Open Access
Add to BibSonomy
Abstract
High-purity samples of the new Ge36Ga5Se59 glass composition doped with 2×1019, 5×1019 and 1020 cm−3 of Pr3+ atoms were synthesized and their luminescent properties were investigated in the 1.6–7.5 µm spectral range. The luminescence registration method involved has made it possible to separate the spectra of the overlapping transitions in Pr3+ in the 2–2.7 µm emission band and to determine the respective lifetimes as well as their concentration dependencies. High sensitivity of the detection system has allowed to detect weak luminescence band at ∼7.2 µm corresponding to 3F3→3F2 transition for the first time in Pr doped glasses. The experiments showed that intense cross-relaxation processes in investigated glasses can cause a steep ∼ 5 µm emission rise with Pr3+ concentration increase. This effect can be of use for the design of efficient ∼ 5 µm luminescent sources with population efficiency significantly exceeding 100% (when pumped at ∼ 1.5 µm) and luminescence quantum yield close to unity.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Rare earth activated chalcogenide glasses are promising materials for the development of infrared luminescent and laser media that can emit at wavelengths longer than available from oxide and fluoride materials. Such luminescent and laser materials, as well as nonlinear wavelength converters based on these glasses, may be of interest for a number of infrared spectroscopy applications in environmental monitoring, medicine and security. The search for the most appropriate chalcogenide glass composition that should combine low optical losses and high (∼1020 at/cm3) concentration of rare earth dopants is the aim of an increasing number of explorers [1]. Despite big efforts still no mid-infrared laser action in chalcogenide glasses have been demonstrated yet. Papers [2,3] overview the luminescent properties of chalcogenide glasses with different rare-earth dopants as well as their potential applications.
Pr3+ ions are some of the most popular rare earth dopants for chalcogenide glasses. Their complicated system of electron levels involves a large amount of optical transitions in the near- and mid-infrared (see Fig. 1).
One of the latest papers [4] presents a detailed model describing the population of Pr3+ excited states in GeAsGaSe glasses when pumped at 1.45 µm. Nevertheless Pr3+ luminescent properties of materials with low-frequency phonons can’t be considered as well-investigated until now. One of the difficulties of their investigations is caused by the spectral overlapping of multiple electronic transitions in both 2–2.7 µm and 3.4–6.5 µm (5 µm) emission bands. Moreover, most investigators have paid no attention to radiationless energy transfer processes between Pr3+ ions that inevitably take place in heavily doped glasses. The authors of [5] were the first who have pointed out the concentration quenching of the luminescence from Pr3+ levels: (3F4, 3F3), (3F2, 3H6) and 3H5 in BaInGaGeSe glass that can contain large (up to 2% wt.) amounts of rare earth ions. Such quenching may be caused both by cross-relaxation processes in Pr3+ ions system and by non-radiational energy transfer to uncontrollable impurities (mostly Se−H, Ge−H groups) in the glass. These impurities can lead to the optical losses also. The above mentioned problems seem to be the reason that mid-infrared lasing has not been achieved yet on Pr3+ ion doped chalcogenide glasses and fibers.
The progress in the technology of chalcogenide glasses has made it possible to fabricate high-purity glasses and optical fibers with Se−H, Ge−H groups concentration not exceeding 5 ppm [6,7]. The minimal optical losses measured in Pr-doped and undoped selenide glass fibers have reached the values of ∼ 0.5–1 dB/m in the 2–4 and 5.5–7 µm spectral ranges. Such glass quality should already enable to observe the cross-relaxation processes clearly.
The aims of the present work were:
- – the investigations of Pr3+ luminescence peculiarities in the 1.6–7.5 µm spectral range in high-purity Ge36Ga5Se59 glass under 1.5 µm excitation. This task included the development of luminescence registration method that has made it possible to separate the emission spectra of the overlapping Pr3+ transitions and to determine the respective lifetimes.
- – the investigation of the radiationless energy transfer processes between Pr3+ ions in Ge36Ga5Se59 glasses with different Pr3+ concentration.
