- Email: [email protected]
Laser potential of calcium aluminate glasses
Journal of Non-Crystalline Solids 496 (2018) 29–33
Contents lists available at ScienceDirect
Journal of Non-Crystalline Solids journal homepage: www.elsevier.com/locate/jnoncrysol
Laser potential of calcium aluminate glasses B.I. Denker a b
a,b
, B.I. Galagan
a,b
, S.E. Sverchkov
a,⁎
T
Prokhorov General Physics Institute of the Russian Academy of Sciences, Moscow, Russia Center of Laser Technology and Material Scince, Moscow, Russia
A R T I C LE I N FO
A B S T R A C T
Keywords: Calcium aluminate glass Laser glass, tunable thulium laser
Deeply dehydrated calcium aluminate glasses doped by rare-earth ions were synthesized and investigated. Spectroscopic properties of Tm3+ and Ho3+ doped glasses were found attractive for fiber lasers emitting in 2–2.5 μm spectral range. Cascade laser action (at 2300 and 2100 nm) in Tm doped bulk calcium aluminate glass was demonstrated for the first time.
1. Introduction Glasses based on CaO-Al2O3 system (calcium aluminate glasses, CAGs) are relatively new materials that have already found their place in optical industry. Multicomponent CAG compositions can be crystallization stable enough to be produced in commercial volumes (several kilograms). Their attractive features are high infrared transmittance (up to 5 ÷ 6 μm) combined with excellent mechanical properties and thermal conductivity exceeding those of fused silica as well as excellent moisture resistance. In contrast to fused silica glass, these glasses can easily include up to 3 ÷ 5% mole of lanthanide oxides. Nevertheless, the amount of investigations devoted to spectroscopy and to potential laser applications of the rare earth activated CAGs is rather limited [1–5]. Limited interest may be explained by tough competition with well-developed activated silica fiber lasers and fluoride (mainly ZBLAN) fiber lasers. Indeed, the ecological niche suitable for activated CAGs and fibers may be rather narrow. But we believe it do exist because CAGs have much better infrared transmittance and rare earth solubility than fused silica as well as much better mechanical properties and moisture resistance than ZBLAN. This niche should include fiber lasers exploiting the rare-earth ions (RE3+) transitions that are noticeably or almost completely quenched by multiphonon relaxation in silica but still radiative in CAGs. The papers [3,4] have already disclosed some part of the area where activated CAGs are of interest: Tm-doped lasers and amplifiers operating at ~ 1.8 μm (3F4e3H6 transition) and at ~1.45 μm (3H4e3F4 transition). Of course, in case of low-gain glass lasers and amplifiers mostly fiber configuration is of interest. Optical fibers drawing of CAGs is a rather tricky procedure because of easy glass devitrification. Nevertheless, it was shown to be possible. For example, a relatively simple approach was realized in the dissertation [6]. The custom modification of the fiber draw furnace was
⁎
Corresponding author. E-mail address: [email protected] (S.E. Sverchkov).
https://doi.org/10.1016/j.jnoncrysol.2018.04.047 Received 19 March 2018; Received in revised form 24 April 2018; Accepted 25 April 2018 0022-3093/ © 2018 Elsevier B.V. All rights reserved.
required to achieve fiber drawing from the crystallization-prone CAG compositions. This included minimization of the heat zone susceptor width and adding water-cooling to narrow the heat zone temperature profile further. Using this technic the Schott calcium-aluminate IRG11 glass sample was successfully drawn into one-meter long Teflon coated 550 μm diameter fibers. The delivery of Er:YAG laser power via this fiber has also been also demonstrated. The fibers prepared in this work had no any glass cladding and no rare-earth activation of them was considered. A more advanced approach to fabrication of CAG optical fibers was developed in [5], where composite (multimaterial) fibers with Tm-activated core in a crystallization–stable silicate cladding were made. Their first attempt to fabricate a fiber with CAG core started from a rodin-tube preform consisting of a Tm-doped aluminate glass jacketed by a cladding tube of B-doped silica glass (10B2O3: 90SiO2). The preform was drawn into 125 μm diameter fiber with a 25–30 μm diameter core. Unfortunately, the core contained ~50 wt% SiO2 and ~5 wt% B2O3 due to intense diffusion during high-temperature drawing. The presence of boron and silicon in the core inevitably spoils the unique IR properties of the CAG glass. In order to overcome this serious obstacle the authors of [5] have developed special cladding glass compositions having softening temperatures close to that of the core glass. Using these special glasses and drawing at lower temperatures, fibers with Tmdoped silica-free calcium-aluminate core were successfully fabricated. Unfortunately, paper [5] gives no information about gain or laser properties of these fibers. All the cited papers show activated CAGs and fibers made on their base are novel interesting materials for near- and mid-infrared lasers and amplifiers. In the present paper, we would propose a laboratory synthesis procedure of deeply dehydrated CAGs. Deep dehydration is required for two purposes. First purpose is the adequate luminescence
Journal of Non-Crystalline Solids 496 (2018) 29–33
B.I. Denker et al.
