quite low (≤1%) based upon the measured lifetimes.
The long lifetimes and strong integrated cross sections tabulated in Table 1 indicate that Pr BIGGSe is a promising candidate for sources in the mid-IR region of the spectra. The 100 μs lifetime of the (3F3,3F4) level and the strong integrated emission cross sections and good quantum efficiency compares well with Pr in LaCl3, which has been successfully lased at 1.6 μm, 5.2 μm and 7.2 μm on the 3F3 ⇒ 3H4, 3F3 ⇒ 3H6 and 3F3 ⇒ 3F2 transitions, respectively [2,3]. The strong integrated cross section of the 3F4 ⇒ 3H6 transition allow the possibility of lasers at 4.0 μm based upon this transition . This wavelength is in the atmospheric transmission window and would be useful for remote sensing and countermeasure applications. Also of interest is the 3.0 μm 1G4 ⇒ 3F3 transition and the 3.4 μm 1G4 ⇒ 3F3 transition. The 1G4 state has a good lifetime, good quantum efficiency and the transitions show strong integrated cross-sections.
The broadband nature of the Pr fluorescence in the 3–5 μm region opens the possibility of using these materials as an alternative to blackbody sources for generating light in this wavelength region. Pr doped BIGGSe fibers pumped by diode lasers at 1.5 μm or 2.0 μm would be a viable phosphor source for chemical sensor or spectroscopic applications. Furthermore, since the phosphor source is essentially a “cool” source, they can be inserted into thermally sensitive environments which preclude the use of hot blackbody sources.
We have spectroscopically characterized the IR transitions of Pr3+ in Barium Indium Gallium Germanium Selenide glass. IR absorption and fluorescence spectra were taken and decay lifetimes of the lower lying states measured. A Judd-Ofelt analysis was performed to characterize the radiative rates of these manifolds. The spectroscopic measurements reveal several potential lines for IR lasers, amplifiers, and high brightness sources.
References and links
1. Shaw, S.R. Bowman, B.J. Feldman, and J. Ganem, “Radiative and multiphonon relaxation of the mid-IR transitions of Pr3+ in LaCl3,” IEEE J. Quantum Electron. , 32, 2166–2172 (1996). [CrossRef]
2. J. Ganem, S.R. Bowman, and B.J. Feldman, “Excited state dynamics of Pr3+:LaCl3,” J. Lumin. , 58, 298–302, (1994). [CrossRef]
3. S.R. Bowman, L.B. Shaw, B.J. Feldman, and J. Ganem, “A 7-μm praseodymium solid-state laser,” IEEE J. Quantum Electron. , 32, 646–649 (1994). [CrossRef]
4. S. R. Bowman, J. Ganem, B. J. Feldman, and A. W. Kueny, “Infrared laser characteristics of praseodymium-doped lanthanum trichloride,” IEEE J. Quantum Electron. , 30, 2925–2928 (1994). [CrossRef]
5. J. S. Sanghera, V. Q. Nguyen, P. C. Pureza, R. E. Miklos, F. H. Kung, and I. D. Aggarwal, “Fabrication of long lengths of low-loss IR transmitting As40S(60-x)Sex glass fibers,” J. Lightwave Technol. , 14, 743–748, (1996). [CrossRef]
6. B. R. Judd, “Optical absorption intensities in rare earth ions,” Phys. Rev. , 127, 750–761 (1963). [CrossRef]
7. G. S. Ofelt, “Intensities of crystal spectra of rare earth ions,” J. Chem. Phys. , 37, 511–520 (1963). [CrossRef]
8. M. J. Weber, “Spontaneous emission probabilities and quantum efficiencies for excited states of Pr3+ in LaF3,” J. Chem. Phys. , 48,. 4774–4780 (1968). [CrossRef]
9. S. Tanabe, T. Hanada, M. Watanabe, T. Hayashi, and N. Soga, “Optical properties of dysprosium-doped low-phonon-energy glasses for a potential 1.3 μm optical amplifier,” J. Am. Ceram. Soc. , 78, 2917–2922 (1995). [CrossRef]
10. J.R. Hector, D.W. Hewak, J. Wang, R.C. Moore, and W.S. Brocklesby, “Quantum efficiency measurements in oxygen containing gallium lanthanum sulphide glasses and fibers doped with Pr3+,” IEEE Photonics Technol. Lett. , 9, 443–445, (1997). [CrossRef]
11. Y. Ohishi, A. Mori, T. Kanamori, K. Fujiura, and S. Sudo, “Fabrication of praseodymium-doped arsenic sulfide chalcogenide fiber for 1.3 μm fiber amplifiers”, Appl. Phys. Lett. , 65, 13–15, (1994). [CrossRef]