2. Glass composition choice and fabrication procedure
For our experiments, we have chosen the glass of the basic composition Ge36Ga5Se59. This choice is caused by several factors. Glasses based on Ga and Ge chalcogenides have good glass-forming ability [8,9]. Rare earth ions solubility in these glasses is much better than in glasses based on As chalcogenides [10]. An additional advantage of As-free glasses is their lower toxicity. Glass samples containing 2×1019, 5×1019 and 1×1020 cm−3 of praseodymium, were prepared according to the method described in [11].
As was already mentioned, hydrogen is the most harmful impurity that may induce absorption losses and the luminescence quenching in chalcogenide glasses. Using the results of [8], we estimated hydrogen content in the form of Se−H and Ge−H groups in our Pr-free glass samples as 1 ppm. It is difficult to measure these groups absorption in Pr-doped glasses because of their absorption peaks placed at λ=4.5 µm and at λ=4.9 µm respectively [7]. Both these wavelengths overlap with Pr3+ absorption band (3H4→3H5 transition). But we estimate the total content of hydrogen atoms for doped samples on the level of 1 ppm also.
The samples for investigations were prepared in the form of cylinders with a diameter of 12 mm and a length of up to 80 mm. Polished discs and rods with a thickness of 2, 5 and 35 mm with plane-parallel faces were used in optical measurements. The list of synthesized Pr-doped glasses is presented in Table 1.
Table 1. Investigated glass samples.
3. Overview absorption and emission spectra of Pr 3+ ions
The transmission range of Ge36Ga5Se59 glass extends from ∼ 1 to ∼ 15 µm.
Figure 2 presents the overview absorption (bottom) and emission (top) spectra of the glass sample #1 with minimal (2×1019 cm−3) Pr content in the 1–8 µm region. The absorption spectrum consists of:
- - two overlapping but well-resolved bands peaking at 1.48 and 1.59 µm corresponding to 3F4 and to 3F3 states respectively. The energy gap between them is 468 cm−1. This value will be used below for identification of some luminescent transitions;
- - the band peaking at ∼ 2.03 µm corresponding to non-resolved transitions to 3F2 and 3H6 states;
- - the wide 3.7–5.3 µm band corresponding absorption to 3H5 state.
Fig. 2. The absorption (bottom) and emission (top) spectra of the glass sample #1 with 2×1019 cm−3 Pr content. (Blue lines show selected portions of the spectrum with 10-fold multiplication. Weak 3F2→3H5 and 3F3→3F2 transitions are indicated by arrows. The recession in the emissions spectrum at ∼ 4.25 μm is due to CO2 absorption in the air. The recession at 4.85 μm is an artifact caused by the specificity of the registration system).
The presented emission spectrum is not corrected for the spectral response of the registration system but gives the overall qualitative picture.
The spectrum consists mainly of the three emission bands peaking at ∼ 1.68, 2–2.7 and ∼ 5 µm. 1.68 µm band corresponds to 3F3→3H4 transition. 2–2.7 and 5 µm bands are complex and corresponds to several overlapping electronic transitions. The band that peaks at 2–2.7 µm consists of 4 electron transitions (3F3, 3F4→3H5 and 3F2, 3H6→3H4). It is difficult to distinguish them spectrally. Nevertheless we have managed to separate these transitions by their lifetimes (see the next section). The mid-infrared emission band at ∼ 5 µm is the most complicated. In the case of ∼ 1.5 µm pumping it should include 4 optical transitions (see Fig. 1).
Besides the three mentioned intensive emission bands we have also detected two weak emission peaks that should correspond to 3F2→3H5 (λ = 3.4 µm) and 3F3→3F2 (λ = 7.2 µm) transitions.
The 7.2 µm Pr3+ luminescent peak is registered, to the best of our knowledge, for the first time in chalcogenide glasses. It should be noted that 7÷8 µm emission bands of Sm3+ and Tb3+ ions were also observed recently in rare-earth activated chalcogenide glasses [9,10]. 3F3→3F2 transition of Pr3+ ions is of special interest as it is the longest-wavelength rare earth transition at which laser action in LaCl3 crystal was ever demonstrated [12]. Worth noting, that we were able to observe the two mentioned weak transitions in the sample with the smallest doping level only.
4. The luminescence kinetics measurements and the component separation of the 2÷2.7 µm emission band
An important characteristic of luminescent transition is its decay kinetic function that may give information about the ways of excitation and relaxation mechanisms.