decay characterization of activated glasses in the mid-infrared. The second purpose is the need for low absorption losses in lasing tests. Then we would investigate the luminescent properties of Tm, Ho and Er-activated deeply dehydrated CAGs in the spectral range where these glasses seem the most interesting - 2÷2.7 μm. Analysis of the laser potential of these activators and lasing tests in bulk glass will be added.
2. Glass fabrication procedure Our first task was to develop a laboratory fabrication procedure that gives CAG samples of high optical quality (absence of light scattering inclusions, deep dehydration and minimal striae). The developed CAG laboratory synthesis procedure noticeably differs from the methods used for it before. The main feature of this procedure is the usage of inductively heated vitreous carbon crucibles. The big advantage of vitreous carbon is its total non-wettability by the glass melt. For this reason, the prepared glass ingot easily drops out of the crucible. It should be noted, that the CAG compositions prepared in such utensils should not content components that can be reduced by carbon. In our glasses doped by Tm3+, Ho3+ or Er3+ no reduction was detected. Addition of Yb2O3 (that could be used for sensitization of all three above-mentioned dopants) resulted in partial reduction of Yb3+ to Yb2+. Therefore, hereinafter we have avoided using Yb. Infrared laser applications require deep dehydration of the active medium to prevent absorption losses and dopants luminescence quenching by OH-groups. Dehydration of the CAG melt was implemented by low-pressure (~0.05 at.) dry nitrogen atmosphere during synthesis procedure and by ~5% mol. of fluorides incorporated into the glass composition. Dry nitrogen atmosphere was prepared by liquid nitrogen evaporation. The basic glass composition we have used (46 CaO, 5 CaF2, 36 Al2O3, 6 MgO, 6 SrO, 1 Ln2O3, where Ln means this or that combination of La, Tm, Ho, and Er), was close to one of the most crystallization stable compositions that were found in [7]. The main difference was the usage of strontium oxide SrO instead of BaO since barium oxide turned to be too volatile at low atmospheric pressure. The ~30 g charge was loaded into the crucible and kept for 3–4 h at 1550–1600 °C and at lowered nitrogen pressure. Then, after the nitrogen pressure was risen to the atmospheric value, the high frequency power was switched off and the crucible with the melt was dragged out of the heat insulating ceramics for rapid cooling. Fig. 1 shows the appearance of as-melted glass ingots after their annealing at 750 °C. The ingots volume was clear from crystalline of
Fig. 2. Transmission spectra of a non-dehydrated (black curve, #1) and welldehydrated (#2) glass samples. Sample thicknesses are indicated near the curves.
gaseous inclusions (though the upper surface of them may be spotted by glassy foam and crystals). No light scatter of the red HeeNe laser can be visible in the samples volume by naked eye. Hydroxyl groups in CAG exhibit wide and intensive absorption band that can hinder laser action in glass. Therefore, deep dehydration is a necessary and very important condition for mid-infrared laser glasses. Fig. 2 shows the transmission spectra of two Tm-doped CA glasses recorded with Cary 5000 spectrophotometer. The first 6.5 mm thick sample (black curve) wasn't vacuum treated. Its spectrum exhibits hydroxyl absorption band (extinction ~0.2 cm−1) peaking at ~2920 nm. Blue and green curves correspond to glass plates of different thicknesses (19 and 6.5 mm respectively) made of a well-dehydrated glass sample #2. It can be seen that the spectra of dehydrated glass plates despite their different thicknesses coincide with the accuracy of about 1%. The difference between the blue and the green curves is purely stochastic and has nothing common with the hydroxyl absorption band. For this reason, we can state that OH-group absorption at 2920 nm in this glass is < 1% or 0.01/(1.9–0.65) = 0.008 cm−1. Practically we were not able to detect OH absorption in any of the dehydrated CAG samples having thickness of about 2 cm. Fig. 3 shows the results of differential thermal analysis of one of the
Fig. 1. The appearance of the CAG ingots. The ingots (left to right) are doped with Er2O3, Yb2O3, Ho2O3 and Tm2O3. Note the yellow-brownish color of Yb-doped sample due to Yb2+ presence. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) 30
Journal of Non-Crystalline Solids 496 (2018) 29–33
B.I. Denker et al.
Fig. 4. Emission spectra of Tm3+ and Ho3+ in CAG.
Fig. 3. Differential thermal analysis of CAG.
prepared glasses. The glass transition temperature Tg was determined to be 796 °C. The intensive exothermic peak at 992 °C corresponds to glass crystallization and the endothermic peak at 1372 °C corresponds to crystals melting. Due to high crystallization ability, fiber drawing of this glass is possible either in a narrow temperature gap between Tg and crystallization onset temperatures, or above the crystals melting temperature. Naturally, fabrication of optical fiber with this glass core needs in the appropriate cladding glass selection.
3. Spectral peculiarities of rare Earth ions iN CAG Luminescent lifetime measurements of Tm3+ and Ho3+ and Er3+ in CAGs were made using pulsed (~10 ns) widely tunable OPO excitation, a grating monochromator and a double-stage TEC cooled 2.6 μm cut-off InGaAs photodiode. In case of 3F4 Tm3+ and 5I7 Ho3+ levels lifetime measurements special precautions were taken to avoid reabsorption of the luminescent light in the samples. These measurements were fulfilled using tiny (< 1 mm3) glass splinters immersed into highly refracting liquid - glycerin, n~1.55. Otherwise, the reabsorption effect can cause quite noticeable lifetime overestimations. The decay functions were averaged over 256 laser pulses at least. All the lifetime measurements were statistically checked by using several glass samples obtained in different synthesis processes. All the registered luminescence decay functions were close to exponential. Taking into account the statistical data (the lifetime measurements in several independently synthetized glass samples) and the characteristics of the registration system we have estimated the lifetime measurements accuracy as ± 5 μs for the lifetimes shorter than 1 ms, ± 20 μs for Tm3+ 3 F4 excited state and ± 100 μs for Ho3+ 5I7 level. The doping concentrations were 1% mol. For Er2O3 and Ho2O3 (that corresponds to ~4.75 × 1020 of dopant ions per cm3) and 0.21% mol. (1020 ions/cm3) for Tm2O3. The latter concentration is small enough to neglect the effect of cross-relaxation of 3H4 manifold [4]. Still it was
Fig. 5. Transmission spectrum of the laser rod. Arrow indicates the pumping wavelength.
high enough for laser tests (see below). Table 1 presents the measured e-fold lifetimes in CAG in comparison with literature data for fused silica. It contains also the radiative lifetimes of Tm3+ levels in calcium aluminate glass of a close composition. These data were calculated using Judd-Ofelt method for calcium aluminate glass of a close composition [4]. The ratio of the experimental and calculated values gives the quantum yields of the corresponding levels: 26% for 3H4 and 60% for 3F4. Both these values are big enough for efficient and low-threshold lasing. The same can be said about 5I7 Ho3+ level lifetime. All the mentioned lifetimes are about an order of magnitude longer in CAG than in fused silica. It means that the lasing
Table 1 Metastable levels lifetimes, μs. Dopant ion and its metastable manifold Tm
3+ 3
H4
Tm
3+ 3
F4
Ho3+ 5I7 Ho3+ 5I6 Er3+ 4I11/2
CAG
Fused silica
220 ± 5 (experimental, this work, coincides well with [4] data) 845 (radiative, [4]) 2500 ± 20 (experimental, this work) 4162 (radiative, [4]) 9400 ± 100 50 ± 5 100 ± 5
20 [8]
31
420 [8] 600 [9] Totally quenched Totally quenched
Journal of Non-Crystalline Solids 496 (2018) 29–33
B.I. Denker et al.
Fig. 6. A photo of the laser setup.
4. Bulk CAG lasing demonstration The lasing test was held with Tm doped CAG since we had a suitable high-power pulsed pump source for it – a flashlamp pumped ruby laser. A 30 mm long flat-parallel polished Tm:CAG sample was pumped by a pulsed ruby laser (incident pump pulse energy up to 0.7 J, pump beam diameter ~2 mm). Fig. 5 presents the transmission spectra of laser element. The laser cavity was formed by a concave (R = 100 mm) mirror with high reflectivity in a wide spectral range from ~2.0 to ~2.4 μm, and flat outcoupler (0.15% at 2300 nm and 1% at 2100 nm). Fig. 6 is a photograph of a CAG glass element being pumped by a ruby laser in the cavity. The lasing wavelengths were analyzed using the same monochromator and photodetector as in luminescent measurements. Double-wavelength cascade lasing was observed. The lasing wavelengths varied from pulse to pulse in the ranges shown in the Fig. 7 that presents the oscillograms of the pump and Tm: CAG lasing pulses. 5. Conclusion Fig. 7. Oscillograms of the ruby laser pump pulse and lasing at two optical transitions of thulium ions (3H4 - 3H5 and 3F4 - 3H6).