First of all it should be noticed, that correct luminescence time dependence measurements in high refractive index materials require special accuracy. Wide angles of total internal reflection cause the capture of the luminescent light inside the sample until it is reabsorbed by the dopant and re-emitted with some time delay. In case of high luminescence quantum yield such reabsorption effect may cause very big lifetime overestimation. The ways to avoid this mistake were discussed in [13]. Briefly, small-size (< 1 mm3) samples in suitable immersion should be used. Unfortunately, there is a problem to find a liquid with high (> 2) refractive index and good transparency up to ∼ 6–7 µm. That is why we have immersed small glass splinters in molten sulfur and quenched the melt rapidly.
It is well known that the 2–2.7 µm Pr3+ emission band is complicated and consists of several electronic transitions. Experiments proved that the rise-and-decay functions of different transitions at various registration wavelengths within the 2–2.7 µm emission band differed radically. Despite strong overlapping of multiple electronic transitions we have found that some of such functions at selected wavelengths correspond to definite electronic transitions as presented in Fig. 3.
Fig. 3. Luminescence rise and decay functions of the glass sample #1 with minimal (2×1019 cm−3) Pr content. Black curve (registration wavelength 2.5 µm) - 3F3→3H5 transition; Red curve (registration wavelength 1.68 µm) - 3F3→3H4 transition; Blue curve (registration wavelength 2.03 µm ) - 3F2→3H4 transition.
Since the two curves with the registration wavelengths 2.50 µm and 1.68 µm have no measurable rise time and coincide with each other within the measurement accuracy, we can conclude that both emission lines originate from the same electronic state 3F3. Both curves are more or less exponential with e-fold lifetimes of 135 µs. The luminescent response measured at 2.03 µm is quite different. This curve in Fig. 3 shows slowly (∼140 µs) rising from zero buildup with further exponential relaxation with characteristic e-fold time in the order of 1 ms. Looking at the energy level scheme in Fig. 1 we can conclude that emission at 2.03 µm should correspond to 3F2→3H4 transition.
Taking into account a big difference in the lifetimes of 3F3 and 3F2 levels we have suggested a way to separate the components of the 2÷2.7 µm emission band. For this reason we have registered the emission spectra at different pump modulation regimes. The curve 1 in Fig. 4 is the emission spectrum registered at the pump modulation frequency of 65 Hz.
Fig. 4. Luminescence spectra of the glass sample #1 with minimal (2×1019 cm−3) Pr content. Curve 1 – pumping diode modulated at 65 Hz. Curve 2 – pumping diode modulated at 4 kHz. Curve 3 – the difference between them. Vertical lines indicate the registration wavelengths of the decay functions in Fig. 3.
This frequency is much less than the reciprocal lifetimes of both 3F3 and 3F2 levels and the resulting spectrum is the sum of the transitions originating from (3F4, 3F3) and (3F2, 3H6) manifolds. The curve 2 corresponds to pump modulation frequency 4 kHz. This value is much higher than the reciprocal lifetime of 3F2, 3H6 states. Thus the luminescence originating from these levels is practically not registered by the selective noise-rejecting voltmeter, and the curve 2 in Fig. 4 corresponds to emission originating from (3F4, 3F3) manifold only.
Worth emphasizing, that the luminescent response at λ=2.03 µm in Fig. 3 grows smoothly from zero. It means that emission at this wavelength includes practically no photon flux from (3F4, 3F3) levels that is excited by the pump laser directly. In contrast, the luminescent response at 2.5 µm in Fig. 3 reaches its maximum instantly. We can conclude that at this wavelength the emission includes only the flux from (3F4, 3F3) manifold that is excited directly.
Subtraction of the normalized experimental curves in the Fig. 4 gives the emission spectrum from 3F2 and 3H6 levels (curve 3 in Fig. 4). We can see that the emission from these levels in our glass demonstrate peaks at ∼ 2.2 and ∼ 2.6 µm respectively which practically do not overlap with each other. The gap between these peaks is about 700 cm−1. We have used this value when indicating the observed Pr3+ transitions in Fig. 1 (see above).