Deeply dehydrated rare-earth ions doped samples of calcium aluminate glasses were prepared by high-frequency melting in vitreous carbon crucibles. The differential thermal analysis showed that fiber drawing of this glass is quite possible at relatively low temperatures (~900 °C). Luminescent properties of thulium and holmium in CAG were found promising for tunable 2–2.5 μm laser applications. Cascade laser action (at 2300 and 2100 nm) in Tm doped bulk calcium aluminate glass was demonstrated for the first time.
thresholds in CAG should also be an order of magnitude lower all other parameters being equal. The lifetimes of 4I11/2 Er3+ and 5I6 Ho3+ levels in CAG (optical transitions at 2.7 and 2.9 μm) seem to be too short for any practical laser. Therefore, we would not consider them further. The emission spectra presented in Fig. 4 were measured using CW laser diode excitation at 800 nm (for Tm doped glass) and 668 nm (for Ho doped glass) and the same monochromator and photodiode as in lifetime measurements. In comparison to other oxide or fluoride glasses, the presented spectra of Tm3+ and Ho3+ in CAG look rather unusual due to enhanced width and red shift. It is especially notable for 3F4 - 3H6 Tm3+ transition that extends up to ~2.4 μm and overlaps with ~2.3 μm 3 H4 – 3H5 transition. Thus, the total tuning range of Tm-doped CAG laser can extend from ~1.9 to ~2.4 μm that is hardly possible in any other type of glass. In case of Ho3+ doped glasses the tuning range is narrower but also extends to ~2.2 μm.
Acknowledgements The authors are thankful to A. M. Kut'in and A. D. Plekhovich for DTA measurements. The investigation was supported by RFBR grant 1702-00369. The authors have no competing interests to declare. References [1] D.F. de Sousa, J.A. Sampaio, L.A.O. Nunes, Energy transfer and the 2.8-μm emission of Er3+- and Yb3+-doped low silica content calcium aluminate glasses, Phys. Rev. B 62 (5) (2000) 3176–3180 https://doi.org/10.1103/PhysRevB.62.3176.
32
Journal of Non-Crystalline Solids 496 (2018) 29–33
B.I. Denker et al.
Dissertation Submitted to the Graduate School - New Brunswick Rutgers, the State University of New Jersey, https://rucore.libraries.rutgers.edu/rutgers-lib/24549/ PDF/1/play/, (2008). [7] E.V. Uhlmann, Glass Forming Ability, Structure and Spectroscopic Properties of Silica-free Calcium Aluminate Based Glasses. A Dissertation Submitted to the Faculty of the Department of Materials Science and Engineering for the Degree of Doctor of Philosophy in the Graduate College, The University of Arisona, https://core.ac.uk/ download/pdf/143732190.pdf, (1995). [8] B.M. Walsh, N.P. Barnes, Comparison of Tm:ZBLAN and Tm: silica fiber lasers; spectroscopy and tunable pulsed laser operation around 1.9 μm, Appl. Phys. B Lasers Opt. 78 (2004) 325–333 https://doi.org/10.1007/s00340-003-1393-2. [9] D.C. Hanna, R.M. Percival, R.G. Smart, J.E. Townsend, A.C. Tropper, Continuouswave oscillation of thulium-doped silica fibre laser, Electron. Lett. 25 (1989) 593–594 https://doi.org/10.1049/el:19890403.
[2] D.F. De Sousa, L.A.O. Nunes, J.H. Rohling, M.L. Basso, Laser emission at 1077 nm in Nd3+-doped calcium aluminosilicate glass, Appl. Phys. B 77 (2003) 59–63 https:// doi.org/10.1007/s00340-003-1247-y. [3] S.L. Oliveira, S.M. Lima, T. Catunda, L.A.O. Nunes, J.H. Rohling, A.C. Bento, M.L. Baesso, High fluorescence quantum efficiency of 1.8 μm emission in Tm-doped low silica calcium aluminate glass determined by thermal lens spectrometry, Appl. Phys. Lett. 84 (3) (2004) 359–361 https://doi.org/10.1063/1.1640782. [4] B.G. Aitken, M.J. Dejneka, M.L. Powley, Tm-doped alkaline earth aluminate glass for optical amplification at 1460 nm, J. Non-Cryst. Solids 349 (2004) 115–119 https:// doi.org/10.1016/j.jnoncrysol.2004.08.217. [5] B.G. Aitken, M.L. Powley, R.M. Morena, B.Z. Hanson, Tm-doped aluminate glass fibers for S-band optical amplification, J. Non-Cryst. Solids 352 (2006) 488–493, http://dx.doi.org/10.1016/j.jnoncrysol.2005.11.067. [6] P.R. Foy, Fabrication and Characterization of Calcium Aluminate Fibers. A
33
Contents lists available at ScienceDirect
Journal of Non-Crystalline Solids journal homepage: www.elsevier.com/locate/jnoncrysol
Laser potential of calcium aluminate glasses B.I. Denker a b
a,b
, B.I. Galagan
a,b
, S.E. Sverchkov
a,⁎
T
Prokhorov General Physics Institute of the Russian Academy of Sciences, Moscow, Russia Center of Laser Technology and Material Scince, Moscow, Russia
A R T I C LE I N FO
A B S T R A C T
Keywords: Calcium aluminate glass Laser glass, tunable thulium laser
Deeply dehydrated calcium aluminate glasses doped by rare-earth ions were synthesized and investigated. Spectroscopic properties of Tm3+ and Ho3+ doped glasses were found attractive for fiber lasers emitting in 2–2.5 μm spectral range. Cascade laser action (at 2300 and 2100 nm) in Tm doped bulk calcium aluminate glass was demonstrated for the first time.