5. Luminescence concentration quenching effects
Figure 5 presents the 1.68 µm (3F3→3H4 transition) luminescence decay functions of the samples with various Pr doping level. The observed reduction in lifetime at Pr content increasing indicates the presence of luminescence concentration quenching effects. Pay attention that all 3 decay curves are non-exponential.
Fig. 5. 1.68 µm luminescence decay functions of the samples ##1–3 with various Pr doping level: 2×1019 cm−3 (curve 1), 5 ×1019 cm−3 (curve 2) and 1020 cm−3 (curve 3).
Provided that a few contaminants content, the exponential asymptote of the sample #1 (with the lowest doping level) decay should be close to Pr3+ spontaneous-radiative decay function in this glass. Still supposing minor contaminants content we can assume that the cross-relaxation should be the main reason for the observed luminescence concentration quenching.
Figure 6 presents the luminescence rise and decay functions in the mid infrared. They were recorded using liquid N2 cooled CdHgTe photoresistor with a special made bandpass filter transmitting at 3.2–5.3 microns with no monochromator and include all the optical transitions of this spectral range.
Fig. 6. Mid-infrared luminescence rise and decay functions. Curve numbers correspond to sample numbers in Table 1.
It is seen that the short-lifetime component (t < 1 ms) presents only in the sample #1 with the lowest Pr content. The samples 2 and 3 show no such decay components. To the opposite, they demonstrate noticeable emission intensity rise during the first 50–100 µs after the pump pulse. It means, that the luminescence of samples #1 corresponding to emission from high-lying excited states (3F4, 3F3) and (3F2, 3H6,) to 3H6 and 3H5 correspondingly but the luminescence of samples #2 and #3 corresponds mostly to 3H5→3H4 transition. Since the observed lifetime explicitly decrease with concentration, we have to concede that the luminescence quenching of 3H5→3H4 transition could be explained by energy migration to some uncontrollable impurities. The most probable impurities are of course Se−H, Ge−H groups having absorption band (4.6–4.8 µm) in good resonance with 3H5→3H4 transition [7].
The lifetime measurement results for all the samples are summarized in Table 2.
Table 2. e-fold lifetimes (in µs).
Figure 5 and Table 2 show also significant concentration shortening of the (3F3, 3F4) manifold lifetime – from 135 down to 35 µs. The lifetimes of the (3F2, 3H6) manifold also decrease significantly. In these cases, the reasons of such concentration quenching can be also the cross-relaxation processes. Such concentration quenching may be caused by an interaction of a Pr3+ ion excited by 1.48 µm radiation to 3F3 state with a nearby non-excited Pr3+ ion as illustrated in Fig. 7 (left). As a result of this cross-relaxation process one ion turned out in the 3H5 state and other − in 3H6 state. Then emerged Pr3+ ion in 3H6 state can interact again with another ground-state Pr3+ ion as illustrated in Fig. 7 (right) and two more ions turn out at 3H5 state. Thus one Pr3+ ion excited into high-lying 3F3 state may generate up to three Pr3+ ions at 3H5 level. In other words, the totality of cross-relaxation processes can lead to population of the 3H5 excited state and to growth of luminescence intensity at ∼ 5 µm.
Thus the totality of cross-relaxation processes can lead to efficient population of the 3H5 excited state and to growth of luminescence intensity at ∼ 5 µm. This effect was really observed in our experiments (see Fig. 8). The spectra in Fig. 8 were detected using a single grating and a single photodetector (CdHgTe) so that the ratio between ∼2.5 and ∼5 microns transition intensities was well fixed, though the signal-to-noise ratio in this case was lower than in Fig. 2.
Fig. 8. Relative intensities of 2–2.7 and 5 µm emission bands depending on Pr3+ concentration. Curve numbers correspond to sample numbers in Table 1.