1. Introduction Glasses based on CaO-Al2O3 system (calcium aluminate glasses, CAGs) are relatively new materials that have already found their place in optical industry. Multicomponent CAG compositions can be crystallization stable enough to be produced in commercial volumes (several kilograms). Their attractive features are high infrared transmittance (up to 5 ÷ 6 μm) combined with excellent mechanical properties and thermal conductivity exceeding those of fused silica as well as excellent moisture resistance. In contrast to fused silica glass, these glasses can easily include up to 3 ÷ 5% mole of lanthanide oxides. Nevertheless, the amount of investigations devoted to spectroscopy and to potential laser applications of the rare earth activated CAGs is rather limited [1–5]. Limited interest may be explained by tough competition with well-developed activated silica fiber lasers and fluoride (mainly ZBLAN) fiber lasers. Indeed, the ecological niche suitable for activated CAGs and fibers may be rather narrow. But we believe it do exist because CAGs have much better infrared transmittance and rare earth solubility than fused silica as well as much better mechanical properties and moisture resistance than ZBLAN. This niche should include fiber lasers exploiting the rare-earth ions (RE3+) transitions that are noticeably or almost completely quenched by multiphonon relaxation in silica but still radiative in CAGs. The papers [3,4] have already disclosed some part of the area where activated CAGs are of interest: Tm-doped lasers and amplifiers operating at ~ 1.8 μm (3F4e3H6 transition) and at ~1.45 μm (3H4e3F4 transition). Of course, in case of low-gain glass lasers and amplifiers mostly fiber configuration is of interest. Optical fibers drawing of CAGs is a rather tricky procedure because of easy glass devitrification. Nevertheless, it was shown to be possible. For example, a relatively simple approach was realized in the dissertation [6]. The custom modification of the fiber draw furnace was
⁎
Corresponding author. E-mail address: [email protected] (S.E. Sverchkov).
https://doi.org/10.1016/j.jnoncrysol.2018.04.047 Received 19 March 2018; Received in revised form 24 April 2018; Accepted 25 April 2018 0022-3093/ © 2018 Elsevier B.V. All rights reserved.
required to achieve fiber drawing from the crystallization-prone CAG compositions. This included minimization of the heat zone susceptor width and adding water-cooling to narrow the heat zone temperature profile further. Using this technic the Schott calcium-aluminate IRG11 glass sample was successfully drawn into one-meter long Teflon coated 550 μm diameter fibers. The delivery of Er:YAG laser power via this fiber has also been also demonstrated. The fibers prepared in this work had no any glass cladding and no rare-earth activation of them was considered. A more advanced approach to fabrication of CAG optical fibers was developed in [5], where composite (multimaterial) fibers with Tm-activated core in a crystallization–stable silicate cladding were made. Their first attempt to fabricate a fiber with CAG core started from a rodin-tube preform consisting of a Tm-doped aluminate glass jacketed by a cladding tube of B-doped silica glass (10B2O3: 90SiO2). The preform was drawn into 125 μm diameter fiber with a 25–30 μm diameter core. Unfortunately, the core contained ~50 wt% SiO2 and ~5 wt% B2O3 due to intense diffusion during high-temperature drawing. The presence of boron and silicon in the core inevitably spoils the unique IR properties of the CAG glass. In order to overcome this serious obstacle the authors of [5] have developed special cladding glass compositions having softening temperatures close to that of the core glass. Using these special glasses and drawing at lower temperatures, fibers with Tmdoped silica-free calcium-aluminate core were successfully fabricated. Unfortunately, paper [5] gives no information about gain or laser properties of these fibers. All the cited papers show activated CAGs and fibers made on their base are novel interesting materials for near- and mid-infrared lasers and amplifiers. In the present paper, we would propose a laboratory synthesis procedure of deeply dehydrated CAGs. Deep dehydration is required for two purposes. First purpose is the adequate luminescence
Journal of Non-Crystalline Solids 496 (2018) 29–33
B.I. Denker et al.
decay characterization of activated glasses in the mid-infrared. The second purpose is the need for low absorption losses in lasing tests. Then we would investigate the luminescent properties of Tm, Ho and Er-activated deeply dehydrated CAGs in the spectral range where these glasses seem the most interesting - 2÷2.7 μm. Analysis of the laser potential of these activators and lasing tests in bulk glass will be added.