It shows the relative intensities variation of the 2–2.7 µm and 5 µm emission bands with Pr concentration changing. The intensities of the ∼ 2–2.7 µm band in Fig. 8 are calibrated proportional to the integrated 3F3 level decay functions shown in Fig. 5. It can be seen that despite a possible presence of quenching of the 3H5 excited state by uncontrollable impurities the ∼ 5 µm emission intensity increases approximately 3-fold with concentration increasing. Thus we have shown that in heavily Pr-doped glass the 3H5 level can be populated by cross-relaxation mechanism with high quantum efficiency (ideally up to 300%). This effect can be of use for the design of highly efficient ∼ 5 µm fiber luminescent sources. Despite the fact that the quantum yield of ∼ 5 µm Pr3+ luminescence in these glasses is a few % only, Ga5Ge20Sb10S65 glass doped with Pr and Dy has already been used as a wideband infrared luminescent fiber light source [14]. Calculation by method Judd-Ofelt (see the next section) shown that in our Pr doped Ge36Ga5Se59 glass the quantum yield can be close to unity. This allows to hope that the usage of heavily Pr-doped Ge36Ga5Se59 glass can greatly enhance the efficiency of luminescent fiber sources in ∼ 5 µm spectral range.
6. Judd-Ofelt analysis of Pr3+ in the investigated glass
For Judd-Ofelt analysis we have used the absorption spectrum of the sample #3 and the program RELIC [15]. Some of the calculated parameters are presented in Table 3. The calculated Judd-Ofelt intensity parameters were found to be Ω2 = 13.0×10−20, Ω4 = 5.1×10−20, Ω6 = 7.4×10−20.
Table 3. Calculated energies of bands, oscillator strengths, predicted spontaneous-radiative transition rates, branching ratios and lifetimes of Pr3+ in Ge36Ga5Se59 glass
It is interesting to compare the calculated and the experimental (presented in Table 2) lifetimes of excited states for the sample #1 with the lowest doping level. The experimental value for 3H5 state is not much less than the calculated one. It means that the luminescence quantum yield in our glass is big (comparable to unity).
The calculated lifetimes of 3H6 and 3F2 are quite different. Since these two levels are thermally coupled, their common experimental lifetime is intermediate between the calculated ones.
Similar situation takes place with 3F3 and 3F4 levels, but in this case their common lifetime is mostly determined by the lower 3F3 level only. The calculated lifetime (241 µs) is still comparable with the experimental e-fold lifetime (135 µs) and with the asymptotic lifetime estimation in Fig. 4 (205 µs). It means that even ∼ 7 µm transition in our glass is practically not quenched. The Judd-Ofelt analysis allows us to explain the observed effect of the strong ∼5 µm emission increase with the doping level increasing as shown in Fig. 8. As it can be seen from Table 3, the branching ratio of all the transitions terminating at 3H5, 3H6, 3F2 levels are low. For this reason the population efficiency of 3H5 level in low-concentrated sample pumped at ∼ 1.5 µm is also low. The situation changes in highly-doped samples where the main mechanism of 3H5 level population are cross-relaxation processes.
7. Summary
A set of high-purity Ge36Ga5Se59 glass samples with Pr3+ concentrations of 2×1019, 5×1019 and 1020 cm−3 was synthesized and their luminescent properties in the 1.6–7.5 µm spectral ranges were investigated. The proposed registration method has made it possible to separate the spectra of the overlapping transitions and to determine the respective lifetimes and their concentration dependencies. The highly sensitive detection system has allowed to detect for the first time the weak ∼ 7.2 µm luminescence band corresponding to 3F3→3F2 transition in a Pr3+ doped chalcogenide glass. The experiments showed that intense cross-relaxation processes in high-purity glasses can cause a steep ∼5 µm emission rise with Pr3+ concentration increase. This effect can be of use for the design of efficient ∼5 µm luminescent sources with population efficiency significantly exceeding 100% (when pumped at ∼ 1.5 µm) and luminescence quantum yield close to unity.
Funding
Russian Foundation for Basic Research (18-29-20079mk).
References
1. G. Tao, H. Ebendorff-Heidepriem, A. M. Stolyarov, S. Danto, J. V. Badding, Y. Fink, J. Ballato, and A. F. Abouraddy, “Infrared fibers,” Adv. Opt. Photonics 7(2), 379–458 (2015). [CrossRef]
2. J. S. Sanghera, L. B. Shaw, and I. D. Aggarwal, “Chalcogenide Glass-Fiber-Based Mid-IR Sources and Applications,” IEEE J. Sel. Top. Quantum Electron. 15(1), 114–119 (2009). [CrossRef]
3. L. Shaw, P. Thielen, J. Sanghera, S. Bayya, and I. Aggrawal, “Rare-Earth Doped Chalcogenide Glass for Mid- And Long Wave IR Fiber Lasers,” Advances in Fiber Devices (L.N. Durvasula, Editor, Proc. of SPIE)4974, 60–64 (2003).