2. Glass fabrication procedure Our first task was to develop a laboratory fabrication procedure that gives CAG samples of high optical quality (absence of light scattering inclusions, deep dehydration and minimal striae). The developed CAG laboratory synthesis procedure noticeably differs from the methods used for it before. The main feature of this procedure is the usage of inductively heated vitreous carbon crucibles. The big advantage of vitreous carbon is its total non-wettability by the glass melt. For this reason, the prepared glass ingot easily drops out of the crucible. It should be noted, that the CAG compositions prepared in such utensils should not content components that can be reduced by carbon. In our glasses doped by Tm3+, Ho3+ or Er3+ no reduction was detected. Addition of Yb2O3 (that could be used for sensitization of all three above-mentioned dopants) resulted in partial reduction of Yb3+ to Yb2+. Therefore, hereinafter we have avoided using Yb. Infrared laser applications require deep dehydration of the active medium to prevent absorption losses and dopants luminescence quenching by OH-groups. Dehydration of the CAG melt was implemented by low-pressure (~0.05 at.) dry nitrogen atmosphere during synthesis procedure and by ~5% mol. of fluorides incorporated into the glass composition. Dry nitrogen atmosphere was prepared by liquid nitrogen evaporation. The basic glass composition we have used (46 CaO, 5 CaF2, 36 Al2O3, 6 MgO, 6 SrO, 1 Ln2O3, where Ln means this or that combination of La, Tm, Ho, and Er), was close to one of the most crystallization stable compositions that were found in [7]. The main difference was the usage of strontium oxide SrO instead of BaO since barium oxide turned to be too volatile at low atmospheric pressure. The ~30 g charge was loaded into the crucible and kept for 3–4 h at 1550–1600 °C and at lowered nitrogen pressure. Then, after the nitrogen pressure was risen to the atmospheric value, the high frequency power was switched off and the crucible with the melt was dragged out of the heat insulating ceramics for rapid cooling. Fig. 1 shows the appearance of as-melted glass ingots after their annealing at 750 °C. The ingots volume was clear from crystalline of
Fig. 2. Transmission spectra of a non-dehydrated (black curve, #1) and welldehydrated (#2) glass samples. Sample thicknesses are indicated near the curves.
gaseous inclusions (though the upper surface of them may be spotted by glassy foam and crystals). No light scatter of the red HeeNe laser can be visible in the samples volume by naked eye. Hydroxyl groups in CAG exhibit wide and intensive absorption band that can hinder laser action in glass. Therefore, deep dehydration is a necessary and very important condition for mid-infrared laser glasses. Fig. 2 shows the transmission spectra of two Tm-doped CA glasses recorded with Cary 5000 spectrophotometer. The first 6.5 mm thick sample (black curve) wasn't vacuum treated. Its spectrum exhibits hydroxyl absorption band (extinction ~0.2 cm−1) peaking at ~2920 nm. Blue and green curves correspond to glass plates of different thicknesses (19 and 6.5 mm respectively) made of a well-dehydrated glass sample #2. It can be seen that the spectra of dehydrated glass plates despite their different thicknesses coincide with the accuracy of about 1%. The difference between the blue and the green curves is purely stochastic and has nothing common with the hydroxyl absorption band. For this reason, we can state that OH-group absorption at 2920 nm in this glass is < 1% or 0.01/(1.9–0.65) = 0.008 cm−1. Practically we were not able to detect OH absorption in any of the dehydrated CAG samples having thickness of about 2 cm. Fig. 3 shows the results of differential thermal analysis of one of the
Fig. 1. The appearance of the CAG ingots. The ingots (left to right) are doped with Er2O3, Yb2O3, Ho2O3 and Tm2O3. Note the yellow-brownish color of Yb-doped sample due to Yb2+ presence. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) 30
Journal of Non-Crystalline Solids 496 (2018) 29–33
B.I. Denker et al.
Fig. 4. Emission spectra of Tm3+ and Ho3+ in CAG.
Fig. 3. Differential thermal analysis of CAG.
prepared glasses. The glass transition temperature Tg was determined to be 796 °C. The intensive exothermic peak at 992 °C corresponds to glass crystallization and the endothermic peak at 1372 °C corresponds to crystals melting. Due to high crystallization ability, fiber drawing of this glass is possible either in a narrow temperature gap between Tg and crystallization onset temperatures, or above the crystals melting temperature. Naturally, fabrication of optical fiber with this glass core needs in the appropriate cladding glass selection.
3. Spectral peculiarities of rare Earth ions iN CAG Luminescent lifetime measurements of Tm3+ and Ho3+ and Er3+ in CAGs were made using pulsed (~10 ns) widely tunable OPO excitation, a grating monochromator and a double-stage TEC cooled 2.6 μm cut-off InGaAs photodiode. In case of 3F4 Tm3+ and 5I7 Ho3+ levels lifetime measurements special precautions were taken to avoid reabsorption of the luminescent light in the samples. These measurements were fulfilled using tiny (< 1 mm3) glass splinters immersed into highly refracting liquid - glycerin, n~1.55. Otherwise, the reabsorption effect can cause quite noticeable lifetime overestimations. The decay functions were averaged over 256 laser pulses at least. All the lifetime measurements were statistically checked by using several glass samples obtained in different synthesis processes. All the registered luminescence decay functions were close to exponential. Taking into account the statistical data (the lifetime measurements in several independently synthetized glass samples) and the characteristics of the registration system we have estimated the lifetime measurements accuracy as ± 5 μs for the lifetimes shorter than 1 ms, ± 20 μs for Tm3+ 3 F4 excited state and ± 100 μs for Ho3+ 5I7 level. The doping concentrations were 1% mol. For Er2O3 and Ho2O3 (that corresponds to ~4.75 × 1020 of dopant ions per cm3) and 0.21% mol. (1020 ions/cm3) for Tm2O3. The latter concentration is small enough to neglect the effect of cross-relaxation of 3H4 manifold [4]. Still it was
Fig. 5. Transmission spectrum of the laser rod. Arrow indicates the pumping wavelength.
high enough for laser tests (see below). Table 1 presents the measured e-fold lifetimes in CAG in comparison with literature data for fused silica. It contains also the radiative lifetimes of Tm3+ levels in calcium aluminate glass of a close composition. These data were calculated using Judd-Ofelt method for calcium aluminate glass of a close composition [4]. The ratio of the experimental and calculated values gives the quantum yields of the corresponding levels: 26% for 3H4 and 60% for 3F4. Both these values are big enough for efficient and low-threshold lasing. The same can be said about 5I7 Ho3+ level lifetime. All the mentioned lifetimes are about an order of magnitude longer in CAG than in fused silica. It means that the lasing
Table 1 Metastable levels lifetimes, μs. Dopant ion and its metastable manifold Tm
3+ 3
H4
Tm
3+ 3
F4
Ho3+ 5I7 Ho3+ 5I6 Er3+ 4I11/2
CAG
Fused silica
220 ± 5 (experimental, this work, coincides well with [4] data) 845 (radiative, [4]) 2500 ± 20 (experimental, this work) 4162 (radiative, [4]) 9400 ± 100 50 ± 5 100 ± 5
20 [8]
31
420 [8] 600 [9] Totally quenched Totally quenched
Journal of Non-Crystalline Solids 496 (2018) 29–33
B.I. Denker et al.
Fig. 6. A photo of the laser setup.
4. Bulk CAG lasing demonstration The lasing test was held with Tm doped CAG since we had a suitable high-power pulsed pump source for it – a flashlamp pumped ruby laser. A 30 mm long flat-parallel polished Tm:CAG sample was pumped by a pulsed ruby laser (incident pump pulse energy up to 0.7 J, pump beam diameter ~2 mm). Fig. 5 presents the transmission spectra of laser element. The laser cavity was formed by a concave (R = 100 mm) mirror with high reflectivity in a wide spectral range from ~2.0 to ~2.4 μm, and flat outcoupler (0.15% at 2300 nm and 1% at 2100 nm). Fig. 6 is a photograph of a CAG glass element being pumped by a ruby laser in the cavity. The lasing wavelengths were analyzed using the same monochromator and photodetector as in luminescent measurements. Double-wavelength cascade lasing was observed. The lasing wavelengths varied from pulse to pulse in the ranges shown in the Fig. 7 that presents the oscillograms of the pump and Tm: CAG lasing pulses. 5. Conclusion Fig. 7. Oscillograms of the ruby laser pump pulse and lasing at two optical transitions of thulium ions (3H4 - 3H5 and 3F4 - 3H6).