4. S. Sujecki, L. Sojka, E. Beres-Pawlik, K. Anders, R. Piramidowicz, Z. Tang, D. Furniss, E. Barney, T. Benson, and A. Seddon, “Experimental and numerical investigation to rationalize both near-infrared and mid-infrared spontaneous emission in Pr3+ doped selenide chalcogenide fiber,” J. Lumin. 209, 14–20 (2019). [CrossRef]
5. L. B. Shaw, B. B. Harbison, B. Cole, J. S. Sanghera, and I. D. Aggarwal, “Spectroscopy of the IR transitions in Pr3+ doped heavy metal selenide glasses,” Opt. Express 1(4), 87–96 (1997). [CrossRef]
6. E. V. Karaksina, V. S. Shiryaev, M. F. Churbanov, E. A. Anashkina, T. V. Kotereva, and G. E. Snopatin, “Core-clad Pr(3+)-doped Ga(In)-Ge-As-Se-(I) glass fibers: Preparation, investigation, simulation of laser characteristics,” Opt. Mater. 72, 654–660 (2017). [CrossRef]
7. V. S. Shiryaev, E. V. Karaksina, M. F. Churbanov, T. V. Kotereva, B. S. Stepanov, L. A. Ketkova, I. I. Evdokimov, V. V. Koltashev, V. G. Plotnichenko, A. I. Filatov, and I. N. Antonov, “Special pure germanium-rich GaGeAsSe glasses for active mid-IR fiber optics,” Mater. Res. Bull. 107, 430–437 (2018). [CrossRef]
8. V. G. Borisevich, V. G. Plotnichenko, I. V. Skripachev, and M. F. Churbanov, “Influence of hydrogen impurity on optical losses in As-S and As-Se glasses,” High-Purity Substances (Vysokochistie Veshchestva) 2, 11–21 (1994). (in Russian).
9. F. Starecki, N. Abdellaoui, A. Braud, J.-L. Doualan, C. Boussard-Plédel, B. Bureau, P. Camy, and V. Nazabal, “8 µm luminescence from a Tb3+ GaGeSbSe fiber,” Opt. Lett. 43(6), 1211–1214 (2018). [CrossRef]
10. F. Starecki, A. Braud, N. Abdllaoui, J.-L. Doualan, C. Boussard-Pledel, B. Bureau, P. Vamy, and V. Nazabal, “7 to 8 µm emission from Sm3+ doped selenide fiber,” Opt. Express 26(20), 26462–26469 (2018). [CrossRef]
11. M. V. Sukhanov, A. P. Velmuzhov, T. V. Kotereva, I. V. Skripachev, and M. F. Churbanov, “New approach for preparation of high-purity sulfide-germanium glasses doped with praseodymium,” Opt. Mater. Express 9(8), 3204–3214 (2019). [CrossRef]
12. S. R. Bowman, L. B. Shaw, B. J. Feldman, and J. Ganem, “A 7-µm Praseodymium-Based Solid-state Laser,” IEEE J. Quantum Electron. 32(4), 646–649 (1996). [CrossRef]
13. B. Denker, V. Dorefeev, B. Galagan, V. Koltashev, S. Motorin, S. Sverchkov, and V. Plotnichenko, “Rare-earth ions doped zinc-tellurite glass for 2÷3 µm lasers,” Appl. Phys. B: Lasers Opt. 124(12), 235 (2018). [CrossRef]
14. J. Ari, F. Starecki, C. Boussard-Pledel, Y. Ledemi, Y. Messaddeq, J.-L. Doualan, A. Braud, B. Bureau, and V. Nazabal, “Co-doped Dy3+ and Pr3+ Ga5Ge20Sb10S65 fibers for mid-infrared broad emission,” Opt. Lett. 43(12), 2893–2896 (2018). [CrossRef]
15. M. P. Hehlen, M. G. Brik, and K. W. Krämer, “50th anniversary of the Judd-Ofelt theory: An experimentalist’s view of the formalism and its application,” J. Lumin. 136, 221–239 (2013). [CrossRef]