Deeply dehydrated rare-earth ions doped samples of calcium aluminate glasses were prepared by high-frequency melting in vitreous carbon crucibles. The differential thermal analysis showed that fiber drawing of this glass is quite possible at relatively low temperatures (~900 °C). Luminescent properties of thulium and holmium in CAG were found promising for tunable 2–2.5 μm laser applications. Cascade laser action (at 2300 and 2100 nm) in Tm doped bulk calcium aluminate glass was demonstrated for the first time.
thresholds in CAG should also be an order of magnitude lower all other parameters being equal. The lifetimes of 4I11/2 Er3+ and 5I6 Ho3+ levels in CAG (optical transitions at 2.7 and 2.9 μm) seem to be too short for any practical laser. Therefore, we would not consider them further. The emission spectra presented in Fig. 4 were measured using CW laser diode excitation at 800 nm (for Tm doped glass) and 668 nm (for Ho doped glass) and the same monochromator and photodiode as in lifetime measurements. In comparison to other oxide or fluoride glasses, the presented spectra of Tm3+ and Ho3+ in CAG look rather unusual due to enhanced width and red shift. It is especially notable for 3F4 - 3H6 Tm3+ transition that extends up to ~2.4 μm and overlaps with ~2.3 μm 3 H4 – 3H5 transition. Thus, the total tuning range of Tm-doped CAG laser can extend from ~1.9 to ~2.4 μm that is hardly possible in any other type of glass. In case of Ho3+ doped glasses the tuning range is narrower but also extends to ~2.2 μm.
Acknowledgements The authors are thankful to A. M. Kut'in and A. D. Plekhovich for DTA measurements. The investigation was supported by RFBR grant 1702-00369. The authors have no competing interests to declare. References [1] D.F. de Sousa, J.A. Sampaio, L.A.O. Nunes, Energy transfer and the 2.8-μm emission of Er3+- and Yb3+-doped low silica content calcium aluminate glasses, Phys. Rev. B 62 (5) (2000) 3176–3180 https://doi.org/10.1103/PhysRevB.62.3176.
32
Journal of Non-Crystalline Solids 496 (2018) 29–33
B.I. Denker et al.
Dissertation Submitted to the Graduate School - New Brunswick Rutgers, the State University of New Jersey, https://rucore.libraries.rutgers.edu/rutgers-lib/24549/ PDF/1/play/, (2008). [7] E.V. Uhlmann, Glass Forming Ability, Structure and Spectroscopic Properties of Silica-free Calcium Aluminate Based Glasses. A Dissertation Submitted to the Faculty of the Department of Materials Science and Engineering for the Degree of Doctor of Philosophy in the Graduate College, The University of Arisona, https://core.ac.uk/ download/pdf/143732190.pdf, (1995). [8] B.M. Walsh, N.P. Barnes, Comparison of Tm:ZBLAN and Tm: silica fiber lasers; spectroscopy and tunable pulsed laser operation around 1.9 μm, Appl. Phys. B Lasers Opt. 78 (2004) 325–333 https://doi.org/10.1007/s00340-003-1393-2. [9] D.C. Hanna, R.M. Percival, R.G. Smart, J.E. Townsend, A.C. Tropper, Continuouswave oscillation of thulium-doped silica fibre laser, Electron. Lett. 25 (1989) 593–594 https://doi.org/10.1049/el:19890403.
[2] D.F. De Sousa, L.A.O. Nunes, J.H. Rohling, M.L. Basso, Laser emission at 1077 nm in Nd3+-doped calcium aluminosilicate glass, Appl. Phys. B 77 (2003) 59–63 https:// doi.org/10.1007/s00340-003-1247-y. [3] S.L. Oliveira, S.M. Lima, T. Catunda, L.A.O. Nunes, J.H. Rohling, A.C. Bento, M.L. Baesso, High fluorescence quantum efficiency of 1.8 μm emission in Tm-doped low silica calcium aluminate glass determined by thermal lens spectrometry, Appl. Phys. Lett. 84 (3) (2004) 359–361 https://doi.org/10.1063/1.1640782. [4] B.G. Aitken, M.J. Dejneka, M.L. Powley, Tm-doped alkaline earth aluminate glass for optical amplification at 1460 nm, J. Non-Cryst. Solids 349 (2004) 115–119 https:// doi.org/10.1016/j.jnoncrysol.2004.08.217. [5] B.G. Aitken, M.L. Powley, R.M. Morena, B.Z. Hanson, Tm-doped aluminate glass fibers for S-band optical amplification, J. Non-Cryst. Solids 352 (2006) 488–493, http://dx.doi.org/10.1016/j.jnoncrysol.2005.11.067. [6] P.R. Foy, Fabrication and Characterization of Calcium Aluminate Fibers. A
33